R

Activate · Project walkthrough

One project walked through
the Activation Map

The Corridor shows where knowledge sits. The Activation Map shows how to move a specific AI investment from where it sits today to where it needs to be by go-live. One initiative, four operational steps.

The project

Will the contract-review copilot deliver?

A legal practice is scoping a contract-review copilot.

The organisation wants to deploy an AI copilot into contract review. When a lawyer opens a contract, the copilot brings up relevant precedent, flags unusual deviations from the organisation’s standard positions, and suggests redlines based on how this organisation has negotiated the same issues before.

The ambition is clear. The business case is approved. Less clear: whether the initiative as currently scoped will deliver, or end up as an advisory sidecar nobody uses. The Activation Map answers that question.

Step 1 · Scope

Where does this initiative sit today?

Place the project on the matrix. The X-axis is how close the AI is to the real knowledge that drives contract-review decisions. The Y-axis is the value the initiative would produce if that knowledge were available.

Outcome value

High value · far from knowledge

⚠ Risk

Context Debt

Where this initiative ends up if the ambitious deployment goes live without L3a/L3b grounding. High-stakes calls on general-purpose knowledge alone.

High value · close to knowledge

◎ Target

L3 Advantage

Codified knowledge plus tacit judgement, made available to a reasoning agent.

L3bMajor gap

Partners’ unwritten positions

Low value · far from knowledge

Current

L1 Commodity

L1 + L2

Model-native productivity. Where the AI’s capability sits today.

L2Partial

Commercial-law conventions

L1Achieved

Standard clauses

Low value · close to knowledge

L3a Foundation

Retrieval-grounded work on documented knowledge. The stepping stone toward L3 Advantage.

L3aGap

Organisation’s precedent library

Far
Proximity to knowledge
Close

Currently in L1 Commodity. Risk: ending up in Context Debt by deploying ambition without grounding. Target: L3 Advantage, via L3a Foundation as the stepping stone.

Current position · L1 Commodity (bottom-left)

Where the AI’s capability sits today. L1 (standard clauses) is achieved through pre-training. L2 (commercial-law conventions) is partial. L3a (the organisation’s precedent library) is retrievable but not integrated. L3b (the partners’ unwritten positions) is uncodified. Without that grounding, the ambitious deployment would land in Context Debt: high-stakes calls on general-purpose knowledge alone.

Step 2 · Target

Where does it need to sit by go-live?

Some initiatives genuinely belong in bottom-left (L1 Commodity), and trying to move them to top-right is scope creep. Others must land in top-right or they’ll go into production with visible value and unreliable outputs.

Target position · L3 Advantage (top-right)

High value, close to knowledge. For this initiative to deliver its promised value, the AI needs both the organisation’s codified precedent (L3a) and the partners’ tacit positions (L3b, at least the writeable subset). L3 Advantage is the only quadrant that combines both. L1 Commodity (bottom-left) gives generic outputs that don’t reflect this practice’s positions or judgement. L3a Foundation (bottom-right) is the stepping stone. Retrieval grounds the documented layer, but tacit reasoning isn’t there yet.

Step 3 · Plan

Where are the activation gaps?

Moving an initiative from Context Debt to L3 Advantage means crossing specific rungs of the Activation Ladder. For this project, the arrow crosses all four. The biggest gaps are in the middle.

Rung 1
Access
No companion AI tool deployed that can supplement the work and access the relevant context
Connect an LLM to the organisation’s iManage corpus and SharePoint precedent library. Partly solved: access exists but is thin.
Gap
Rung 2
Embedding
A tool exists, but drawing on it depends on the individual remembering to use it
Current plan is a sidecar web app. Needs to be a Word / iManage add-in so the AI sits inside the lawyer’s document, not in a separate tab.
Major gap
Rung 3
Workflow
The AI isn’t built into how decisions get made, so it’s not able to add value within the business process
The review workflow invites the AI after the partner has redlined. Redesign so the AI is prompted at the open-contract moment, not after.
Major gap
Rung 4
Trust
The AI is embedded and available, but people don’t act on what it suggests
Lawyers won’t adopt suggestions they can’t trace back to precedent. Add citation links so every suggestion is traceable. Verification architecture is mandatory, not optional.
Gap

The critical move: rungs 2 and 3 are the major gaps, and they dominate the arrow. The current scope solves rung 1 (the integration plumbing) but leaves the AI as a sidecar invited in after the lawyer has redlined. No amount of model improvement on rung 4 closes rungs 2 and 3. The activation plan is primarily about workflow and where the AI sits, not about model selection. This is the scoping error the Activation Map catches before the investment is committed.

Step 4 · Monitor

How do we know we’re moving in the right direction?

Active monitoring turns the Activation Map from a one-off diagnosis into a running dashboard. Without it, simple initiatives can quietly drift into Context Debt.

Quarterly checkpoints

  • Position on the Map: has the current position moved toward the target?
  • Rung progress: of the four ladder rungs, which have been cleared, which remain as gaps?
  • Usage at the decision point: what percentage of decisions invoke the AI when they’re being made? Rung 3: workflow.
  • Citation accuracy: what percentage of AI suggestions can be traced back to sources the user verifies as relevant? Rung 4: trust.

Set the benchmark for go-live. The decision is made before launch.

What the Activation Map makes visible

Without the Activation Map, the contract-review copilot looks like a reasonable investment with a plausible business case. With it, the picture sharpens: the project starts in L1 Commodity, the biggest gaps are at rungs 2 and 3, and the risk is drifting into Context Debt rather than landing in L3 Advantage. The scoping is visible before the program starts.

One Corridor per organisation. One Activation Map per initiative. Quarterly reassessment of progress. Run this on every initiative you’re considering, and your portfolio becomes a meaningful dashboard for benefit realisation.

Continue

Decision walkthrough →

See the same kind of thinking applied to a single decision rather than a whole program. The Knowledge Corridor across one call.