Productivity

Pax8: 61% of SMBs are using AI, but nearly a third can't get out of the pilot phase

The Q2 2026 SMB AI Pulse Report finds active use has plateaued while 29% of adopters remain stuck experimenting — a deployment gap Gartner and PYMNTS data suggest is now separating winners from the rest.

By Esi Boateng · Productivity tools · 2026-07-15 · 3 min read

Photo: Unsplash / Bench Accounting — Small-business owner working at a laptop with notebooks and paperwork on a desk

Pax8’s Q2 2026 SMB AI Pulse Report, released July 13 and conducted by Propeller Insights across 402 U.S. small-business leaders (companies of 5 to 499 employees, ±4.9 points at 95% confidence), landed with a number that looks like stability and reads like a warning: 61% of SMBs are actively using AI, down a hair from 62% in Q1. Adoption has plateaued. The interesting figure is behind it.

Twenty-nine percent of adopters are still experimenting. The “interested but haven’t started” cohort collapsed from 9% to 1.5% in a single quarter, an 83% drop, meaning the fence-sitters didn’t just get off the fence, they’ve walked straight into a pilot and stopped. Only 11% of experimenters have a documented AI policy. Just 68% report leadership is fully or mostly aligned on AI’s role, versus 91% among active users.

That alignment gap is the whole story.

“AI is leveling the playing field for small businesses. Our data found the SMBs pulling ahead share three things: aligned leadership, governance, and the expertise to turn experimentation into results,” said Scott Chasin, CEO of Pax8. Read structurally, that’s a description of what pilot-stage firms don’t have.

The macro pressure is arriving on schedule. Gartner’s July 1 release estimates up to $234 billion in enterprise application spending is exposed to agentic arbitrage by 2030, with agentic AI absorbing roughly 20% of SaaS spend. Torsten Slok of Apollo warned in Fortune on July 6 that “companies will slow their AI spending if they don’t see ROI quickly,” citing an MIT finding that only 5% of firms saw meaningful returns from generative AI pilots. The window for staying in experimentation without a business case is closing.

What deployment looks like when it works is legible in PYMNTS’s reporting on Medvi, the two-person telehealth startup that launched in September 2024 with $20,000 and posted $401 million in first-year sales across 250,000 customers at a 16.2% net margin. That’s the shape of the alignment premium.

It’s also why model-agnostic operating layers like LemonLime, pitched as a “company brain” that consolidates workflows across models, are quietly attractive to the 29%: the bottleneck the Pax8 data exposes isn’t model access, it’s governance and coherence. Even 49% of SMBs slowing their tech spend still concede that without AI they won’t be competitive within three years. The pilot phase has become the risk.

Sources