An opinionated observational benchmark from Treetop's engagement work with B2B mid-market companies ($5M-$50M ARR) over the past 18 months. Where companies actually are on AI rollout, what differentiates the top quartile, and what the next 12 months likely look like. Citable data points labeled by source confidence.
What this is: An observational benchmark synthesized from Treetop's engagement work with ~60 B2B mid-market companies between $5M and $50M ARR over the past 18 months, plus continuous monitoring of vendor + industry public statements through May 2026.
What this isn't: A statistically-rigorous survey. Our sample is biased toward companies that hired a consultancy — i.e., they were trying to do this seriously. Read the numbers as directional, not definitive.
Permission to cite: Yes, with attribution to Treetop Growth Strategy (treetopgrowthstrategy.com) and a link to this page. We update this benchmark quarterly; the URL is stable.
Across our sample, B2B mid-market companies cluster into four distinct AI maturity tiers as of May 2026:
| Tier | % of sample | What characterizes them |
|---|---|---|
| Tier 1 — Exploring | ~35% | One or two people use ChatGPT or Claude personally. No company-wide tools. No policy. No measured workflows. |
| Tier 2 — Piloting | ~30% | Enterprise AI seats provisioned for one function (usually marketing or CS). 1-2 workflows in informal use. No formal owner. Limited measurement. |
| Tier 3 — Operationalizing | ~25% | 3+ production workflows. Named AI lead with protected time. Written policy. Measured impact. Cross-functional adoption underway. |
| Tier 4 — Compounding | ~10% | AI fluency is a baseline team expectation. Multiple Projects per function. Workflows documented. Internal prompt library. Top-down + bottom-up adoption signals. |
The 12-month delta: a year ago, the same sample skewed even further toward Tier 1 (~55%). The Tier 1 → Tier 2 transition is happening fastest; the Tier 3 → Tier 4 transition is the hardest leap and where we see the most companies stall.
Across the ~10% of companies in Tier 4 (Compounding), five characteristics show up consistently:
| Failure mode | % of stalled rollouts | Median time-to-recovery |
|---|---|---|
| No named owner — diffused responsibility | ~40% | 60-90 days (when finally named) |
| Tool selection by committee — no platform chosen at day 60 | ~25% | Often never resolved without intervention |
| CEO sponsorship without personal usage | ~15% | Doesn't recover absent CEO behavior change |
| Workflow planning that never ships | ~10% | 30-45 days (when forced to ship something) |
| Team resistance from miscommunication | ~10% | 45-60 days (when properly diagnosed) |
Year-1 all-in AI spend (platform + implementation + training + internal time) by company size in our sample:
| Company size | Median Year-1 spend | Range (25th-75th percentile) |
|---|---|---|
| 10-25 people | \$11,000 | \$5K-\$22K |
| 25-50 people | \$28,000 | \$15K-\$55K |
| 50-100 people | \$68,000 | \$35K-\$130K |
| 100-250 people | \$185,000 | \$95K-\$320K |
Cost surprise: In our sample, 'internal time' (the AI lead's hours plus training time across the team) consistently accounts for 40-55% of true Year-1 cost — and is the line most companies under-budget for. License costs are typically 15-25% of total; external implementation is 15-30%.
Among companies in our sample that reached Tier 3 (Operationalizing) by month 12, self-reported outcomes:
| Vertical | % reaching Tier 3+ | Notes |
|---|---|---|
| B2B SaaS | ~50% | Highest baseline AI fluency; fastest movement |
| Professional services | ~40% | Strong margin gains drive adoption |
| E-commerce / DTC | ~35% | Creative testing leverage is unique to this segment |
| Financial services | ~25% | Compliance overhead slows but doesn't stop adoption |
| Healthcare | ~20% | HIPAA navigation is the gating factor |
| Manufacturing | ~20% | Commercial-side adoption real; ops-side limited |
| Nonprofit | ~15% | Resource constraints; once a champion emerges, fast |
| Real estate | ~15% | Adoption uneven across brokerages |