The News
Banyan Software has published its inaugural State of AI for Vertical Market Software Operators report, drawing on responses from over 260 founders, CEOs, and operators. The central finding is stark: 51% of software operators report that fewer than one in four of their customers actually use the AI features the company built. The report further identifies a maturity gap, with AI leaders generating 45% revenue impact from AI compared to just 9% for the rest of the field, and AI-leading companies nearly 20 times more likely to report improvements in customer retention.
Analyst Take
The Deployment Problem Is the Real AI Problem
The headline statistic here deserves to be read carefully. This is not a story about companies that failed to ship AI. It’s a story about companies that shipped AI and still failed. Half of vertical software operators are carrying the engineering cost, the security overhead, the support burden, and the organizational distraction of AI features that the majority of their customers never touch. That is a meaningful misallocation. The gap between “we built it” and “they use it” is precisely where AI investment goes to die.
This finding aligns with a broader pattern ECI Research has tracked in engineering capacity. According to ECI Research’s 2026 Application Development survey, 65.2% of respondents said that 0–20% of engineering time is spent on net-new innovation. When most engineering cycles are absorbed by maintenance, integration, and operational overhead, the features that do get shipped rarely receive the sustained go-to-market and onboarding investment needed to drive actual customer activation. Building an AI feature and embedding it into customer workflows are two entirely different disciplines, and most software companies are only funding the first one.
What Separates AI Leaders Isn’t What You’d Expect
Banyan’s data pushes back on the intuitive assumption that bigger, better-funded companies win the AI race. Bootstrapped and PE-backed companies progressed at nearly identical rates (32% vs. 33%). The differentiator is execution rigor: assigning ownership, measuring activation, and integrating AI into workflows rather than bolting it onto product surfaces. That is an organizational design problem, not a technology problem.
For developers, the practical implication is that instrumentation of AI feature usage is not optional. If your organization cannot tell you what percentage of customers activated a given AI capability last quarter, you are operating blind. Activation telemetry, cohort analysis, and in-product feedback loops are the engineering investments that actually validate whether AI features are working. This connects directly to where enterprise teams are putting observability dollars: according to ECI Research’s 2026 Application Development survey, 61.7% of respondents reported having AI-driven anomaly detection as part of their observability strategy. That same discipline applied to user behavior and feature adoption would dramatically sharpen AI product decisions.
The ITDMs’ Lens: Revenue Impact and Retention Are the Scoreboard
For IT decision-makers at software companies, the 45% vs. 9% revenue impact split is the number that should anchor budget conversations. AI leaders are not just building more features; they are measuring outcomes and iterating. The 37% vs. 2% gap in customer retention improvement is even more striking. Retention is a compounding metric. A company that improves retention by even a few points through well-adopted AI will outperform one with flashier features and poor activation over any multi-year horizon.
The lesson for ITDMs is that the ROI calculation on AI development needs to include activation cost, not just build cost. Workflow integration, onboarding design, training, and change management for end customers are not soft considerations. They are the primary variables that determine whether the AI investment generates returns or becomes a line item in a future “biggest regrets” survey. Banyan’s own finding that 44% of operators regret not starting sooner should not be read as an invitation to rush. It should be read alongside the activation data: starting sooner only helps if you measure what happens next.
Looking Ahead
Banyan’s benchmark will likely become an annual reference point for the vertical software market, and the maturity index methodology (deployment level, adoption tracking, business impact measurement) gives the industry a practical framework to self-assess. Expect competitors in the vertical software acquisition space to publish similar benchmarks, and expect PE-backed software portfolios broadly to start requiring AI activation metrics as part of their standard operating reviews. The question of how to measure AI success is moving from philosophical to contractual.
The deeper shift this report signals is a reorientation of the AI product conversation from features to outcomes. Over the next 18 months, the vertical software companies that pull ahead will be those that treat customer activation as a product discipline equal to engineering delivery. Those that don’t will find themselves in the uncomfortable position of explaining to investors why their AI roadmap is full but their adoption numbers are empty. The deployment gap is the defining competitive variable in vertical software AI for 2026 and beyond.
Stay Ahead of Application Development Trends
Get weekly analyst insights, research notes, event coverage, and AppDevANGLE updates delivered directly to your inbox.
Subscribe for Weekly Insights
Join technology leaders, practitioners, and GTM teams following the trends shaping modern software delivery.
Looking for deeper research access?
Explore ECI Research reports, survey insights, and market analysis through the ECI Research Portal.
