A practical diagnostic for redesigning complex bids around better judgement, greater capacity and clear human accountability.
Agents prepare and coordinate. People direct, challenge, decide and approve.
Choose one workflow worth testing. Define what the agent may prepare, what a person must decide and what evidence is required before the work advances.
A bid is not a document-production exercise. It is a sequence of commercial, technical and delivery decisions, made under time pressure and with imperfect information.
The useful question is not simply where AI could save time. It is where a bounded agent contribution could improve the quality or timing of a decision while a named person remains accountable.
Use a live, recent or representative pursuit. Avoid rating the organisation in the abstract.
There is no single maturity score. The pattern across these six areas matters more than the total. Mark each statement Visible, Partial or Unclear, then use the prompt to challenge the workflow.
Challenge promptWhich consequential decision should this work make easier to reach?
Challenge promptCould a reviewer trace every important output to an authorised source?
Challenge promptIs the agent's job clear enough without saying “help with the bid”?
Challenge promptHow will you know it is wrong before it affects the pursuit?
Challenge promptWho can say yes, no or not yet?
Challenge promptWhat could this bid teach the system that the next team can use safely?
Your strongest first use case sits where three conditions intersect. A gap in one of them is not failure. It tells you what to resolve before you spend more.
Turn the opportunity into one sentence before discussing tools or platforms.
Does this happen often enough to learn from?
Would improvement matter to growth, risk or capacity?
Are the inputs available, authorised and current?
Can a knowledgeable person judge success?
Can it be tested safely before live reliance?
Complete these six fields with the people who own the work, the evidence and the decision. Keep the first test deliberately narrow.
Name the user, the task or decision, the current friction and its practical consequence.
State the preparation or coordination work it may perform, and what it must not do.
List the authorised sources, owners, access conditions and rules for stale or conflicting information.
Name the person who reviews the evidence, resolves exceptions and decides whether the work advances.
Set the baseline, checks, tolerances, known failure modes and escalation route.
Choose controlled material, reviewers and measures, then define the decision to stop, refine or expand.
The free PDF combines the operating-model poster, the 12-question assessment and the 30-day test canvas in a practical format for a leadership or bid-team working session.
Poster · Diagnostic · First-use-case template · 30-day test canvas
Bring one workflow and the areas you marked Partial or Unclear. We can identify the operating-model gap, the first credible test and the decisions that must remain human.
