Your Portfolio Companies Have an AI Story. Can They Prove It?

Bain & Company's latest tech investing report - The New Era in Tech Investing Starts Now - lands a sharp observation that should stop every PE deal team in its tracks: the most important question in software due diligence is no longer "Do they have an AI strategy?" It's "Can they show me the proof points?"

That shift in emphasis from narrative to evidence is more significant than it first appears. It marks the moment that AI stops being a story you tell and starts being something you have to demonstrate - in working product, in measurable outcomes, in metrics that stand up to scrutiny. The problem is that most software companies aren't ready for that scrutiny. And most engineering organisations don't know how to build toward it. That's the gap Codurance exists to close.

What This Article Explores:

• Why AI due diligence is shifting from strategy and narrative to measurable proof points that can stand up to buyer scrutiny.
• What Bain’s latest tech investing report signals for PE-backed software companies, including the pressure on growth, retention, margins and traditional SaaS metrics.
• Why many portfolio companies face an execution gap when trying to turn AI ambition into working product, measurable outcomes and a credible exit story.

What Bain Got Right

The report is clear-eyed about what's changed in the market. Revenue growth in software has roughly halved. Net revenue retention is in decline. Post-pandemic buyouts are tracking below historical returns. And AI - rather than offering a straightforward path to recovery - introduces a new layer of risk: it threatens the embedded moats, seat-based pricing, and structurally high margins that made software the PE industry's most reliable asset class for a decade.

The Bain authors identify three things GPs (General Partners) need to do differently: 

• Rethink due diligence - traditional SaaS metrics increasingly mislead. ARR and NRR were built for a world of near-zero marginal cost and predictable retention. They don't capture AI impact. The diligence question becomes: is this company actually deploying AI in ways that are measurable, scalable, and defensible or are they spinning a narrative without numbers behind it?
  

• Transform the product during ownership - incremental AI feature additions aren't enough. The winners are the companies that zero-base their product assumptions, go deep on customer workflows, and ask what would genuinely change the outcome, not just speed up a human.

• Build a measurable exit story - buyers are demanding evidence. That means separating AI revenue into distinct buckets, tracking AI-specific cost structures, and answering three questions with precision: Is AI driving incremental revenue? Is it changing cost structures? Are AI products scaling efficiently?
  

All of this is correct. But there's a dimension the report doesn't fully address: who actually does this work inside a portfolio company, and how?

The Execution Gap 

The honest answer, in most cases, is that the engineering organisation isn't ready. Not because the talent isn't there. But because integrating AI at the product level - in a way that generates real proof points rather than impressive demos - requires a set of practices, habits, and architectural foundations that most incumbent software companies haven't had to build before.

It requires engineers who know how to instrument AI features for measurement from day one. Teams that can distinguish between genuinely agentic capability and an LLM wrapper. Technical leadership that can make the hard calls about what to stop, what to rebuild, and what to buy - rather than accumulating AI initiatives that consume capacity without producing signals.

Bain's Zendesk example illustrates the scale of the organisational challenge. In roughly 18 months, Zendesk shifted more than half its workforce onto AI work (up from under 10%), made painful trade-offs to collapse overlapping initiatives, and rebuilt its platform around autonomous customer service resolution - ultimately delivering $200 million in AI ARR. That required extraordinary execution discipline, not just a compelling vision.

Most portfolio companies aren't Zendesk. They're running leaner teams, carrying technical debt, and managing a core product that still needs to generate revenue while the AI transformation is underway. The transformation doesn't happen because someone decided it should. It happens because someone builds the engineering capability to execute it.

What Codurance Does

We are a software craftsmanship consultancy. Our work sits at the intersection of engineering excellence and AI adoption and increasingly, that intersection is exactly where PE value creation happens.

When we engage with a portfolio company, we're not there to produce a strategy deck. We're there to help engineering teams build the proof points that matter.

• Engineering and AI Readiness Reviews give deal teams and operating partners a clear-eyed picture of the current state not just of what AI initiatives are underway, but how instrumented they are, how the team is organised to execute, where technical debt is constraining AI deployment, and how realistic the AI roadmap actually is. This is due diligence that goes beyond product demos and strategic narratives.

• Embedded AI Champions accelerate the organisational shift that Bain describes. We place experienced practitioners inside engineering teams, working alongside them to build the habits, the practices, and the measurement culture that turns AI experimentation into scalable, demonstrable traction.

• Engineering Capability Uplift addresses the underlying capability gap. We work with engineering organisations to raise the technical baseline - the test coverage, the architectural discipline, the deployment practices - that makes rapid, confident AI iteration possible in the first place.

  The output isn't a report. It's working software, measurable outcomes, and the internal capability to keep building after we've left.

The Proof Point is the Exit Story

Bain is right that 2026 is the year AI truly redefines the software industry. The companies that emerge from this transformation with strong exit stories won't be the ones with the boldest AI vision. They'll be the ones that built an engineering organisation capable of turning that vision into evidence - and then built the metrics to prove it.

If you're a GP looking at a software asset that has an AI story but can't yet answer the three questions - incremental revenue, cost structure change, scaling efficiency - the gap isn't strategic. It's operational. And it's closeable.

That's the work we do. Codurance is a software craftsmanship consultancy helping engineering organisations build the technical foundations for AI-driven growth. We work with PE-backed software companies at the point where strategy meets execution turning AI ambition into proof points that hold up in diligence.

Get in touch with us today to learn how we can support PE value creation. 

Frequently Asked Questions

1. What is the AI execution gap in portfolio companies?

The AI execution gap is the difference between having an AI ambition and having the engineering capability to deliver it. Many companies have promising ideas or demos, but lack the architecture, measurement practices, delivery discipline or technical foundations to turn them into scalable products.

2. How can portfolio companies build a stronger AI exit story?

They need to show clear evidence that AI is improving the business. This could include AI-specific revenue, reduced support or delivery costs, improved customer outcomes, higher product adoption, better margins, or stronger defensibility against competitors.

3. What AI proof points should PE-backed software companies be able to show?

They should be able to show how AI is being used in the product, how its impact is measured, whether it is creating new revenue or efficiency gains, and whether the underlying engineering organisation can scale AI features reliably.

About the Author

Alan is a Director of Client Solutions at Codurance, with more than 20 years’ of experience in technology. Alan works with organisations to shape technology strategies and software delivery approaches that enable meaningful business change. Alan also works with clients to explore how AI-enabled engineering practices and AI-powered systems can unlock new opportunities while maintaining the quality, reliability and discipline required for production systems.