95% of funds say AI is working- the other 5% are being honest
Hey there!
I recorded a new episode of PE Data Guy this week with Ryan Krook, who has one of the more interesting career arcs I’ve come across. Ontario Teachers pension plan, McKinsey inside PE portcos, data science at Uber and Shopify, now building data and AI tools for PE-backed companies.
We covered a lot of ground, but the phrase that stuck with me was about vibe coding. If you haven’t heard the term, it’s the wave of non-technical people using AI to build software tools by just describing what they want in plain English. It sounds great until you think about what happens next.
I hope your respective Spring Breaks are going well! Enjoy the rest of the memo.
Cheers,
Graeme
Three Things I Learned This Week
Vibe coding is creating the Excel problem all over again
This came up in my conversation with Ryan Krook on PE Data Guy this week. Vibe coding lets anyone build tools and dashboards by talking to an AI. No engineering background required. On the surface, that sounds like progress. And in some ways it is.
But Ryan made a sharp observation. Twenty years ago, Excel gave everyone the power to build their own reports and models. What we got was thousands of ungoverned spreadsheets, no version control, no audit trail, and finance teams spending weeks reconciling numbers that should have matched. Vibe coding has the same structural risk. You get fast creation with zero governance. The tool works on day one. On day ninety, you have fifteen dashboards built by fifteen people, all pulling from slightly different data, and nobody can tell you which one is right.
If you’re running or overseeing a portfolio company, this is worth paying attention to. Not to block it, but to get ahead of the governance question before it becomes an audit finding. Watch the full conversation here.
56% of PE firms are now using alternative data, up from 31% in 2022
KPMG’s latest numbers show PE firms are treating data as an asset class in its own right. Not just financial statements and management decks. Alternative data. Web traffic, satellite imagery, credit card transaction trends, and employee sentiment signals. The firms using it are making faster decisions, catching problems earlier, and building conviction on deals where the traditional data is ambiguous.
The shift from 31% to 56% in three years is not gradual adoption. That is a market that has decided this works. And it means the firms not doing this are now behind, not neutral. If the fund buying your portfolio company is running alternative data analysis on your business before they even call, you want your own data house in order before that call happens.
The discipline of starting small
Something else Ryan said during our conversation: the most successful data programs he’s seen in PE-backed companies all started the same way. Not with a grand strategy or a transformation roadmap. They started with getting reporting right.
I think ambitious people, and I count myself in this group, have a bias toward big moves. It feels productive to design the end state. The architecture diagram. The platform. But the companies that actually get somewhere start with one accurate, trusted report, then build from there. It is boring work. It does not look impressive in a board deck. But it compounds in a way that big plans announced and never finished never do. Something I keep reminding myself of in my own business, too.
Two News Stories From This Week in Mid-Market PE & Data
FTI Consulting reports 95% of funds say AI is meeting or exceeding expectations (source)
What happened. FTI Consulting released their 2026 PE AI Radar, surveying funds across the market. The headline number is striking. 95% of respondents say AI is meeting or exceeding their original business case. But the detail underneath tells a different story. Talent is the number one constraint, cited by 35% of respondents. Adoption is deeply uneven across portfolio companies within the same fund. Revenue acceleration is the top priority for 41% of firms. And the performance gap between funds that have figured out AI deployment and those that have not is widening, not closing.
Why you should care. The 95% number will make it into pitch decks and board presentations by next month. But before you take it at face value, think about what “meeting expectations” actually means when talent is the biggest bottleneck and adoption is uneven across portfolios. If your best portco is running AI well and your other four are not, the fund-level number looks fine while most of the portfolio is stuck. The real question for your next board meeting is not whether AI is working. It is whether your specific portfolio company has the data infrastructure and people to make it work. That is a much harder question with a much more useful answer.
PE deal value hit $904B in 2025 with revenue growth driving two-thirds of the value bridge
What happened. Bain’s 2026 Global Private Equity Report puts 2025 deal value at $904 billion, a 44% year-over-year increase. Average deal size hit a record $1.2 billion. But the number that matters most for operators is from MSCI. Revenue growth now accounts for two-thirds of the value bridge in PE exits, not multiple expansion. Separately, Bank of America launched a dedicated Private Capital M&A unit specifically for PE exits, and Ares is capping withdrawals from its $10.7 billion fund.
Why you should care. For years, multiple expansion did a lot of the heavy lifting in PE returns. That is over, at least for now. If revenue growth is driving two-thirds of the value bridge, your ability to prove that growth with clean, auditable data is more important than it has ever been. Buyers will not take your word for it. They will trace revenue from CRM to invoice to cash receipt and expect the numbers to tie. Every gap in that chain is a negotiating lever for the buyer and a discount on your exit. The firms that can tell a clean, defensible revenue growth story with data that holds up will trade at a premium. Everyone else is in for a longer process and a harder conversation.
Free Tool of the Week - VCP Data Score
Most value-creation plans assume the data exists to execute them. Very few actually check. The VCP Data Score is an interactive assessment that scores portfolio company data readiness across five dimensions, from revenue data to governance maturity. Takes two minutes. Gives you something concrete to bring to the next operating partner review.
Sign-off
If this resonates and there’s something you think we can help with, just reply to this email. We read everything that comes through.
As always, please forward this on to your favorite PE-backed friend.
Cheers, Graeme