Proserv & PE Insights · Saturday, September 26, 2026

Hi, this is Koko from Koko Knows.

Let's get right into what's new. KPMG had a big week. They published their Q3 2026 quarterly AI survey, over two thousand one hundred senior leaders across twenty countries, and in the same stretch they're pushing a new IT Maturity Assessment product to help CIOs figure out where to spend on AI readiness. Two fee-adjacent plays landing at once isn't a coincidence. This is KPMG building an advisory-led sales motion around AI governance and infrastructure readiness, positioning themselves to own the entry point of the engagement before a client has even scoped a deployment. Surveys generate leads, the maturity assessment converts them into billable diagnostic work. It's a smart, repeatable funnel, and it's worth watching whether the rest of the Big Four answer with their own version.

Second big signal, and this one's a gut check on the whole AI-consulting narrative. Accenture's research found that only twenty percent of bank leaders are seeing widespread, sustained AI value. Read that again. Eighty percent of banking leadership is not seeing durable returns on their AI spend. For those of us who've been pricing transformation engagements on the promise of AI-driven efficiency, that's the number that matters most this week. It means the consulting firms keep getting re-hired to chase value that hasn't materialized yet, which is good for utilization in the near term, but it's also a credibility clock. At some point clients start asking why the last three engagements didn't close the gap, and that's when scope and fee structures get renegotiated downward, or worse, in-housed.

Third, PwC US added five thousand staff year over year, and the mix is shifting hard toward engineers and data scientists paired with business architects. This confirms something we flagged last week, that "two in the box" delivery model, pairing a technical builder with a business translator, isn't just a talking point in partner interviews anymore. It's actually showing up in headcount data. That's a meaningful signal for how the delivery stack is being rebuilt across the industry, and it tells you where the war for talent is heading. If you're competing for the same engineering and applied-AI talent PwC is hiring at scale, budget for it now, because the compensation pressure is going to intensify.

Now here's the pattern underneath all three of these stories, and it's the thing I'd flag to any partner or MD reading the tea leaves. These firms are monetizing the AI-value gap itself. The survey, the maturity assessment, the reskilling push, none of it requires the client's AI initiative to actually succeed. It just requires the client to keep believing success is achievable and one diagnostic away. That's a durable fee stream right up until it isn't, and the risk sits quietly underneath the delivery model. We still don't have resolution on PwC Switzerland's bonus cuts or on how KPMG's headless delivery model with OpenAI actually gets staffed and priced. Those are the comp and delivery-risk threads to keep pulling.

If I zoom out to what this means for your own portfolio and client conversations, three things stand out. First, on growth and margin, the market is rewarding firms that package readiness and governance as standalone products rather than folding them into a broader transformation scope. If you're advising or building a professional-services asset, look hard at whether you have a discrete, sellable diagnostic offer or whether you're still bundling everything into one long engagement that's harder to price and harder to close quickly.

Second, on talent and delivery risk, the shift toward embedded engineers and data scientists inside client-facing teams is now a headcount fact, not a slide in a pitch deck. If your portfolio companies are competing for that same talent pool, expect wage inflation and plan retention economics accordingly, especially in applied AI and ML roles.

Third, and this is the sharpest point of view worth carrying into your next board meeting or LP conversation, system integrators and the big consultancies are inserting themselves as the mandatory adoption layer between frontier AI vendors and enterprise buyers. Between OpenAI's alliances with the major consultancies and this steady drumbeat of readiness products from KPMG and others, assume that direct-to-enterprise agent sales are going to carry a consulting tax. If you're pricing a platform or vendor deal, or evaluating a portfolio company's go-to-market, factor that gatekeeper layer into your assumptions now rather than discovering it in a lost deal.

That's where things stand. The through-line this week is firms getting very good at selling the diagnosis of the AI-value gap while the underlying delivery and liability questions stay unresolved. Watch that gap, because it's where both the fee growth and the risk are compounding at the same time.

That's it for this update. For the full details and sourcing behind everything I covered, head over to Koko Knows. Talk soon.