A senior consultant uses Claude to draft a first-pass analysis. Another builds a ChatGPT workflow that helps structure client interviews. A third has a prompt library that cuts their weekly reporting time in half.
Individually, it looks like progress. At the firm level, nothing has really changed.
This is Level 1. And for most professional services firms, it's exactly where they've been stuck for the past two years.
AI Organizational Maturity Model
Tribal
Knowledge lives in 2–3 people's heads. Processes undocumented, decisions person-dependent. Everything works "somehow" because it always has.
Experimenting
Individuals use AI tools, but nothing accumulates. Workflows disappear when people leave. Individual productivity grows; the firm doesn't move forward.
Legible
The firm can describe its work in a form machines understand. Processes documented, terms defined, decision thresholds explicit.
Knowledgeable
The firm knows what it knows — and can prove it. Data is trustworthy, connected, and auditable. Clear sources of truth for every key process.
Adaptive
The system acts before being asked. Proactive anomaly detection, automatic escalations, problems anticipated before they surface.
Self-improving
The organization learns from its own operations. AI improves processes through feedback loops. In 2026, still a vision — not a current reality.
What Level 1 actually looks like
At Level 1, AI adoption is personal and fragile. Smart individuals find tools that work for them. The gains are real — but they don't compound. When that consultant leaves, the workflow leaves with them. The firm hasn't learned anything.
The problem isn't motivation or tool access. It's that the firm's most valuable knowledge — the diagnostic frameworks, the ERP red flags, the decision logic that took 20 years to develop — was never written down in the first place. It lives in people's heads. And AI can't work with what it can't read.
What the L1→L2 transition actually requires
Moving to Level 2 is not about picking better tools or buying more licenses. It requires something more uncomfortable: making the organization legible to machines.
That means being able to describe, in explicit terms, how work actually gets done. Not the org chart version. The real version.
Which SAP fields does your consultant actually look at when assessing procurement health? What are the thresholds that flag a supplier as a risk? What sequence of questions does a good diagnostic follow — and why?
If the honest answer is "it depends on who you ask," you're at Level 1. The expertise is real. It's just locked inside individuals in a format no tool can use.
Why this is harder than it sounds
There's a reason so much institutional knowledge never gets documented. When a process lives in someone's head, that person is irreplaceable. When expertise isn't explicit, it can't be questioned. Making work legible means making it auditable — and that's a form of vulnerability that many organizations quietly resist.
This isn't a technology problem. It's an organizational one. The firms that navigate this transition aren't necessarily the ones with the best AI tools. They're the ones willing to ask: what do we actually know, how do we know it, and can we write it down in a way that holds up?
A concrete example
Consider a diagnostic engagement in manufacturing. An experienced consultant can walk into an SAP environment and within a few hours identify the three signals that indicate a planning process is broken — stock coverage patterns, order confirmation lags, a specific set of master data gaps. That knowledge is genuinely valuable.
But if another consultant in the same firm — or an AI tool — can't access that logic, the firm hasn't created value. It has created a dependency.
The L1→L2 work, in this case, is straightforward to describe and genuinely hard to do: sit down with that consultant, extract the diagnostic framework, define the fields and thresholds and decision sequence, and write it down in a form that can be shared, trained on, and used. The first-pass diagnostic that currently takes two days can eventually run in hours. But not until the logic is out of someone's head and into a structure a tool can work with.
What this means in practice
The output of Level 2 doesn't look impressive. It's documentation. Process maps. Defined terms. Structured templates. Nobody makes a keynote out of it.
But without it, everything above — agents, automation, AI-assisted delivery — collapses. The firms that skip this step and jump straight to AI tooling are the ones who end up with demos that work and deployments that don't.
The good news: everything you do to make your organization legible for machines also makes it better for people. Onboarding gets faster. Delivery becomes less dependent on who's staffed. The firm gets more resilient. This isn't overhead on the way to AI. It's organizational hygiene that AI finally gives you a reason to invest in.
Where is your firm on this maturity curve right now — and what would it take to move to Level 2? Let's talk about it.