SuperReturn 2026: The AI maturity gap is becoming a valuation gap

AI is no longer an innovation topic: it is a valuation topic. In Q1 2026, approximately $300 billion of market value evaporated across SaaS, data, and software-heavy companies. This was not a general market correction. It was a structural repricing. Investors were making a judgment about which businesses AI would strengthen and which businesses AI would commoditize. 

That same question dominated SuperReturn International in Berlin this week. AI is increasingly becoming a multiple-expansion lever, not just a productivity lever. What emerged was a picture of an industry at a genuine inflection point. Not the inflection point of AI adoption, which has already happened, but the inflection point of AI economics: who is actually changing their financial profile through AI, and what specifically are they doing that everyone else is not.

A joint session between Permira and AWS laid out the most precise answer this conference has produced on that question. OMMAX’s latest whitepaper on B2B software repricing in the AI era provides the analytical framework that makes sense of what the most advanced portfolios are now demonstrating in practice.

1. The $300 billion signal the market has already sent

The repricing was not random. Investors were discounting a specific type of asset: businesses where AI adoption has produced interface improvements and feature additions without changing what the business actually does or what it is worth. The assets being rewarded are those where AI has moved the business toward controlling operational workflows, transactional rights, and the data generated by execution. 

The market has made a judgment that the software industry is bifurcating. On one side are platforms that have secured execution authority: the ability to initiate actions, enforce business rules, automate decisions, and capture recurring revenue from operational outcomes rather than seat licenses. On the other are companies that have added AI features to existing architectures without fundamentally changing what they do or what they are worth. 

For private equity, this is not a software-sector observation. It is a portfolio-wide valuation signal. The assets most exposed to this repricing are those where AI adoption has produced demos and dashboards rather than embedded workflow control. The assets most protected are those where AI has moved the business along the spectrum from system of record toward system of action: from storing operational data toward autonomously executing operational decisions. 

As Konstantin Kugler, Partner Strategy and Transactions at OMMAX, noted in the firm’s whitepaper on B2B software repricing:

"AI readiness and maturity are becoming a structured diligence dimension, especially in the B2B software sector. Buyers increasingly assess automation depth, execution embedding, monetization resilience, and dependency on seat-based expansion. AI is no longer a product roadmap topic. It is a multiple protection topic."

2. Software economics are being restructured, and the numbers are precise

The most consequential insight from the Permira session was one that tends to get lost in conversations about AI tooling and use cases. As a service or software company embeds AI more deeply into the operational workflows it controls, its financial profile changes structurally. A conventional service business operates at around 30% gross margin. A pure software company operates at 85 to 98%. AI is the mechanism by which businesses can move along that spectrum, and the pace of movement is accelerating. 

The OMMAX whitepaper identifies five KPIs that now separate companies structurally capturing AI value from those merely adding AI features. Net revenue retention above 110 to 120% is the clearest indicator of compounding value, reflecting successful cross-sell of AI modules, deeper workflow penetration, and growing share of wallet. Gross revenue retention above 95% signals mission-critical embedding and high switching costs. When more than 50% of new ARR comes from expansion rather than net new customers, growth becomes self-reinforcing. Sustained EBITDA margins of 25 to 35% indicate scalable unit economics and productized AI that drives operating leverage. And sustained organic growth above 15% signals genuine product-market fit and pricing power, not just experimentation.

Together, these five metrics describe a business that has made the transition from tool provider to workflow owner. The operating leverage effect that accompanies this transition is what makes it particularly valuable for PE investors. As gross margins expand through tech enablement, a company gains the capacity to add revenue at a structurally lower incremental cost. Growth accelerates while unit economics improve simultaneously. For investors with defined hold periods and specific exit targets, actively managing a portfolio company along this spectrum is one of the most direct sources of multiple expansion currently available. 

In practical terms, AI is increasingly becoming a multiple-expansion lever in its own right. The firms that successfully convert software from workflow support into workflow execution are not only improving earnings but also changing how those earnings are valued. A business with an NRR above 110%, GRR above 95%, and more than half of new ARR coming from expansion does not trade on service company multiples. It trades on software company multiples. That is the financial prize at the end of the execution authority journey, and it is why the most sophisticated buyers are now assessing AI maturity as a structural component of valuation, not a feature of the investment narrative. OMMAX's analysis highlights a significant gap between cloud-native incumbents and companies operating at the agentic workflow layer. The latter are expanding addressable markets, improving retention, increasing expansion revenue, and strengthening operating leverage. As a result, investors increasingly reward these businesses with premium valuation multiples. The difference between those two positions is not a feature gap. It is a valuation gap. 

3. What conviction looks like in practice 

Two transformations from the Permira portfolio illustrate what that margin journey actually looks like when executed with real capital and real accountability. 

Octus, the global credit intelligence provider formerly known as Reorg, invested approximately 12% of revenue into technology and launched Credible AI, a product that fundamentally changed how credit analysts access and synthesize information. The financial results were not incremental. ARR per client tripled, driven by higher engagement and price increases justified by output that was qualitatively incomparable to anything the market offered before. Publishing time fell by 90%. The company has sustained growth above 30%. This was not AI added as a feature. It was a rebuilt core product that changed the economics of the relationship with every client and moved the business firmly toward the workflow-owner end of the spectrum. 

Acuity, a knowledge process outsourcing firm with 6,000 analysts, presented a different kind of test. When Permira evaluated the investment in April 2023, the obvious concern was existential: were they acquiring a business that AI would displace? Permira committed 20 million euros to build Agent Fleet, an AI platform designed to sit alongside the analyst workforce and automate the workflows that did not require human judgment. Today, 15% of company revenue is generated by fully automated processes with no human involvement. 30% of total workforce capacity runs through automation. Core workflow efficiency has improved by around 20%. The fear of disruption was converted into the primary mechanism of value creation. 

What both cases share is a decision-making posture that the OMMAX whitepaper defines precisely: moving from feature delivery toward execution authority. Octus and Acuity do not simply use AI. They have operationalized it as the execution layer of their businesses, which is exactly what the five-KPI framework identifies as the signal of structural, compounding value. 

4. The agentic shift is already live, and most portfolios are behind 

One data point from the Permira portfolio warrants particular attention. 25% of portfolio companies have already moved agentic AI into live production environments. Not tests. Not pilots. Live, revenue-generating production systems that can initiate actions, run multi-step processes, and operate with humans in the loop only where required. 

The OMMAX AI maturity framework maps this progression across five levels: from point solutions and embedded AI toward workflow AI, cross-system orchestration, and ultimately autonomous AI operations achieving 80 to 95% automation. Each level represents a structurally different set of economics, and the gap between where most PE-backed companies currently sit and where the value is concentrated is the central challenge of this vintage. 

The AWS perspective reinforces the scale of this shift. Across approximately 18,000 AI projects deployed with around 120 PE firms globally since the start of the generative AI era, the projects generating the clearest financial return are concentrated in product embedding and front and middle-office operations. Software vendors lead, with 93% of AI project volume directed at their own products and services. The back office, as a proprietary investment priority, has largely been commoditized by third-party vendors. The funds still treating it as a competitive AI investment are misallocating both capital and management attention. 

The whitepaper also draws an important asymmetry between horizontal and vertical software that has significant implications for portfolio construction. Horizontal SaaS, built around generic cross-industry workflows, faces commoditization as general-purpose AI agents replicate broad use cases. Vertical SaaS benefits from defensibility rooted in proprietary, regulated domain workflows and structured data. For vertical leaders, AI accelerates expansion. For horizontal players without execution authority, it poses structural risk. Understanding which category each portfolio asset occupies is now a prerequisite for accurate valuation. 

5. How the best funds are using AI risk to shape what they buy 

One of the most significant structural changes in how leading PE funds are operating is the integration of AI risk directly into the investment committee process. Permira has made this explicit. The firm has developed clear guidelines for evaluating AI opportunity against business durability, and it has made deliberate decisions to exit or avoid certain market segments as a result. 

The logic is precise. Generic tax advice is an example of a segment where AI dramatically reduces the cost of the service but does not drive incremental demand: individuals still file one tax return a year regardless of how cheap or fast the advice becomes. Certain media software categories face similar constraints. In these segments, AI disruption compresses margins without creating growth, which means the investment case deteriorates even if the technology is adopted successfully. 

The OMMAX whitepaper formalizes this into a framework built around two strategic imperatives: protecting TAM and expanding TAM. Protecting TAM means reinforcing vertical depth, system-of-record anchoring, proprietary data ownership, and compliance-native architecture. These are prerequisites for durability. Expanding TAM means evolving the product toward execution authority: the ability to convert AI-driven automation into scalable, outcome-based ARR. The two imperatives are interdependent. Defending the core creates the foundation and credibility needed for sustainable AI-driven expansion into operational budgets. 

The implication for diligence is direct. The question is no longer whether a target company uses AI. The question is where it sits on the maturity stack, whether it controls mission-critical workflows and structured operational data, whether its retention metrics signal execution-layer embedding or discretionary tooling, and whether its revenue growth is compounding through expansion or dependent on continuous net-new customer acquisition. 

6. Token spend is the new cloud spend, and the right response is not what most firms think

The token maxing problem is real, but it is being misdiagnosed. The instinct across the industry has been to apply hard limits: rollbacks, consumption caps, leaderboards tracking individual token usage. Amazon itself rolled back an internal software engineering leaderboard tracking token consumption just two weeks before SuperReturn, a visible signal that blunt cost suppression is not the right mechanism. 

The OMMAX whitepaper makes the connection explicit: R&D efficiency must be tracked as a complementary performance metric. In the agentic AI era, it is not enough to increase engineering and AI-tooling spend. What matters is whether that investment translates into scalable expansion ARR. Only when R&D cost, AI tooling, and token spend remain aligned with commercially relevant product output does AI create true engineering leverage rather than simply adding cost. 

The right approach has three components. First, apply FinOps principles: introduce genuine visibility into what is being spent and why, with the same governance rigor applied to cloud spend a decade ago. Second, route tasks to cost-appropriate models. A modern AI workflow involves multiple model interactions within a single process, not all of which require the most capable or most expensive model available. Third, train and incentivize. The organizations making durable progress are rewarding efficient spending rather than penalizing consumption. Cost optimization built through organizational capability compounds. Cost suppression built through hard stops does not. 

The funds that build this discipline early will compound the advantage as AI embeds more deeply into every commercial and operational function. The ones that apply blunt suppression will find that their teams route around the controls or stop experimenting entirely, and the innovation gap with more disciplined competitors will quietly widen. 

What this means for PE funds and their portfolios 

The conversation at SuperReturn this week was not about whether AI creates value in private equity portfolios. That question has been answered. The most advanced portfolios and OMMAX's work across 4,000 digital value creation projects and more than 500 transaction advisory engagements have produced evidence that is substantial and growing. The conversation this week was about the choices that determine which funds capture that value and which ones accumulate cost while their competitors pull ahead. 

Four choices separate the leaders from the rest:

1. Judge AI by whether it creates execution authority, not by feature count. The OMMAX whitepaper scorecard is precise: NRR above 110 to 120%, GRR above 95%, expansion revenue above 50% of new ARR, EBITDA margins of 25 to 35%, and organic growth above 15%. Companies meeting these thresholds are capturing execution-layer value. Companies missing them are adding AI features.
2. Move from pilots to production with real capital. The Permira examples are not proofs of concept. They are proofs of conviction. Twenty million euros committed to Acuity. 12% of revenue invested at Octus. Results that are durable because the decisions behind them were treated like any other investment decision in the portfolio.
3. Apply FinOps discipline to token spend now, not after costs become a political problem. Visibility, intelligent routing, training, and positive incentives. Not hard stops. The organizations that build this capability early compound the advantage. The ones that apply blunt suppression slow their own innovation and widen the gap with more disciplined competitors.
4. Integrate AI risk into the investment thesis from entry. The durability of an AI investment case matters as much as the size of the opportunity. The funds building AI risk assessment into their IC process are making better entry decisions. The funds that do not discover the problem at exit, when the gap between the AI narrative and the verifiable financial evidence is no longer something sophisticated buyers will overlook. 

The question is no longer whether your portfolio is running AI. The question is whether it is building the financial profile, the organizational conditions, and the investment discipline required to make AI a durable source of competitive advantage rather than a source of cost that has yet to find its return. 

From AI strategy to value creation 

OMMAX is a leading AI-first management consultancy and AI-engineering platform supporting investors and portfolio companies across AI strategy, business transformation, transaction advisory, and value creation in the age of AI. Our work spans the full transformation journey: identifying value creation opportunities, assessing AI readiness during due diligence, designing operating models and governance structures, building organizational capabilities, and scaling AI initiatives into measurable growth, profitability, and exit outcomes. 

If you would like to discuss any of these themes or explore how OMMAX supports PE funds and their portfolio companies in building durable AI economics, please reach out to the authors of this article.