Warning: The article below is a bit of a monster (c. 30 pages) and contravenes all guidance on blogposts needing to be short, digestible expansions of one or two salient points. I've spent weeks writing it but I, honestly, don't recommend that you read it. Rather, I suggest you skim the titles and perhaps dig a bit into the aspects that interest you. Then
How will AI change sustainable investment?
I have spent a lot of time recently:
- assessing how AI will change the dynamics, practices and processes that make up the sustainable investment value chain (Spoiler: Fundamentally and quickly!) and
- exploring what practitioners can do to ensure that these changes are directed in a way that are beneficial to their clients, to wider society and to environmental sustainability (Spoiler: Also, a lot, if we move fast).
This article contains my work-in-progress thinking on the different aspects and areas of the value chain that I think are likely to be affected and the ways that change will evolve and manifest.
I have organised these under the questions that I have been asking myself in recognition of the huge amount that we still don’t know and the fact that many of my ‘answers’ are still provisional.
Then again, development is happening so fast that ‘wait and see’ is simply not an option. ‘Analyse, take a view, act and constantly re-assess’ appears to be the only appropriate modus operandi.
In which spirit, here’s my first cut (published 20 July 2026). I’ll update the article as I explore further and learn more.
Feel free to get in touch to discuss and challenge any of these ideas or to share your own thoughts on the relationship between sustainable investment and this revolutionary suite of technologies.
1] How will AI change the sustainable investment value chain?
1a] AI enables sustainable investment to ‘reset’ to what it always should have been
In AI, we have a tool that enables sustainable investment to become what it always should have been: a set of investment strategies that allocate capital in ways that:
- deliver the totality of investor preferences – not only their financial preferences.
- contribute to and benefit from alignment with specific sustainability trends,
- facilitate the broader achievement of sustainable development and
As I discuss below, it is a tool which – if applied with a clear-eyed capital allocation perspective - should be able to rescue sustainable investment from the quagmire of data minutiae, single-issue activism and regulatory reporting that some areas of the value chain seem to find themselves trapped within.
These ‘reset’ benefits will be achieved, I believe, if we choose to apply AI with:
- a clear focus on investment objectives,
- a rigorous commitment to empowering the human and,
- strong attention to environmental, social, ethical, economic and financial considerations
For too long, too much (but not all) of the sustainable investment value chain has been driven by the inputs that are possible (What data can we get?) rather than by outcomes that we seek (What are we actually trying to achieve?).
Now that AI delivers all of the data inputs that we can ever imagine (and more) directly to our fingertips, we have a fresh opportunity to focus on these outcomes.
1b] Focus on the upside promise
By using AI to improve the depth, breadth and quality of contextual information (about trends and companies) that investors and analysts are able to access (at negligible cost), we can:
- Integration: improve the effectiveness with which investors can integrate sustainability factors into valuation and capital allocation,
- Engagement: focus engagement activity on issues that are material from a sustainability or financial perspective,
- Communications: improve significantly the efficiency of communications between companies and investors of sustainability issues
- Operations: improve significantly the efficiency of investment operations
- Reporting: improve both the efficiency and quality (hence readability and access) of investor reporting to clients on sustainable investment activity
Importantly, information alone does not deliver change. Information only catalyses change when it combines with:
- (client) demand for change
- the availability of an articulated alternative and the confidence that it can be attained
- tools for reaching this alternative
- … in a way that surmounts the (inevitable inertia) barriers to change and is cognisant of any ‘worse case’ scenarios to be avoided
In respect of ‘integration’, ‘engagement’, ‘communications’, ‘operations’ and ‘reporting’, these other conditions exist such that AI should land in a sustainable investment value chain that is ready for change.
1c] Of course, there’s a less positive possibility
The existence of positive outcomes does not mean that they will necessarily be delivered. Worse case scenarios exist and it is important that we are self-aware as an industry and aware of these potential pitfalls and sidetracks so that we can avoid these.
- Integration: It will be possible use AI to generate ever broader and more granular ESG datasets in continued pursuit quants-based linkages between companies’ sustainability ‘performance’ and investment performance. This is (IMHO) misguided and will result in further time lost in pursuit of a fictional silver bullet solution.
- Integration: AI, by its nature, identifies the most commonly-given answers rather than the correct answers or the modal ideas rather the most forward-looking, innovative or relevant ideas. This reinforces groupthink and increases conviction around what is already priced in rather than in ‘investable insight’. Self-evidently, this should be avoided
- Engagement: AI tools could be used to identify every environmental or social transgression committed by every company on even the most marginal topics as a precursor to ever more micro-managed engagement across portfolios
- Engagment: We might be tempted to allow AI-automation to increase the quantity of investor-company contact without first ensuring that the focus, direction and quality of this contact also improves. This will lead to wasted effort throughout the value chain with minimum impact.
- Communications: Investor to company communications: It will be possible to use AI to speed up the process and hence volume of our interactions without improving the quality. Again, this will result in wasted effort.
- Operations: There is a risk that we could use AI to create duplicate (belt and braces) processes or spend time automating processes because we can rather than because it delivers improvement. Again, work will be done without improvement being delivered.
- Reporting: Finally, there is a risk that might use AI to develop ever-contracting circles in which machine-generated data from companies flows to machine-reading at research providers and investors, and through machine-written reports to regulators and to beneficiary investor clients without anyone paying any attention to it whatsoever.
So, let's avoid as many of these situations as we can and focus instead on virtuous circles whereby machines empower humans at companies to communicate better with human decision-making by investors and to financial and real-world impact that is understood by beneficiary investors.
1d] Mapping the adoption landscape (the baseline)
At this stage, much is unknown about the ways that AI is being applied within sustainable investment today and what the immediate development trajectory looks like for investors, for research providers and for companies.
From conversations so far, I understand there to be a wide variation in adoption levels across the sustainable investment value chain from…
- people who use AI tools as ‘advanced Google’ through to…
- people who are using AI to develop thinking, to reconfigure information flows and to systematically automate workflows and practices
So, over the summer, Neil Brown and I (www.coduspartners.com) plan to build out from this anecdotal picture – through surveys and in-depth interviews. We will map the extent and depth of AI-adoption by sustainable investors in a way that:
- provides immediate ideas and feedback to the people we speak with
- identifies specific investment applications for individual firms and
- helps the wider sustainable investment value chain adopt AI in a way that adds value to industry participants, to their clients and to wider society and the environment.
If you fancy bouncing development ideas around with us or mapping your own progress against that of the wider value chain, do get in touch. (
2] How can we maximise the benefit and minimise the harm of AI to sustainable investment?
2a] First and foremost, empower the human
There are two ways of thinking about AI. We can see it as a way to:
- Replace the human, or
- Empower the human
These are not mutually exclusive. Indeed, both are already happening and will continue to happen simultaneously.
However, the order in which we are motivated to apply them matters enormously. For multiple (financial and effectiveness as well as social) reasons, I believe that we must seek to:
- First, empower the human
- Secondly, offload low-value human tasks to AI and free-up humans for ‘higher-order’ work.
- Finally, see ‘replace the human’ as a margin-enhancing side benefit of the first two
While this is true for all types of investment, it is particularly true for sustainable investment which involves:
- A large degree of human judgement about how material sustainability factors will become for the valuation of companies if / as / when society decides or is forced to confront environmental and social realities.
- A shared cultural understanding between those buying sustainable investment products and those delivering them – that operates behind and beyond the contractual relationship between them.
Both facets are instinctively understood by most sustainable investment professionals but none are particularly well-articulated – certainly not in ways that communicated to (or inferred by) machines.
These two facets of sustainable investment involve high degrees of (often nuanced) human judgement:
- In the former case, to discern signal from noise;
- In the latter case, to understand the ‘intent’ of human investors
The result is that we are likely to end up with much better final outcomes (for everyone) if we use AI as part of a process that involves articulating and improving how humans deliver sustainable investment outcomes and - as this understanding evolves – gradually and incrementally automate where appropriate.
The alternative approaches:
- allowing AI to infer desirable outcomes from currently articulated practice
- allowing AI to develop practices based on currently observed outcomes
… would miss important (but often not articulated) aspects of practice and would also automate sub-optimal current practices.
Put another way, if we seek to replace human processes without understanding them first, it’s not likely to go well for anyone.
We need to use AI to integrate sustainability factors better into human investment decision-making; we need to use AI to improve the way companies and human investors communicate and engage; we need to use AI to improve the way that human investors communicate to their human clients.
Then, as we do this, we need to partner with AI and teach it to automate this better version of our practices and the outcomes that we seek.
2b] Get ahead of and control how AI replaces the human
To an extent, AI will enable investors to improve quality and to grow revenues. However, it is also likely that AI will be used to increase margin by cutting (human) costs. As ever there are constructive and destructive ways to cut costs.
I hope that sustainable investment practitioners will get ahead of the curve within their businesses and quickly cut the costs that do not contribute to social, environmental, ethical, economic and financial outcomes so that they can maintain their spending on things that do contribute to these core deliverables.
Doing this, however, will need to be an active process. Passivity and non-participation will, inevitably, lead to human costs being cut detrimentally.
2c] The cost of people vs the cost of tokens
Media comment appears to have swung recently:
- From: “AI will replace humans”
- To: “Aaaargh. Tokens cost as much as humans”
At this stage, I am unconvinced by either poles or the terms in which this is presented – particularly within the context of sustainable investment teams.
We need first to think about how people with token can be more effective than people without.
Seeking one-for-one replacements may be suitable for IT firms with hundreds of coders, it doesn’t work for sustainable investment teams of 3-4 people.
Achieving 20% better output with 20% fewer resources seems like a more reasonable starting target.
I’m fine with the tokens working our Fridays!
3] (When) will AI replace humans in the sustainable investment value chain?
3a] Investment management = information processing + button pressing …
Fundamentally, the process of investment management can be reduced to information processing and button pressing.
With AI, we now have a technology that can gather, process and weigh information more quickly, for longer and with far greater bandwidth than any human brain.
… and Claude in Chrome … can press buttons.
Hmmm …
3b] … + trust
However, the other less tangible but hugely important aspect of investment management is the human characteristic of trust.
Trust is partly engendered by information, partly by reputation and track record, partly by regulation, and partly by deeply-human factors such as: the ability to meet a person, to challenge them, to gauge their responses and to form a view about whether that person can be trusted to deploy your money in line with your ethics.
AI can enhance trust - by adding a verification layer to human claims, by extending the information available about counterparties and by surfacing inconsistencies in data and reports.
However, AI also introduces its own trust challenges: its information has known (and possibly unknown) reliability issues such that the technology cannot be a substitute for human trust. It is, at best, a complement to it.
Notably, humans do learn to trust machines to make decisions (or else the aviation industry would struggle). However, our willingness to trust our money to decision-making by machines alone is something that will likely evolve steadily rather than immediately.
Neil Brown argues that “agents don’t get invited to dinner parties” to convey the lack of constraints that autonomous agents experience. He explains that a human being contemplating fraud will experience a heady brew of dopamine, cortisol and endorphins as they weigh up the evolutionary boost of providing for their family with social anxiety around being shunned. By contrast, an AI-agent has no such experience and no such self-regulation – such that it is, by design, less ‘trustworthy’.
In this respect, it is notable that sustainable investment requires (and therefore reinforces) a second bond of trust between client and investment manager and between investment manager and company. Not only does there need to be a bond of trust around financial matters, but also around social, environmental, and / or ethical priorities. Arguably, for this reason, sustainable investment has a deeper ‘moat against the machines’ than other forms of investment.
3c] … + judgement
The additional factor between information processing and ‘button-pressing’ is judgement.
Judgement – in the context of sustainable investment - comprises practices such as:
- the weighing of scenarios,
- the prioritisation of valuation factors,
- the development of investment theses,
- engagement with and influence on companies and
- the production of final investment recommendations and decisions
- assessment over what and how to communicate to clients
In some respects, these practices can be reduced to statistically-weighted decision-gates (and hence automated by AI); in other respects the human brain will likely prove the most effective place for decisions requiring judgement to be taken.
Even under a situation where automation of judgement is sought, human beings will be needed to identify and track the development of these decision-gates.
Some automation will be delivered by inferring the decision-gates from observed outcomes; much, I suspect, will be developed by articulating and codifying human processes from bottom-up. Ultimately, it seems likely that applying both top down (inference) and bottom up (articulation) processes together will deliver the best results.
While I can envisage the ways that human judgement will be gradually replaced within the investment process, I find it hard to predict how quickly this will happen.
I suspect that:
- To remain competitive, investors will need to move proactively now to train models and to start identifying which of their judgement calls might be effectively automated and which should remain within the human domain
- Human judgement will be an integral part of any ‘training’ the machines to ‘judge’
- The more nuanced the judgement required, the longer the human will likely remain ‘in-the-loop’.
- The direction of travel (towards greater automation) is clear
- We must not, however, expect humans to become merely ‘checkers-in-the-loop’. This does not play to the strength of the human mind.
- It is important not to see ‘in-the-loop’ / ‘out-of-the-loop’ as a binary condition. The most effective fundamental analysts are likely to be those that manage progression from AI-out-of-the-loop, through AI-in-the-loop, centaur-like-thinking, cyborg-like-thinking and on to human-in-the-loop practice – adjusting to whatever ‘mode’ delivers most effective outcomes.
While I find it difficult to predict (for the various reasons described above) how quickly the evolution towards automation will occur, I feel confident that considerable levels of human input will be required before it is achieved.
3d] … + creativity
AI generates the highest-probability-predicted answer to any given prompt. If the human prompt is creative, the answer may seem creative. However, it is not.
A time will likely come when AI has sufficient breadth, experience, processing and (most important) filtering power to produce genuinely creative ideas by extrapolating from apparently disconnected ideas. We are not there yet. We still need humans to feed creativity into the process.
3e] Summarising and the implications for sustainable investment
While progress will not be linear, it seems reasonable to conclude that, the order in which human-processes will be replaced by automation runs thus:
- First, information processing
- Second, judgement (provided that an effective combination of bottom-up instruction and top-down inference is applied)
- Third, creativity. Although fully-creative ‘what ifs’ can never be generated by machines, it can be replaced by judgment over an expanded range of possibilities
- Last, trust, which rather depends on social preferences whose importance I recognise but which I feel unqualified to assess
Sustainable investors have long downplayed the role of ‘judgement’ & ‘trust’ in their investment processes. Perhaps they believe that that these will be perceived as ‘subjectivity’ or as the dominance of ‘values’ over ‘value’.
In their place, we have lionised ‘quantitative data’ to confer a pseudo-financial legitimacy to our activity.
In reality, of course, judgement (which, if analytically reached, does not have to be the same as ‘subjectivity’) has always been at the heart of the (sustainable active) investment proposition.
As more and more data processing is subcontracted to machines, we will need to get better at articulating the role and contribution of judgement, creativity and trust within sustainable investment – particularly as these features are stronger in sustainable investment than in other investment strategies and are important aspects of competitive advantage.
If we consider the application of AI to sustainable investment to focus on (or to be portrayed as focusing on) the costs and efficiency of data-processing, we all lose. If we consider it to be about unlocking and improving judgement, creativity and trust, we (all) win.
4] How should sustainable investment professionals act today?
4a] Move now; move fast
(NB: ‘Now’ means today. Not next month or next quarter. If you wait until next year, you may not have a business or a role left to which you can apply your understanding of AI).
Sustainable investors need to move fast to adopt AI for three reasons:
- Set-up takes time – while AI can automate processes in a way that saves time and improves impact, there is no denying that the set-up takes time. Although AI agents can infer and can prompt proactively, they are most effective when they are dealing with articulated processes. While some firms have formally documented their research and management processes, most haven’t. Most need to surface and assess their own processes before they are able to expose these to automation by AI.
- The deeper changes - to culture and to working practices - take longer to implement than the changes to data handling, to workflows and to reporting
- The fintech ‘bros’ and their AI agents aren't waiting for conservative investors to evolve steadily. They are cracking on fast and breaking things. Some will go bust. Others will sweep the table while investment institutions are stuck in the third of six risk evaluation and compliance review meetings.
- Significantly, moving and moving fast generates its own momentum as learning makes more learning possible and humans start to capture the exponential capabilities of AI-assisted work. “I’ve missed the AI opportunity” becomes a self-fulfilling prophecy. However, “I started a bit late but am catching up” is also one. So, get on it. Experiment. Move fast and … you don’t actually need to break things if you think a bit as you go.
4b] Focus on outcomes and the (human) resources and workflows that deliver these
As discussed earlier, sustainable investment has struggled for too long with an ‘inputs-focused’ model whereby ‘whatever data we can get’ is appended to a loose thesis about 'responsibility being rewarded'. There are two problems with this approach:
- It is clear that – in the current political environment – responsibility often isn’t rewarded; in fact, it’s often the opposite: irresponsible, unsustainable and anti-societal practice can thrive and can be rewarded. You have to be selective. The fact that a general top-down thesis doesn’t work does not mean that there are not plenty of bottom-up case-specific theses that do.
- This ‘spaghetti-to-the-wall-and-see-what-sticks’ approach is hugely inefficient as it requires companies to gather and report on swathes of data that might just be (but rarely is) financially-material. The opportunity cost of time spent reporting rather than doing seems to outweigh the management focus that a need to report delivers.
If we seek to automate this ‘inputs-focused’ model, we will simply accelerate the production of non-material information (until companies, research providers and investors all drown).
Rather than focusing on ‘inputs’, investors should begin their AI journey with a clear focus on their desired outcomes (often comprising client satisfaction, regulatory compliance, financial performance and environmental sustainability), and identify how AI can support workflows that empower humans to deliver these.
This will prove hugely more efficient and ultimately more impactful.
It may be that (some) comparable quantitative data contributes to these outcomes and needs to be collected. It may be, however, that it turns out to be less necessary than is commonly (and, IMHO, erroneously) assumed.
4c] Make yourselves visible (as organisations, teams and individuals
Dear humans,
A case can be made that your primary value proposition lies in the fact that you are human beings and that you can think in deeper and more nuanced ways than machines and that you can empathize with and understand other humans.
(This applies whether the human you face is a CFO stating 5% growth but with unconvincing eyes or a client who is scared or optimistic about our ability to tackle climate change and wants an investment strategy to match.)
Not to put too fine a point on it, if your primary value lies in being human, you would do well to display that humanity by putting a picture of your human face on your firm’s website and by describing your coverage, interests, priorities and needs.
I have now spent fifteen years trying to persuade investment analysts and companies to make themselves and their interests and priorities more visible to each other as I believe that:
- visibility is a precondition to connection
- connection improves understanding, and
- understanding improves investment decision making.
I have had successes and I have fallen short on occasion. I have always made the case in positive terms.
However, AI now also makes the case in negative terms: “If you don't make yourself and your interests visible to AI, you will not be presented to the workflows of counterparties, and you will effectively not exist in some domains that are significant to you.”
(There is not yet an AI-version of me running around and checking that the analyst covering the Food Producers sector for Investor A is found and invited to Food Company B’s sustainable investment roadshow … although I am building one.)
(Notably, LinkedIn doesn't count as it largely prohibits machine reading.)
- If you are an investment analyst, this matters as your ability to participate in and extract ideas from the investment debate depends on you being included in that debate.
- If you are an IRO or CSO at a company that wants sustainable investors to properly understand your company, you'll need to make yourself and your story visible to them before AI reduces your efforts and activities to bare metrics and numbers.
- If you work for a research provider, your whole future value proposition likely centres on your ability to infer, to intuit, to prioritise and to judge. Data processing will be done by machines; human skills have human faces. Use yours.
(PS I have resolved to get a new headshot of myself taken and uploaded to all of my online profile as soon as I have finished this article!)
4d] Train, train, train.
Some people (‘beginners’) mainly use AI as an enhanced search engine.
Other people (‘confident explorers’) use it to structure and refine their thinking and to automate their workflows, their information processing and their decision-making
Over a relatively short period of time, these more sophisticated practitioners will develop considerable and accelerating competitive advantage over the former – not least because they will be supported by an army of agents working at the speed of light.
However, today, the gap between the two is little more than a small amount of empowering training, and some enthusiastic experimentation.
If you're in the ‘beginners’ category, now is the time to leapfrog into the ‘confident explorers’ category before it becomes too late.
4e] Think existentially and act immediately.
AI’s current capabilities should be causing investors to ask themselves: Does AI present an existential threat to my firm / to my role?
It should also cause them to ask: How can I use AI today to improve the way I work tomorrow?
The answer to the existential question should not prevent anyone from asking the immediate action question. It is only by addressing the immediate action question and embarking on a journey of discovery that you will ever be able to accurately answer the existential question and certainly the only way you will be able to find a positive answer to it.
Of course, wherever possible, you will be likely to try to create a ‘line-of-sight’ between the immediate and the long-term. This ‘line-of-sight’ is desirable and it will change. However, it should certainly not be a pre-condition to immediate action.
4f] Challenge yourself every day; push on through the setbacks
Understand the pathway to proficiency (from search, via ‘Centaur’ and ‘Cyborg’ thinking to various phases of ‘X-in-the-loop’ practice) and push yourself to progress a little further each day.
… and remember, AI is really annoying! The way it operates (under the bonnet) means that it makes errors and does not follow instructions and takes shortcuts that end up being long deviations for you.
Your new digital colleague is keen to help but is coming at the subjects that are relevant to you and the practices that you deploy fast with only a basic contextual understanding of ‘what generic investors might need’ – not what your specific firm needs or how it goes about getting this.
It needs training in the way that you were trained – over an extended period, through clear guidance, repetition and firm correction.
Think of AI like an exceptionally gifted child: full of ability and full of energy but often misguided, error prone, frequently annoying and requiring clear direction.
Belligerence is a virtue. Rage against the machine until the machine does what you want.
4g] Does SpaceX matter?
It is easy to be distracted by newsflow around SpaceX's valuation (I can’t wrap my head around its value proposition yet), around OpenAI’s potential IPO date, around the arrival date (or not) of artificial general intelligence or around the percentage of global output that will be supported by / supplanted by AI-derived technology.
These are all critical questions for investors to answer in respect of their investment decision-making.
They are not, however, critical to their investment processes. The technology that exists today (even you buy none of future prospects story for the technology or the companies who deliver it) is more than enough to revolutionise significant swathes of investment-related processes.
We can move ahead with these whether or not SpaceX flies.
5] How will AI affect DATA gathering and processing (and, by extension, RATINGS) within sustainable investment?
5a] AI will improve speed and efficiency of data processing and reduce its cost
The easy answer: AI will make data gathering and processing faster and cheaper.
That is true and it matters.
However, it is neither the most important nor the most interesting change. It’s just something that is obvious and needs to be executed.
5b] How AI will change data gathering and processing? Who will lead the change?
For answers to this question, I am indebted to Krista Tukiainen of Arctal and her post ‘The cost of data is collapsing but its value isn't’. I would encourage anyone interested in this question to read her thoughtful ‘practitioner’s read on what AI does to the economics of building sustainability and thematic datasets’.
That is not to say that I agree entirely with Krista. In particular, I think I differ from her view on the ultimate value of such datasets.
However, I find her evaluation of the economics and competitive dynamics that will drive evolution instructive and helpful.
What, perhaps, I would add to Krista’s evaluation is that:
- Cost and quality are only two dimensions of investors’ rationale for selecting services from specific data providers timeliness
- Incumbency and client inertia are also critical determinants in this regard
I fully endorse and highlight the points that Krista makes about:
- the need for judgement before (rather than after) data collection
- the importance of building services around client needs
- the likelihood that many datasets will not (after judgement is applied) be built at all
She writes “When production is close to free, the constraint rises from whether you can build a dataset to whether you should, and if it is the right outcome”.
Everything in her evaluation screams ‘hybrid’ at me.
I think we will see:
- hybrids between incumbents and start-ups
- hybrids between humans and machines
- hybrids between in-house and outsourced data collection
It seems likely that the winners over the medium terms will be those firms and individuals that select most efficiently and hybridise most effectively.
5c] Pricing pressure from clients and new entrants
AI will give asset managers the ability to gather and refresh data directly themselves (at negligible cost). This will empower them to press suppliers to price reductions.
At the same time, AI-first new entrants to the market will increase the pressure on incumbents.
Even so, I don’t think this will lead to widespread switching. Other factors (Liability offloading, brand, incumbent installation, regulatory requirements, inertia etc) have all proved strong barriers to switching in the past and will likely do so again.
5d] AI will kill the ‘data monkey approach’ to ESG.
The more important change is this: AI will finally and definitively demonstrate that many in the sustainable investment value chain have been placing unrealistic and inappropriate demands on ‘data’ and on companies’ sustainability ‘performance’. These demands have never been met; they will never be met and AI is about to show us why.
For years it has been argued that ‘if only we had better, more reliable, more comparable, more granular data, the investment case for sustainability would become self-evident.’
This argument hasn’t worked. IMHO, it will never work. It is based on several false premises.
In financial markets, data is imperfect. That’s the point. If data were perfect, pricing would be automatic and we would have no need for analysts or active investors at all.
AI will not resolve this by challenging head-on the argument for granular comparable data. I have tried this on hundreds of occasions over the years and – in my experience – there is no convincing the believers in data. It is an article of faith that is now embedded over a generation of sustainable investment professionals and supported by regulation.
In spite of being nonsensical, it is also unopposable.
Instead, AI will resolve the question by providing the data, testing it to destruction, then asking to provide more … and more … and more … until the argument that ‘if only we had better data’ finally collapses under its own weight.
You can’t prove a negative. (You can’t disprove the ‘if only…’). Well, you couldn’t until AI – within its ability to search to exhaustion – arrived.
5e] On balance, AI will improve the accuracy of ESG data
ESG data has known reliability issues. AI-generated data has known reliability issues (and, possibly, some unknown ones).
AI’s own reliability issues will likely introduce errors to ESG data. However, it will also correct errors generated by current ESG data practices.
That said, the ability of AI to run multiple agents, in parallel, at the speed of light in a way that cross-checks for errors is a powerful addition to the dynamic.
It leads me to conclude that the application of AI to ESG data gathering will – in the first instance - reduce errors in ESG data more than it increases them and ultimately result these to a negligible quantity.
This will be significantly enhanced when companies’ own disclosures are made more readily machine-readable.
Although, as above, whether we will need (or ever needed) that data remains, for me, an open question…
5f] AI could lead the shift from an ‘input-led’ approach to information gathering an ‘investable impact’ approach
… particularly, if we stop asking: How does this company perform against a constructed notional of sustainability best practice?
… and start asking: To which sustainability themes and trends is this company exposed? To what degree? How well is it managing those exposures? What are the catalysts for this exposure and management (or lack of) to become relevant to price?
AI gives us the processing power to pursue these second questions at scale, for the first time.
It enables analysts to access and understand the contextual information that is needed for them to fill in the gaps and draw clear ‘lines of sight’ between sustainability trends and stock price performance rather than simply depend on an assumed blanket relationship between responsibility and returns.
5g] Beavers, beetles and contextual information
Over the years, I have come to the conclusion that the vast majority of errors in ESG data and its use arises not from the accuracy of the data itself (although this can be an issue). Rather, errors arise from a failure to contextualise data appropriately. Beavers and beetles spring immediately to mind, but hundreds of other examples tell the same story.
Equally, the difficulties some investors appear to have linking sustainability factors to valuation lies (I suspect) in their focus on quantitative data rather than on contextual information.
As our work for WBCSD last year (Demystifying Investor Sustainability Information Needs and Use), it is ‘contextual information’ that does the heavy lifting around the integration of sustainability into valuation. ‘Quantitative data’ is a marginal player.
AI makes it much easier to gather relevant, in-depth, contextual information about industries, competitive dynamics, value chain pressures and sustainability issues. All you need to do is know how to ask it the right questions…
6] (How) will AI change the supply of sustainable investment RESEARCH?
6a] The practice of research
At SRI-Connect, we see the provision of sustainable investment RESEARCH as being a fundamentally different product from the provision of ESG DATA and RATINGS.
Research is distinguished by its objective of:
- identifying the specific sustainability-related opportunities and risks that face sectors and individual companies
- setting these within the context of the business environment and objectives of those companies
- with the objective of making investment decisions and/or identifying engagement situations.
Whereas the information gathered and processed for DATA and RATINGS typically aims to be systematic and comparable; the information gathered for RESEARCH is often company-specific and idiosyncratic.
The different nature of information sought for these different product outputs means that AI is likely to be used in different ways:
- Whereas DATA and RATINGS will value the speed, breadth and accuracy that AI delivers
- RESEARCH will value the depth, context and judgement enhancing information that it delivers
6b] The research market (in theory; in practice)
In theory, the research powers conferred by AI could open the market up to supply by a far greater number of specialist providers.
In theory, the market could start to reward the nimble, the creative, the client-focused and the innovative.
In theory, it could enable one analyst with a deep understanding of an industry supported by an AI-specialist (or their own understanding of AI) to compete with research teams many times larger.
In practice, however, investment research (including sustainable investment research) has over many years shown itself to be a remarkably closed market with multiple attempts to open it to wider competition failing.
Ultimately, I take the rather pessimistic view that expertise, counter-consensus creativity and insight in investment research come rather a long way down the list of features that asset managers seek from their research suppliers – certainly when set against factors such as systems inertia, regulatory reporting, responsibility outsourcing, sales intensity, balance sheet strength of suppliers, etc)
While it would be nice to see AI as a significant disruptor, I am not optimistic.
Alongside this, we have to place the rising tide of AI-generated information (both genuine insight and slop).
Two factors may prove significant:
- In a world of endless content, the ability to sell research messages (which depends as much on the quality of the sales effort as of the quality of the underlying research) will be critical
- In a world where the price of data tends to zero, there will be greater competition in (oversupply at) the ‘judgement’ end of the spectrum
7] (How) will AI change DEMAND for sustainable investment?
7a] Not materially — not more than other forces.
The trajectory of sustainable investment demand is already being shaped by forces considerably more powerful than AI: political & cultural preferences across the USA, the EU and in other parts of the world, consumer and beneficiary preferences and their investment firepower and regulatory flux.
While AI may inform and influence these at the margin, it seems unlikely that the technology’s influence will overpower some of these social and investment factors.
Then again, this rather depends on the extent to which sustainable investors can engage AI to enhance investment returns and deliver impacts. If we do that, demand for sustainable investment will grow. If we fail to do that, it will not.
8] (How) will AI change the supply of sustainable investment funds & strategies?
8a] A greater variety of thematic investment products?
Through its ability to gather and process granular data, AI could enable faster development of thematic investment products. The speed at which it delivers information could enable us to address one of the significant flaws in thematic investing: By the time the product is developed, much of the performance juice has already been squeezed from the investment lemon).
That said, I find it hard to envisage environmental or social themes that have not yet been identified and productized. For this application, AI feels a bit like 'a solution in search of a problem'.
In a similar vein, these data processing capabilities will make single issue screened funds (e.g. biodiversity funds, DEI funds etc) easier to launch and manage. However, I suspect that the inability to articulate a clear line of sight between the sustainability criteria applied and the investment logic is what will always constrain these funds to niches rather than an absence of information.
By contrast, as I'll discuss below, AI's ability to deep dive rapidly into issues has considerable potential when it comes to valuing sustainability issues and integrating them into broader valuation.
8b] AI could enable deeper integration of sustainability factors into valuation
This is one of the areas where AI has genuine transformative potential.
As discussed above, transformation does not depend on information or technological tools alone. It also requires:
- Demand for change
- Awareness of the alternative
- Tools that enable the transformation
- Awareness of other ‘worse case’ applications
- Barriers that are surmountable
In respect of ‘integration’, I believe these various conditions are largely met.
Current status
Current practices for integrating sustainability into valuation vary widely:
- From fundamental approaches where sustainability trends are fully evaluated for the influence they have on the key value drivers of stocks
- Through quantitative approaches which are used academically to demonstrate generalities around sustainability factors and investment performance
- To superficial approaches where scores are assigned, boxes ticked and capital sometimes allocated without much specific linkage between the sustainability issues and valuation being established
Demand for change
There is currently strong demand for better integration and demonstration of integration as part of the wider (and ongoing) pressure on sustainable investors to demonstrate the contribution they make to ‘mainstream’ investment processes.
Awareness of the alternative
Work remains to be done to broaden understanding of how sustainability factors can be integrated from bottom up into valuation. Although the practice has been developed and is deployed by some investors, many still assume that systematic quants-based approaches (as opposed to intrinsic valuation-based approaches) are the way forward.
Tools that enable the transformation
AI now provides sustainable investment analysts with a tool that enables them to quickly and cheaply explore the operating context of businesses and sustainability trends and to cross reference these two with each other and with valuation models. As such it empowers them to explore sustainability-driven investment ideas at much greater speed and with much greater depth than was hitherto possible.
Worse case applications (to be avoided)
It would be disappointing (certainly to me as I think it they are dead-ends) if AI were used (as it technically could be) to pursue systematic data-driven quants approaches.
Barriers (that are surmountable)
The primary barrier, I believe, to widespread adoption of integration is awareness and application of the techniques required. The techniques work and need to be demonstrated and applied. Once this happens, we should see wider adoption as demonstration causes better supply of information which causes further demonstration opportunities etc.
AI could be the catalyst that enables us to turn case studies into common practice.
8c] AI could enable more contextualised and more efficient engagement
Engagement is a second area where, I believe, AI has transformative potential.
I have long argued that investors could achieve twice as much engagement impact with half the resources if they aligned their activity more closely with investment drivers and encouraged companies to lead the communications process in the same way as occurs in ‘mainstream’ practice.
I now feel confident that AI could double this ratio again. Four times the impact with half the resources.
Although, it sounds punchy, I think it is achievable and we won’t know until we try.
To apply the same conditions to the practice of engagement:
Current status
Currently engagement practises vary widely across the spectrum:
- From deeply-considered identification of risks and opportunities faced by companies and the management (or not) of these
- Through financially-immaterial but sustainability-relevant practices that may be very important to the ultimate beneficiary investors expressing their moral preferences
- To, in the worst cases, engagement initiatives that are developed and prosecuted with little apparent relevance to the priorities of the company concerned or the market environments within which it operates.
Demand for change
Engagement strategies are currently under cost pressures, under political pressures and under pressure from the companies targeted. These all demand a better display of the relevance of the issues raised via engagement to the strategic and business priorities of the companies. Demand for better alignment is certainly in place.
Awareness of the alternative
In this case, the alternative is simply a filtering to ensure that more high-quality (strategically- and financially-relevant) and less low-quality (tangential or unrelated) engagement activities are prosecuted.
Tools that enable the transformation
By providing investors with a better contextual understanding of the real world context within which companies operate and the sustainability impacts of this, AI will enable them to align their engagement expectations more closely to companies’ positioning and strategic priorities.
At the same time, its ability to identify and profile roles and responsibilities within companies will make the process of engagement more efficient.
Similarly, AI's ability to inform companies in advance about investors’ priorities, needs and expectations will focus communications effectively.
Beyond these, AI-supported enhancements to communications practice and reporting will improve efficiency and deliver better ‘lines of sight’ between beneficiary investor client expectation, engagement undertaken by asset managers and sustainability outcomes achieved by the company.
AI will enable investors who want to engage more deeply to do so: arrive at company meetings better briefed, ask better questions, process the responses with a better understanding of its strategic and sustainability relevance etc.
Worse case applications (to be avoided)
AI will enable investors who want to run more engagement cases with less effort to do exactly that: to generate engagement rationales at scale using scraped data and to construct and prosecute intervention cases in bulk at minimal cost for the sake of ticking reporting boxes.
Barriers (that are surmountable)
As above, AI could improve the focus, quality, relevance and depth of investor to company engagement on sustainability issues or it could be used to supercharge superficial engagement and greenwashed reporting.
Which of these pathways is followed depends on the extent to which asset owners and beneficiary investors interrogate the engagement activity of asset managers. That is a human decision, not a technological one.
Overall
Individually, the changes identified above are incremental. However, taken together the improvement to engagement focus, practice and processes that they deliver could be little short of revolutionary. I stick with my claim that an eight-fold improvement is possible.
8d] Will AI change the balance between active and passive investment?
Although I have given this some thought, I’m not sure I can see much in AI that disrupts or materially changes this dynamic.
The long march to passive is on; no-one seems inclined to enforce a levy on passive investment to pay for the price finding ‘service’ that active investors perform and that passive investors depend on.
However, a few factors do seem to merit further consideration:
- The contortions (to their rules) that some index providers seem to be performing to give passive investors access to Limited float IPOs perhaps the first impact of AI is to make passive investors become more active?
- If AI can bring the cost producing alpha through fundamental analysis down significantly then active strategies become more viable at lower AUM thresholds, and the case for passive strategies narrows.
- If / when AI can see through the value chain to match the actual liabilities of investors with the returns available from companies then new low-cost investment strategies that do not rely so heavily on the relationship between equity markets and inflation may emerge.
9] How will AI change investor reporting (to clients) on sustainability?
9a] A spectrum of possibility
My hopes are that AI could be deployed to improve investor reporting on sustainability by delivering:
- More accessible reporting (whether by content choice or format)
- More clarity
- More personalisation (to the needs of the individual reader)
- More transparency
- More transparency and engage-ability with ‘real-activity’ focus
My fears are that it could be used to deliver:
- Unreadably encyclopaedic reports with more data and narrative than anyone has the bandwidth to absorb
- Volumes of bland greenwashing
- Data-loaded reports that hide a lack of fundamental ‘real-world’ activity
It might be worth lighting a beacon on a hill whereby a beneficiary investor is able – via an AI-chatbot – to interrogate all sustainability-related actions of an asset manager – not because I think that any client will ever be interested in that level of detail but because it is a laudable direction to head in.
10] How will AI affect company – investor communications on sustainability?
Beyond the recommendations that apply to all sustainable investment value chain participants, AI can be applied in a number of specific ways by IROs and CSR managers at companies.
10a] Better mutual understanding => better interaction => better investment decision-making and more significant sustainability outcomes
===
Dear IRO/CSO,
Imagine that every interaction you ever have with any investor or research provider is informed in advance with a full understanding (for you) of that investor’s / research provider’s real needs, priorities, capabilities and the potential follow-up actions available to them.
(Importantly, this won’t be delivered as a 40-page report for you to ignore. It will be easily accessible in whatever format you choose to access it)
===
Dear sustainable investment analyst,
Imagine that every company that you ever interact with has, in advance, a good level of understanding about your needs and focus.
Also imagine that you are prepared for each interaction with a full understanding of the company’s business context, value chain position, strategic focus, sustainability narrative (and the gaps in it) and the market’s current views on these.
Dear both,
Now imagine that these two perspectives are brought together ahead of the interaction such that matches and disconnects between investor priorities and company positions are identified in advance.
It goes without saying that any interaction prepared for in this way will be unrecognisably superior to what we experience today.
Encouragingly, all of the technology needed to deliver this is available to everyone today at modest cost. Also, the wiring / processing steps required to deliver it universally are minimal.
10b] Sustainability reporting … has already changed
It is noticeable that - this year, particularly - companies have stopped producing sustainability / CSR reports in favour of integrating sustainability information within their annual reports.
It would be nice to gloss this as belated recognition that sustainability factors should be considered as integral to mainstream business operations and, hence, to reporting. I fear, however, that it is actually belated recognition by companies that very few people were reading these reports and that the political climate is currently such that they can get away with it.
As one of the few people (I suspect, in the world) that still considers standalone sustainability / CSR reports to be of great value as measures of companies’ preparedness for sustainability trends, I mourn their passing. (In ‘analyst-mode’, I regard the way that a company wants to present itself in respect of sustainability as a useful lead indicator for its underlying ability to manage exposures.).
However, I have resigned myself (as I believe investors and research providers also must) to creating our own windows into company practice now that we are deprived of the window created by the company itself.
There are self-evident ‘pros’ and ‘cons’ to both companies and investors of this new approach. What is clear, however, is that companies need to be planning to make their reporting much more machine accessible than is currently the case to ensure that the information that they want to present fits the multiplicity of different windows onto their activities what will be considered.
10c] Action for IROs / CSOs: Get to know your investors and their real needs and priorities
With a small amount of considered prompting, AI tools can deliver to companies clear and accessible summaries of which sustainable investors matter to them and what the real needs and priorities of these investors are.
Ultimately, only about forty individuals matter to an individual company's sustainable investment positioning. So, it's nice to have a tool that enables us to find out who these people are and what they want.
10d] Action for IROs / CSOs: Make your reports machine-readable
Companies often don't publish clearly the date on which their sustainability / CSR (or even annual) reports are published. It’s extremely annoying!
(Note: The same companies then complain that ESG rating agencies don't access the most up-to-date information on them. This is also annoying!)
What this illustrates (at the most basic level) is a need for companies to make information more accessible and readable by machines.
Although machines are better at finding reporting dates than humans, there is no harm in helping them and making all information within such reports fully machine-readable.
Anecdotally, the AI tools that I use don't like massive PDFs (like Annual Reports). Rather than dig out sustainability information when asked, they often give up and start making stuff up.
This is all avoidable if companies simply ensure that the information that machines need to track their sustainability practices are included in reports that are machine-readable.
11] How will AI affect the regulation of sustainable investment?
11a] Regulators will, most likely, miss the point again
I have never been a fan of the way that regulators have tackled sustainable investment.
I have frequently argued that they have swallowed unquestioningly the false narrative around quantitative data and have misunderstood, at a fairly fundamental level, the way that sustainable investment can deliver sustainability outcomes.
As a result, the regulation that has been developed and applied has been hugely inefficient and often counter-productive.
Optimistically, AI presents a tremendous opportunity to regulators – if they can drag themselves away from their rules- and disclosure-driven instincts and focus instead on market-enabling activity.
They could use AI technologies to develop tools that facilitate transparent markets that inform and protect users and consumers alike. They could create ‘case-to-be-answered’ agents. They could create a sustainable investment regulatory super-agent.
They probably won’t. Sigh!
12] Will AI bring about the end of sustainable investment?
12a] AI may eventually replace (active) investment altogether
When AI is able to process and price effectively, in real time, all information that is relevant to industries and businesses and to predict accurately the market's reaction to new information, human agency in capital allocation will no longer be needed.
While this development process has started, it is hard to see whether this day arrives in five, ten or twenty years’ time.
I have argued elsewhere that, in the context of AI, we should all think existentially about our roles and the roles of our firms in the capital allocation process. (We should also think incrementally about what we can do tomorrow.)
In respect of the former, it is helpful to be alive to the possibility that AI might eventually replace the need for active investment. However, ‘eventually’ is a key word here and an awareness of the potential ‘endgame’ must not paralyse us in any respect (or else it creates a self-fulfilling prophecy). Considerable value will be created and will be destroyed between now and this time as investors and research providers position themselves on these shifting sands.
(It’s not dissimilar to the notion that ‘eventually’ renewable energy will replace fossil-derived energy).
In human terms, much of our judgement, creativity and patience will be needed to train AI and its models before we can sit on beaches and enjoy their work.
The amount of value that will be transferred…
- between different asset managers,
- between different parts of the investment value chain
- between investment value chain incumbents and new entrants
… is what makes immediate and active engagement an absolute necessity for anyone who wants to play any role in managing money in five years’ time.
12b] Sustainable investment may be most resilient
While sustainable investment uses and is subject to all of the same information processing trends as ‘mainstream investment’, it also operates in an additional dimension: it incorporates human preferences, values, hopes, fears, and choices that go beyond the purely financial.
A world in which all financial information is processed instantly and efficiently will still be a world in which people agree and disagree about social priorities and what they want their capital to do. This agreement and disagreement are the irreducible human elements that active sustainable investment serves. Again, sustainable investment’s ‘moat against the machines’ may be wider and deeper than that of ‘mainstream’ investment.
13] Do the risks of AI outweigh the opportunities?
13a] Massive risks and opportunities require open-minded evaluation
The environmental, social and ethical risks posed by AI are massive.
I do not need to rehearse here the huge environmental pressures imposed by data centres, the challenges to workforces of the socio-economic revolution presented by the technology impact or the ethical challenges presented by autonomous artificial general intelligence. These are well covered by others. However, for companies, I do need to recognise them.
However, it is worth noting that – in the face of new technology - ESG practitioners have often (historically) shown a bias towards identifying and managing downside risks and away from identifying and capturing upside opportunities.
This should be avoided with AI. We need to be open-minded about the balance of downside risks and upside opportunities presented by these technologies and to adapt appropriately.
13b] Sustainable investment criteria will need to adapt to the novel challenges of AI
In the segment below, I reflect on the wide variety of ways that sustainable investment criteria will need to develop to ensure that AI related risks and opportunities are fully understood by sustainable investors and incorporated into valuation.
13c] The risks should not prevent us from adopting AI for our practices and processes
We need to consider separately:
- the downside risks that AI brings to businesses in which we invest and to wider society
- the upside opportunities that AI brings to businesses and to wider society
- the downside risks that AI brings to the operating practices and processes of our own investment businesses
- the upside opportunities that AI presents to the operating practices and processes of our businesses
Importantly, we must allow these to coexist in our minds and not allow one category to cloud our judgment about the other.
14] How will AI change the issues addressed by sustainable investment?
As noted above, the emergence of AI presents a swathe of complex environmental, social, economic, ethical and financial issues that companies and investors must face – particularly as we seek to integrate these factors into valuation and capital allocation.
To comprehend and engage with these, we will need to expand somewhat our conceptual 'ESG' frame of reference in four (or possibly more?) ways:
14a] Expand ESG: Add an economic ‘E’
AI presents society with both economic opportunities and social risks. On the one hand, the technology could deliver productivity with significant and widespread economic impact. It could also deliver social upheaval of scale and significance. Alternatively, social protest could prevent its delivery altogether.
In order to understand the economic dimension effectively, sustainable investors first need to separate their understanding of companies’ ‘economic contribution’ from their understanding of companies’ ‘financial return’. (I have written elsewhere about the difference and how a failure to differentiate has repeatedly disempowered sustainable investors.)
At a macro level, the economic case for AI rests on its ability to improve economic productivity. Whether or not this delivers positive social outcomes in the short or long term remains to be seen. Either way it is a factor that we need to form a view on independently of the impact of the technology on individual companies - not least because it appears to have become an explicit consideration for the Fed.
In ‘ESG’ terms, we need another ‘E’.
14b] Expand ESG: A better ‘S’
It is widely held that the social factors that contribute to ‘ESG’ are harder to apply than environmental factors because they are less easily quantified.
I find this to be a misdiagnosis of the situation. I think that 'social' factors are less-explicitly applied to sustainable investment because
- more of them have already been explicitly articulated within legislation and regulation such that they are already priced
- no-one has effectively articulated a unifying set of parameters for companies (my preferred approach applies stakeholder evaluation across multiple timeframes - but I am not aware that this is commonly applied
Put another way, no-one has ever really articulated holistically what ‘good’ looks like for ‘social’ impact and how this could be applied to sustainable investment practice.
Better late than never … and, hopefully, in time for its application to AI, let me suggest that ‘level of contribution to social inclusion, cohesion and equity’ would be a useful top-line measure and one that could be applied to every business from defence to healthcare and from utilities to (AI) software development.
14c] Expand ESG: Add an ethical ‘E’
We can’t ignore - but I can’t pretend to understand - the ‘ethical’ issues associated with autonomous artificial general intelligence.
Then again, I don’t need to understand this as the somewhat-more-qualified Pope Leo is already on the case and – in a more immediate investment context - Laurie Fitzjohn-Sykes’ Investor AI Resource Hub is developing resources to help investors to understand and apply these factors into their evaluation of and engagement with companies.
14d] Expand ESG: Separate the ‘CG’
Although I’m no corporate governance expert, I can see that Silicon Valley’s approach to shareholder rights is (shall we say?) ‘somewhat creative’. This will need to be factored into valuation by investors for the protection of their own assets and because it removes some of the checks and balances that lead to the effective allocation of capital and the wider economic order.
IMHO, corporate governance (and specifically the rights and responsibilities that comprise the architecture of the relationship between investors and companies) deserve specific focus and not to be bundled in with other E&S factors.
This is particularly true as these factors need to be applied differently to the valuation of assets. If we're going to teach machines to price and value the different aspects of sustainability within their models, we need first to understand ourselves that:
- environmental, social, economic and ethical factors can be applied to the P&L and balance sheet in forecasts and valuation models while
- corporate governance factors are best valued by adjusting the risk premium and, by extension, the WACC.
The forthcoming IPOs of AI companies emphasises the importance of this – not least because who owns (and has ultimate decision-making power over) technologies of such societal significance matters to us all!
15] What are AI’s implications for SRI-Connect?
www.sri-connect.com is an online research network for the global sustainable investment industry. It has >5,000 professional members and is free (to join and use) for anyone with current professional exposure to the sustainable investment value chain.
15a] A glass half … both empty and full
At first sight, SRI-Connect is an information business serving other information businesses (investment management and research) – facing a technology that commoditises information and reduces its cost to zero. So, it is not looking promising!
However, SRI-Connect focuses specifically on:
- People in sustainable investment and their profile, ideas, needs and priorities
- The distribution of fundamental research and analytical insight.
(We don’t deal in data or ratings – only the contextual information and differentiated ideas that lead to capital allocation) - Direct (person-to-person) communications between companies and investors on sustainability issues
These points of focus, we believe, makes SRI-Connect…
- resilient to
- able to benefit from and
- well-positioned to help the sustainable investment value chain navigate …
… the next AI-driven chapter of our industry’s evolution. So, bring it on!
15b] Introducing Codus
I have recently – with Neil Brown – launched Codus – a partnership to help investors, companies and research providers position their firms, business models and workflows for the arrival and evolution of AI.
Codus will operate alongside SRI-Connect such that I am able to continue delivering focused support to individual firms and also contribute to wider market development.
Codus operate through five streams (Strategy, Execution, Training, Nexus and Tooling). Of these three (Training, Nexus and Tooling) are focused on wider market development.
Outstanding questions
- What is the current state and trajectory of investor adoption of AI?
- What is the current state and trajectory of research provider adoption of AI?
- What is the current state and trajectory of company (IRO & CSO) adoption of AI?
- How can we use AI capabilities to re-engage the experience of sustainable investment professionals who have recently left paid employment?
- How might AI be used to improve the quality of sustainable investment regulation? What could a ‘case-to-be-answered’ agent look like?
- … and so, so, many more!
