The Connected AI GTM Gap: Why PE-Backed Companies Have AI Tools But Not AI Intelligence
Most PE-backed companies are running AI tools that do not talk to each other, and that gap is costing them revenue they cannot see. The fix is better foundations.
One of the most common strategies in private equity is the platform model: a sponsor backs a strong core business, then grows it deliberately through tuck-in acquisitions that add customers, capabilities, or geographic reach.
It is an efficient path to scale, but it creates a predictable side effect. Every acquisition arrives with its own revenue team and its own definition of how go-to-market gets done. By the time the portfolio company is three or four deals deep, the go-to-market motion is fragmented by design, and no one has stopped to rebuild it as a single, unified system. Whether you are a deal partner evaluating the next add-on or an operating partner trying to drive performance across the portfolio, this problem is likely in front of you.
For example, this is what it looks like: we audited the go-to-market motion of a company with $150 million in top-line revenue. The company had absorbed four acquisitions over a few years, and each one brought its own Leadership, RevOps management, and their own tools and processes. By the time we finished counting, the number was 57 different go-to-market tools. That is roughly 40 too many.
On top of all of that, the company had added Claude and Copilot. Both were running and reading the same data, but were giving different answers.
Both are looking at 2 plus 2, but no one's getting 4.
The AI was not broken, but the foundation underneath it was. Neither model was anchored to clean processes or integrated around a shared system of record, which resulted in everyone being misinformed.
This is not a story about a bad RevOps hire, but a structural problem, and it is showing up across PE-backed portfolio companies right now.
The real problem underneath the tools
The infrastructure in place becomes what I call "shoestrings and bubblegum": getting worse with every new addition. The deeper error is that PE portfolio companies have treated revenue predictability as a tooling problem. The foundational work, things like buyer segmentation, personas, sales process, and territory alignment, came after the technology.
You can't fix AI and tool fragmentation until you understand what you want those tools to enable inside your go-to-market motion.
Operating partners are facing a harder version of this than most realize. They have limited bandwidth to evaluate every AI initiative crossing their desk.
When it all goes into ChatGPT
The process failure shows up in two layers, and one feeds the other. The first is visible. The second is the one that spiderwebs out.
In the visible layer, reps bypass the system. When a CPQ isn't built to match how customers actually buy, reps work around it. They generate quotes outside the system, and finance never gets notified. Customer success never gets cued for onboarding, which causes handoffs to break, and the dashboard never shows it.
In the company we audited, a $527,000 revenue gap never made it into the ERP or the revenue numbers.
Imagine reporting up to a board and saying, oops, we missed over half a million in the last board meeting.
For a company heading toward an exit, the problem compounds.
If you can't report numbers accurately in your system, it becomes harder to tell a great story when you're going to exit. This results in KPIs filling the screen with little to no integrity. Now that untrustworthy data is feeding something new: LLM models.
This is the second layer underneath GTM execution; it is moving faster than Operating Partners have the bandwidth to inspect. AI adoption is already underway across portfolio companies, with some having little to no governance in place. Sales and marketing are licensing their own tools, which puts security and intellectual property at risk. Then someone loads that same data into ChatGPT or Claude, without the business context or foundation that would make the output useful.
The result is poor messaging to potential customers and current clients, AI slop marketing campaigns, poor sales enablement tools (i.e. battle cards, sales plays, etc.). All of this eats away at NRR and revenue per head, metrics that most PE firms are using to evaluate AI effectiveness.
This leaves every PE and Portco management team asking the same question: How do I get ahead of this?
Foundation first, not tools
There is a sequence to this. Most companies skip it.
Before adding AI, start by understanding your go-to-market problems and desired organizational results. This is how you do it.
The diagnostic runs five steps, in order, each one building on the last:
- Stakeholder interviews: what are the actual problems? Remove AI from the framing first.
- AI usage review: what is already running inside the company, managed or not?
- Tech stack review: optimize or remove tool usage based on business requirements
- Workflow and process analysis: Evaluate your automations and how these systems are interconnected
- Data integrity assessment: Bad data and invalid inputs yield poor results
The one rule that governs everything is context. Run the diagnostic first. Your foundations, segmentation, personas, clean data, and unified processes, are the inputs to that engine. Get the context engine right, and better output follows. Get it wrong, and every AI tool layered on top amplifies the problem.
Successful AI isn't a novelty; it is a practical lever that solves real business problems. Here are two examples:
Case Study 1: Water treatment company, 15 sellers, zero wasted bets.
The client was a water treatment company with municipal customers. Fifteen sellers covered the United States. The core challenge was clear: government budgets cannot be stimulated. Money only moves when government officials vote to spend it.
By providing AI with real context, this company was able to scrape publicly available board meeting transcripts for trigger keywords that precede RFP issuance. Those keywords became leading indicators inside Salesforce. The rep knows the signal is building before the RFP drops.
What made it possible: clean process documentation, Salesforce as the system of record, and defined personas. The foundation came first. The AI layer worked because it had something real to run on.
Case Study 2: Cross-sell across 100+ portfolio companies
The same logic applies at the fund level.
A large PE firm with more than 100 portfolio companies. The firm was struggling to systematically map where referral opportunities existed across the portfolio.
"B2B deals sourced through referrals are 4x more likely to close than cold outreach." (Harvard Business Review)
We used AI to map customer profiles across portcos and surface cross-sell paths. One portco sold IT management services, and its strongest customers were in higher education. Another portco in the fund operated in that same sector, and AI surfaced the match. The operating partner now had a lever that few individuals portcos could see on their own.
The first move
Neither of these companies had an AI problem but rather a sequencing problem. The companies getting this right are not the ones with the most AI tools. They are the ones who did the hard work first and did it in the right order.
Where is your go-to-market organization on the AI journey?
Answer a few questions and get a read on how ready your go-to-market setup actually is for AI. 14-question assessment across 7 dimensions - from Guesser to Enabler.
