When founders talk about AI ROI, the conversation almost always lands on efficiency. How much time did we save? What tasks are we doing faster? Can we avoid hiring that next person? These are real gains — the data backs them up — but they describe a category of benefit that turns out to be the smaller of the two available categories.
PwC's 2026 AI Performance Study, which surveyed 1,217 senior executives across 25 sectors, produced a result worth sitting with. Three-quarters of AI's economic value — 74% — is flowing to just 20% of companies. The remaining 80% of businesses are sharing the leftover 26%. That is not a marginal gap. And when the researchers went looking for what explains it, the answer was not efficiency. Leaders generate 7.2 times more revenue and efficiency gains than the average competitor. But the primary driver of that performance is not doing existing work faster.
It is doing work they could not have done before — in markets they could not previously afford to enter.
PwC calls this "industry convergence." It is the single strongest factor in AI-driven financial performance, ranking ahead of operational efficiency gains in predictive power. AI leaders are two to three times more likely than their peers to use AI to identify and pursue growth opportunities arising from industry convergence — expanding beyond their traditional sector boundaries to offer services or reach customers that previously sat outside their addressable market. Leaders using AI this way are 2.5 times more likely to post revenue growth above 10% and 3.6 times more likely to run at margins above 15%.
Most founders are optimising the machine they already have. The top performers are building a different machine entirely.
What the Research Is Actually Describing
Industry convergence sounds like something that happens to large enterprises — a bank entering insurance, a retailer entering healthcare. But the mechanism that AI enables applies at every scale. It is not about sector-level M&A. It is about the cost structure of capability.
Before AI, adding a new service line meant adding expertise. A marketing agency that wanted to offer competitive intelligence had to hire an analyst — someone who could spend days researching competitor positioning, synthesising industry reports, and producing actionable findings. That is a $70,000 salary minimum, and the workload to justify it would need to be recurring and substantial. Most founder-led agencies could not clear that bar, so they stayed in their lane and referred that work out or let it go.
After AI, the constraint changes. A research workflow running on Perplexity, Claude, and a structured prompt library can produce competitive landscape analysis in hours that would have taken a junior analyst weeks. The expertise required to commission and quality-check it is still necessary — but the cost of the underlying capability dropped by roughly 95%. The barrier to entering the adjacent service line dropped with it.
This is the mechanism. Not sector-level transformation. Not sweeping disruption. The unit economics of entering an adjacent market shifted dramatically, and the companies seeing the largest AI returns are the ones that noticed and acted on it.
Deloitte's 2026 State of AI report adds a frame to why most companies are not doing this: only 34% of organisations are "truly reimagining their business" with AI. The other 66% are in efficiency mode — doing existing things faster and cheaper. That is not nothing. But it is also the floor, not the ceiling.
What This Looks Like at Founder Scale
The convergence examples that make most sense for founder-led businesses are not glamorous. They are practical expansions of what you already do well, made possible because AI removed the capability bottleneck that was blocking you.
The accounting firm that now offers forecasting. Basic bookkeeping and tax preparation are commoditised services with low margins and significant price pressure from tools like QuickBooks and Bench. Financial modelling, scenario analysis, and cash flow forecasting — the work a fractional CFO does — command two to three times the effective hourly rate. The expertise gap between bookkeeper and fractional CFO used to require years of additional experience and separate credentials. The execution gap — the ability to actually produce a well-structured three-year projection with scenario toggles — can now be dramatically compressed using AI tools like Runway, or well-prompted models with the right financial context. Firms entering this adjacent service are not competing with full CFOs. They are capturing a market of founder-led businesses that need the output but cannot justify the full-time hire. That market is large, underserved, and willing to pay.
The HR consultant who now delivers training programs. HR consulting typically covers policy, process, and compliance — advice-oriented work that clients value but can only engage on intermittently. Training content development — building onboarding programs, skills curricula, and leadership development tracks — is a different service with different economics: recurring, deliverable-based, and priced per program rather than per hour. Building that content traditionally required instructional designers, writers, and subject matter interviews spread across weeks. AI brings the production cost down by an order of magnitude. An HR consultant who can frame the learning objectives and quality-check the output can now offer a training development service that would have required a separate team to build. The work that needed a content studio now needs a sharp prompt, a structured approach, and the same domain knowledge they already have.
The marketing agency that now offers sales enablement. Marketing agencies produce content. Sales enablement — the collateral, playbooks, battlecards, and objection-handling frameworks that sales teams use to close deals — is a distinct category that sits adjacent to marketing but is typically owned by sales operations or specialist firms. The production requirements are similar: research, writing, structured documents, iteration. The buyers are different and the budgets are often larger, because the output is directly tied to revenue. An agency with working AI content systems already has most of what it needs to enter this market. What it needs is a framing shift and a first client willing to test the expanded scope.
In each case, the expansion is not speculative. It is entry into a service category that already has buyers and already has budget — buyers who currently get that service from someone else or do without it. AI did not create the demand. It removed the cost barrier that was keeping them out.
The Four Questions That Find Your Opening
The practical question for a founder is not "what can AI theoretically enable?" It is "which specific adjacent market is now accessible to my business that wasn't before?" These four questions tend to find it.
What do your existing clients wish you also offered? This is the most reliable signal available. Clients who already trust you and would prefer a single relationship have been selecting someone else for an adjacent need because you didn't offer it. That gap is both a market signal and a warm pipeline — they already know you and have already chosen to work with you. Ask them directly what they're getting elsewhere that they'd rather get from you. The most common answers are the candidates for your convergence move.
What capability were you referring out that AI has now made accessible? Most founders-led businesses have a mental list of things they say no to or refer to partners. Some of those things are genuinely outside your zone of competence and should stay there. Others are things you said no to because the production cost was prohibitive, not because you lacked the judgment to do them well. That second category is where to look. The execution cost changed. The strategic judgment required did not.
What does your expertise enable you to quality-check that you couldn't produce yourself? This is the asymmetry AI creates. You do not need to be able to produce something from scratch to be able to build a service around it — you need to be able to commission, evaluate, and improve it. A former finance executive can quality-check a financial model they did not build. A marketing strategist can evaluate a competitive analysis they did not research. If your existing expertise makes you a credible judge of output quality in an adjacent domain, AI makes you a credible producer of that output. Those are two different things, and the gap between them used to matter. It matters less now.
What adjacent service is currently being delivered at a price point that creates an underserved market below it? Many professional services have a pricing ceiling that excludes a large population of buyers who have the need but not the budget. Fractional CFOs start at $3,000–5,000 a month — accessible to growth-stage companies, but out of reach for businesses under $1M revenue. That segment still has financial planning needs. If your AI-enabled service can deliver 70% of the value at 30% of the price, you are not competing with the full-price incumbents. You are opening a market that currently has no supplier.
The Sequencing That Actually Works
None of this means efficiency work is the wrong place to start. It is the right place to start — with a specific endpoint in mind.
The operational case for AI begins with automating your highest-volume, lowest-judgment recurring tasks: lead qualification, proposal follow-up, client update reports, scheduling coordination. These are the workflows that consume the most time and require the least of the expertise that makes your business valuable. Getting these running reliably creates two things: slack capacity and AI operational fluency. Slack capacity is what you reinvest in the expansion move. Operational fluency is what lets you build and manage the expanded service without building a separate team to run the AI.
Founders who skip the efficiency phase and jump directly to expansion tend to stall — they are trying to build new AI-enabled services at the same time they are learning to work with AI, which doubles the failure surface. Founders who complete the efficiency phase but never move to expansion are leaving the majority of the return on the table. The sequencing is efficiency first, expansion second. Not efficiency instead of expansion.
The urgency is real. Census Bureau data from May 2026 shows that only 17–20% of US businesses are currently using AI in any meaningful capacity, with another 20–23% planning to adopt in the next six months. That adoption wave will close some of the competitive gap. The businesses that moved early and are already operating in expanded markets will have a head start in client relationships, service reputation, and operational efficiency that is hard to replicate from a standing start. But the window is narrower than it was twelve months ago.
Not sure which adjacent market makes sense for your business — or whether AI has actually removed the barrier yet? That is the strategic audit we run. We map what you're capable of now versus what you could offer, identify where the real convergence opportunity is, and help you build the first version without over-investing in a hypothesis. Talk to us. We'd rather tell you no than watch you pursue the wrong expansion at the wrong time. We'd rather tell you no than waste your money.
Related: The AI Performance Gap Is Real — Here's Which Side of It You're On | Your AI Copilot Is the Slow Lane. Here's What the Fast Lane Looks Like.