Most maturity models are written for the Fortune 500. They’re not useful for a $20M service business. Here’s the version that actually matches how mid-market companies move through AI — and how to know what stage you’re in.
Why the standard maturity models don’t fit
Walk into any consulting deck and you’ll find a five- or seven-stage AI maturity model with terms like “AI-augmented enterprise” and “autonomous decision-making at scale.” Useful if you’re a global bank. Useless if you’re a $20M industrial services firm in East Texas.
The mid-market arc is different. It’s shorter, it’s more workflow-centric, and the milestones look more like operating-business milestones than IT-program milestones. After watching roughly 100 of these engagements, the three-stage version below is the one that actually maps to how owner-led service businesses move.
Stage 1: Individual productivity
What it looks like: People are using AI tools — Claude, ChatGPT, Copilot — on their own. Marketing is drafting copy faster. The sales team is writing better emails. Someone in operations figured out how to summarize meeting notes. There’s no policy and no measurement, but there’s energy.
What’s working: Adoption is happening organically. People are getting more comfortable with the tools. You’re building intuition about what AI is good at.
What’s not working: None of it shows up in your P&L. The wins are individual, not operational. You can’t tell a board “AI saved us $X” because it didn’t — it saved 30 minutes here and 45 minutes there across 20 people. This is the trap MIT’s NANDA initiative documented in its 2025 GenAI Divide study: generic tools like ChatGPT shine for individuals because they’re flexible, but they stall in the enterprise because they don’t learn from or adapt to your workflows — and only about 5% of AI pilots ever produce measurable P&L impact.
How to know you’re here: Your largest documented AI win is “someone uses ChatGPT for emails.” Your AI spend is in unmanaged tool licenses. There’s no AI on your roadmap because there is no AI roadmap.
Time in stage: Most companies sit here for 6 to 18 months. Some never leave.
How to move on: Pick one workflow. Get senior judgment on which workflow. Build a real solution inside that workflow with a measured baseline. Ship in 30 days.
Stage 2: Operational workflows
What it looks like: One to five specific workflows have AI embedded in them. Quoting is faster. Field reports are structured. Customer service handles Tier 1 routinely. Each workflow has a measured KPI and a documented runbook. There’s a clear AI roadmap for the next 12 months.
What’s working: Real ROI is documented. The CFO can defend the AI spend. The team trusts the systems. New use cases get scoped through a known process rather than as one-off heroics.
What’s not working: AI is still bolted onto existing workflows rather than reshaping them. The senior team still operates the way they did before AI; AI is a tool their teams use, not a way of thinking. Scaling beyond five workflows starts to feel like running uphill. This is exactly where most companies get stuck: McKinsey’s 2025 State of AI found 88% of organizations now use AI in at least one function, but only 39% report any enterprise-level EBIT impact — and the single strongest correlate of the firms that do capture value is that they redesign workflows rather than bolt AI onto the old ones.
How to know you’re here: You can name three or more AI workflows in production with measurable impact. You have a documented roadmap. AI is line-itemed in the budget. Someone owns it — usually an operations leader.
Time in stage: 12 to 24 months for most companies.
How to move on: Start operating with AI rather than just using AI. The senior team — CEO, COO, CFO — pulls AI into their own workflows. The architecture of the business starts shifting around AI capability rather than the other way around.
Stage 3: AI-native operations
What it looks like: AI is a baseline assumption in how work happens. New workflows are designed AI-first. The org chart has roles that didn’t exist in Stage 2 — an AI operations lead, a prompt librarian, a workflow architect. The CEO uses AI multiple times a day for actual decisions, not just tasks. The competitive advantage is no longer “we have AI” — it’s “we operate differently than companies that don’t.”
What’s working: Output per FTE is materially higher than peers. Senior people are more leveraged. You can take on work that competitors can’t because your unit economics are different. Talent recognizes you as a place where the work is more interesting.
What’s not working: Governance and risk management have to catch up. Customer transparency about what AI does in your delivery becomes a real question. The pace of change inside the org outstrips the pace of change in your customers’ organizations, and you have to manage that gap.
How to know you’re here: Most mid-market service businesses won’t reach this stage until 2028 or 2029. The pioneers will get there in 2027. You’ll know because operating without AI feels as strange as operating without email.
Why stage matters more than tactic
The most common mistake mid-market operators make is buying Stage 3 tactics while they’re operationally at Stage 1. Hiring an “AI Chief of Staff” before you have a single AI workflow in production is Stage 3 thinking applied to a Stage 1 reality. It doesn’t work. The Chief of Staff has nothing to operate, no roadmap to execute, no team to coordinate.
The reverse mistake — staying at Stage 1 because Stage 2 feels too ambitious — is more common and more costly. The companies that pull ahead in 2027 are the ones who got operationally serious about Stage 2 in 2026.
What we recommend by stage
Stage 1: Start with an AI Office Sherpa tier ($2,500/mo) or a Workshop ($4,500). Identify the one workflow that matters most. Ship a first build in 30 days. Build the muscle.
Stage 2: Move to Operator ($5,000/mo) or Embedded ($10,000/mo). Run a Value Sprint or two against the workflows that need the most depth. Get to three to five workflows in production within 12 months.
Stage 3: This is multi-quarter program territory — an Embedded AI Office ($10,000/mo) running an active build queue. By the time you’re here, the question isn’t whether to invest in AI — it’s how to design the next chapter of the business around the leverage you’ve built.
Where are you?
If you can’t tell which stage you’re in, that’s diagnostic information by itself. Most owners we talk to think they’re at Stage 2 because their team uses ChatGPT. They’re actually at Stage 1. The gap between “individuals use AI” and “the business has AI workflows in production” is the gap that matters.
A 30-minute call usually answers this clearly. We can tell you straight what stage you’re at, what the next stage costs, and what it takes to get there.