Dec 26, 2025
Stop Asking 'What's Our AI Strategy?' Start Building Your Intelligence Strategy
The Wrong Question in the Right Room
Walk into any financial services boardroom and you'll hear some variation of the same question: "What's our AI strategy?" It's a question that sounds sophisticated. It gets nodded at solemnly. Consultants get hired. Budgets get allocated. PowerPoint decks bloom like spring flowers.
And yet, it is precisely as useful as asking "What's our electricity strategy?"
Technically, the question is relevant. Conceptually, it's entirely useless. Worse still, it's dangerous because it sends organisations off on a wild goose chase, chasing generic AI capability when what they actually need is domain-specific intelligence.
Think about it. Your firm doesn't have an electricity strategy. You have an electricity infrastructure strategy. You plug things in, the lights come on, people get work done. The electricity itself is a commodity. What matters is what you do with it.
AI is the same. The question isn't whether you use AI. You already do. Neither is it how much AI you should use.
There is one question that truly matters, and it separates firms that benefit from AI from those that waste millions on it:
What problems are we solving within our domain, using intelligence that understands our domain?
The Reality: What CEOs Are Actually Experiencing
Here's where the gap between aspiration and reality becomes rather uncomfortable. According to Teneo's survey of 350+ public-company CEOs, less than half of current AI projects had generated more in returns than they cost. Yet 68% plan to increase AI spending in 2026.
This isn't irrational, quite the opposite. It's the sound of organisations finally understanding that the problem wasn't the technology. It was the absence of domain-specific logic. After spending millions on generic AI and watching accuracy plummet, observing outputs that hallucinate with bewildering confidence, and seeing recommendations that ignore regulatory constraints, leadership has reached a rather obvious conclusion: the technology isn't the problem. The strategy is.
Defining the Distinction
Before we go any further, it's worth clarifying two terms that get conflated:
Domain-specific intelligence is intelligence built to understand the rules, structures, and logic of a particular industry. In Capital Markets, that means understanding market structure, financial instruments, research methodologies, regulatory constraints, and how decisions are actually made. It knows what matters in the domain and, as importantly, what doesn't.
Organisation-specific intelligence goes a step further. It reflects how your firm operates within that domain. Your investment frameworks. Your internal research. Your methodologies, taxonomies, constraints and risk models.
Generic AI understands neither. Domain-specific intelligence understands the industry. Organisation-specific intelligence understands your business.
The firms seeing real value from AI are building both and in that order.
Domain-Specific Intelligence: The Unglamorous Advantage
Domain-specific intelligence isn't sexy. It doesn't get mentioned at Davos. It doesn't sell consulting hours. It's not "AI" or at least, not the kind of AI that gets people excited.
But it works.
The Problem With Generic AI in Capital Markets
Here's what happens when a firm deploys generic AI on Capital Markets data: they get a sophisticated system that is, in the most fundamental way, illiterate about their business. A large language model trained on the open internet knows what "earnings" means in a general sense. It doesn't know what your firm's earnings models should prioritise, what regulatory constraints shape your assumptions, or why a 1% discrepancy in a forward estimate matters.
The result: between 60–80% accuracy on critical financial decisions. Not 90%+. Not "highly reliable." In trading, in wealth management, in asset allocation, that's not "pretty good," it’s "dangerously misleading."
Generic AI systems hallucinate. They produce outputs that are so plausible, so confidently stated, that they look like the product of analytical reasoning when they're actually statistical guesses wrapped in persuasive language. They lack grounded knowledge about Capital Markets terminology, market structure, and the workflows that actually matter. They can't explain their reasoning in terms that compliance officers accept. They can't audit their own outputs. They certainly can't justify them when things go wrong.
And when things go wrong in Capital Markets, they go very wrong.
The Three Foundations of Domain-Specific Intelligence
Domain-specific intelligence is built on three foundations:
First: Context-Aware Data Integration
Your firm has internal data scattered across systems. Your firm ingests research, market data, regulatory filings, sentiment data. Some of it's structured. Much of it isn't. A domain-specific intelligence system takes all of that and creates a unified, contextualised intelligence graph. The system understands how data points relate to one another, not just where they're stored. A report from your research team isn't just another text file. It's a validated institutional view. A regulatory filing isn't just a document. It's a signal. A market datapoint isn't just a number. It's context for decision-making.
Second: Domain-Specific Analytics and Reasoning
Once structured and unstructured data is integrated, domain-specific intelligence applies quantitative models, machine learning, and AI reasoning that respects the rules of your industry. Your taxonomies? Embedded. Your quantitative models? Embedded. Your methodologies? Embedded. The system doesn't guess, it reasons within your domain's logic.
Third: Persona-Specific Delivery
A research analyst needs different insights, at different levels of detail, with different explanations, than a wealth adviser or a trader does. Domain-specific intelligence delivers the same underlying intelligence through different workflows, suited to how each persona actually works.
The Wrong Question in the Right Room
Walk into any financial services boardroom and you'll hear some variation of the same question: "What's our AI strategy?" It's a question that sounds sophisticated. It gets nodded at solemnly. Consultants get hired. Budgets get allocated. PowerPoint decks bloom like spring flowers.
And yet, it is precisely as useful as asking "What's our electricity strategy?"
Technically, the question is relevant. Conceptually, it's entirely useless. Worse still, it's dangerous because it sends organisations off on a wild goose chase, chasing generic AI capability when what they actually need is domain-specific intelligence.
Think about it. Your firm doesn't have an electricity strategy. You have an electricity infrastructure strategy. You plug things in, the lights come on, people get work done. The electricity itself is a commodity. What matters is what you do with it.
AI is the same. The question isn't whether you use AI. You already do. Neither is it how much AI you should use.
There is one question that truly matters, and it separates firms that benefit from AI from those that waste millions on it:
What problems are we solving within our domain, using intelligence that understands our domain?
The Reality: What CEOs Are Actually Experiencing
Here's where the gap between aspiration and reality becomes rather uncomfortable. According to Teneo's survey of 350+ public-company CEOs, less than half of current AI projects had generated more in returns than they cost. Yet 68% plan to increase AI spending in 2026.
This isn't irrational; quite the opposite. It's the sound of organisations finally understanding that the problem wasn't the technology. It was the absence of domain-specific logic. After spending millions on generic AI and watching accuracy plummet, observing outputs that hallucinate with bewildering confidence, and seeing recommendations that ignore regulatory constraints, leadership has reached a rather obvious conclusion: the technology isn't the problem. The strategy is.
Defining the Distinction
Before we go any further, it's worth clarifying two terms that get conflated:
Domain-specific intelligence is intelligence built to understand the rules, structures, and logic of a particular industry. In Capital Markets, that means understanding market structure, financial instruments, research methodologies, regulatory constraints, and how decisions are actually made. It knows what matters in the domain and, as importantly, what doesn't.
Organisation-specific intelligence goes a step further. It reflects how your firm operates within that domain. Your investment frameworks. Your internal research. Your methodologies, taxonomies, constraints and your risk models.
Generic AI understands neither. Domain-specific intelligence understands the industry. Organisation-specific intelligence understands your business.
The firms seeing real value from AI are building both and in that order.
Domain-Specific Intelligence: The Unglamorous Advantage
Domain-specific intelligence isn't sexy. It doesn't get mentioned at Davos. It doesn't sell consulting hours. It's not "AI" or at least, not the kind of AI that gets people excited.
But it works.
The Problem With Generic AI in Capital Markets
Here's what happens when when a firm deploys generic AI on Capital Markets data.: they get a sophisticated system that is, in the most fundamental way, illiterate about their business. A large language model trained on the open internet knows what "earnings" means in a general sense. It doesn't know what your firm's earnings models should prioritise, what regulatory constraints shape your assumptions, or why a 1% discrepancy in a forward estimate matters.
The result: between 60–80% accuracy on critical financial decisions. Not 90%+. Not "highly reliable." In trading, in wealth management, in asset allocation, that's not "pretty good," it’s "dangerously misleading."
Generic AI systems hallucinate. They produce outputs that are so plausible, so confidently stated, that they look like the product of analytical reasoning when they're actually statistical guesses wrapped in persuasive language. They lack grounded knowledge about Capital Markets terminology, market structure, and the workflows that actually matter. They can't explain their reasoning in terms that compliance officers accept. They can't audit their own outputs. They certainly can't justify them when things go wrong.
And when things go wrong in Capital Markets, they go very wrong.
The Three Foundations of Domain-Specific Intelligence
Domain-specific intelligence is built on three foundations:
First: Context-Aware Data Integration
Your firm has internal data scattered across systems. Your firm ingests research, market data, regulatory filings, sentiment data. Some of it's structured. Much of it isn't. A domain-specific intelligence system takes all of that and creates a unified, contextualised intelligence graph. The system understands how data points relate to one another, not just where they're stored. A report from your research team isn't just another text file. It's a validated institutional view. A regulatory filing isn't just a document. It's a signal. A market datapoint isn't just a number. It's context for decision-making.
Second: Domain-Specific Analytics and Reasoning
Once structured and unstructured data is integrated, domain-specific intelligence applies quantitative models, machine learning, and AI reasoning that respects the rules of your industry. Your taxonomies? Embedded. Your quantitative models? Embedded. Your methodologies? Embedded. The system doesn't guess, it reasons within your domain's logic.
Third: Persona-Specific Delivery
A research analyst needs different insights, at different levels of detail, with different explanations, than a wealth adviser or a trader does. Domain-specific intelligence delivers the same underlying intelligence through different workflows, suited to how each persona actually works.
The result: 90%+ accuracy, transparency that regulators accept, insights that are transparent and drive trust and adoption to make better decisions, and outputs that can actually be acted upon.
Why Most Firms Haven't Yet Cracked the Code
Most organisations embarking on their AI journey follow a predictable pathway, and whilst it's not inherently flawed, it often lacks the strategic foundation needed for success:
Recognition. Leadership recognises that AI capabilities are becoming table stakes in their industry.
Talent Acquisition. The firm invests in AI expertise, often bringing in specialists or creating new roles to drive the initiative forward.
Ambitious Vision-Setting. The organisation articulates bold intentions to embed AI meaningfully across operations.
Infrastructure Deployment. A generic AI platform or large language model is integrated with existing data systems, creating a starting point for experimentation.
Initial Optimism. Early pilots and proof of concepts demonstrate promise, generating internal momentum and buy-in.
Reality Checkpoint. More rigorous testing reveals gaps: outputs lack domain context, accuracy falls short of critical thresholds, or recommendations don't align with operational constraints.
Recalibration. Teams reassess their approach, recognising that the limitation isn't the technology itself, but rather the absence of domain-specific logic and governance.
Strategic Reset. Armed with new understanding, organisations pivot towards building the intelligence layer that actually serves their business.
This journey isn't failure, it's learning. Many of the most sophisticated firms in capital markets are currently at stage 6 or 7, converting initial generic AI deployments into domain-specific capabilities.
What separates firms that successfully navigate this arc from those that stall is strategic clarity: understanding what problems you're solving, why domain-specific expertise matters, and how to build intelligence that compounds over time.
The Better Question
So here's the question that really matters:
"What intelligence do our decision-makers need, within our domain, to move faster and with more confidence?"
That question forces you to think like a business. It forces specificity. It forces trade-offs. It forces strategy.
Once you've answered that question, then you can ask: "What data do we need? What analytics? What workflows? What guardrails?" And suddenly, your "AI strategy" becomes something real—a blueprint for solving specific problems with intelligence that understands your business.
Generic AI won't do that. It can't. It wasn't built for your domain.
But domain-specific intelligence can. And it does.
The firms that will win in the next five years won't be the ones with the best "AI strategy." They'll be the ones with the best domain-specific intelligence strategy and the discipline to focus on outcomes, not generic AI.
Your move.
Why Most Firms Haven't Yet Cracked the Code
Most organisations embarking on their AI journey follow a predictable pathway, and whilst it's not inherently flawed, it often lacks the strategic foundation needed for success:
Recognition: Leadership recognises that AI capabilities are becoming table stakes in their industry.
Talent Acquisition: The firm invests in AI expertise, often bringing in specialists or creating new roles to drive the initiative forward.
Ambitious Vision-Setting: The organisation articulates bold intentions to embed AI meaningfully across operations.
Infrastructure Deployment: A generic AI platform or large language model is integrated with existing data systems, creating a starting point for experimentation.
Initial Optimism: Early pilots and proof-of-concepts demonstrate promise, generating internal momentum and buy-in.
Reality Checkpoint: More rigorous testing reveals gaps: outputs lack domain context, accuracy falls short of critical thresholds, or recommendations don't align with operational constraints.
Recalibration: Teams reassess their approach, recognising that the limitation isn't the technology itself, but rather the absence of domain-specific logic and governance.
Strategic Reset: Armed with new understanding, organisations pivot towards building the intelligence layer that actually serves their business.
This journey isn't failure, it's learning. Many of the most sophisticated firms in Capital Markets are currently at stage 6 or 7, converting initial generic AI deployments into domain-specific capabilities.
What separates firms that successfully navigate this arc from those that stall is strategic clarity: understanding what problems you're solving, why domain-specific expertise matters, and how to build intelligence that compounds over time.
The Better Question
So here's the question that really matters:
"What intelligence do our decision-makers need, within our domain, to move faster and with more confidence?"
That question forces you to think like a business. It forces specificity. It forces trade-offs. It forces strategy.
Once you've answered that question, then you can ask: "What data do we need? What analytics? What workflows? What guardrails?" And suddenly, your "AI strategy" becomes something real—a blueprint for solving specific problems with intelligence that understands your business.
Generic AI won't do that. It can't. It wasn't built for your domain.
But domain-specific intelligence can. And it does.
The firms that will win in the next five years won't be the ones with the best "AI strategy." They'll be the ones with the best domain-specific intelligence strategy and the discipline to focus on outcomes, not generic AI.
Your move.
