Most accounting firms are using AI to solve a cost problem. PwC's 2026 AI Performance Study shows the 20% are solving something different — a revenue problem. Here's how to tell which category your practice is in, and what the 20% are doing instead.
In our firm, we're also using AI mostly on the cost side. We're doing more with less, eliminating subscriptions, speeding up the work we were already doing. I'm not going to pretend that's the wrong answer. Fee pressure is real, and if your practice can't do more with less, the math stops working. Cost savings are how you stay in the game.
We also use AI when we're sitting with clients — drafting website alternatives, mapping ideal customer profiles, generating marketing ideas that strengthen the advisory conversation. That's different work. It doesn't save time. It adds value inside engagements we're already billing for.
What we haven't fully cracked yet is the third thing. And on April 13, PwC published data that makes the third thing impossible to ignore.
Cost is where most of us live — and that's not the problem
PwC's 2026 AI Performance Study surveyed 1,217 senior executives across 25 sectors. The headline finding: eighty percent of the financial returns from AI are being captured by twenty percent of companies. Those companies aren't focused on productivity. They're focused on growth.
Here's the part that lands hardest. Ninety-six percent of AI-investing organizations report productivity gains. But the average net productivity increase is only 11.5%. Productivity gains are real — they're just not what separates the 20% from everyone else.
If you're running a CAS practice, the 80% position is normal. Fee compression, staff costs, and client expectations that the bill should drop when AI shows up — cost-focused AI is how most of us stayed current. It's not the mistake. The mistake is thinking that's the whole job.
The 20% aren't skipping the cost work. They're doing the cost work — and two other things on top.
The second lever — AI inside the advisory work you already deliver
I had a client meeting last month that illustrates what this looks like. A client who does branded clothing — retail storefront, online store, and a growing corporate business. They were deciding whether to close the physical store and lean harder into corporate. As we talked through it, the gap was obvious: when a corporate buyer lands on their website, there's nothing speaking to a corporate buyer. It's a consumer retail site.
So we built the alternative together. We used AI to draft an ideal customer profile for the corporate buyer, pointed it at their existing site, and asked it to produce a new version using the StoryBrand framework — same brand, same products, rewritten for a corporate audience. A few minutes later, they had a working prototype of a corporate-facing variant of their site.
That's not cost work. It took more time, not less. And it wasn't new revenue — we were inside the same advisory engagement we've had with them for years. It was advisory value-enhancement. The client could have built this themselves. They just wouldn't have. That's what the advisor brings: challenging the client to see what's possible, then building it with them in the room.
This is the second lever. Most CAS practices could pull it tomorrow. No new service line required — just bring AI into the room the next time you sit with a client. Your judgment drives the conversation; AI handles what used to take weeks of agency work.
The third lever — and why your niche is the prerequisite
The third lever is the hardest. It's using AI to create services that didn't exist before.
Most CAS practices can't see these opportunities because they're general practitioners. Where AI creates new revenue is industry-specific. If you serve hairdressers, dentists, restaurants, and SaaS companies in the same week, the AI-enabled products you could sell them are different for each — and you don't have the domain depth to identify them for all four.
Imagine you only served hairdressers. The opportunities sharpen. You could sell niche benchmark reports that compare a salon's performance against the industry. You could build tools hairdressers routinely don't have — commission calculators, chair-utilization dashboards, pricing models calibrated to local markets. You could encode your knowledge of the niche into an on-demand coaching program, the way PwC encoded decades of institutional knowledge into PwC One. The niche is the raw material the AI works with.
This goes against how most CAS practices are wired. We're trained to serve who walks in the door. But the firms capturing outsized AI returns aren't generalists — they've picked a niche deep enough that the revenue plays become obvious, and then they've built for it.
Making any of it compound — Champion + Operator
Whichever lever you're pulling, one-off wins don't compound into firm-wide practice without two roles staffed. A Champion discovers what's possible — the experimenter who tries the branded-clothing-site trick, the niche-metric-report prototype, the new AI tool. An Operator turns those experiments into repeatable systems — the one who takes the prototype and makes it something every client gets.
Most firms with "an AI problem" actually have a staffing problem. A Champion with no Operator means experiments pile up but nothing becomes repeatable practice. An Operator with no Champion means existing processes run faster but nothing new is discovered. The 20% have both.
The question you have to answer this quarter
Which of the three levers is your firm pulling? Cost? Advisory value-enhancement? Revenue creation? If you can only name one, that's the answer. If you can name two, you're building. If you can name all three, you're in the 20%.
Cost savings keep you in the game. Advisory value-add keeps your clients. Revenue creation is the only one that changes which game you're playing.
Ready to identify the specific revenue plays in your niche? I've built a structured prompt that walks you through your geography, niche, current services, and best-client profile — then outputs ten AI-enabled revenue ideas specific to your practice. Find your niche's revenue plays by subscribing below. The prompt is yours to use this month.