There is no single maturity level for your bid practice. Maturity belongs to each task.
Your team may be ready for AI to extract requirements while a qualification decision stays fully human. That is healthy. Different tasks carry different evidence, judgement and risk.
The useful question is not, “How mature are we?” It is, “How much of this task can we safely delegate, and what must remain a human decision?”
Use this map to make that choice visible. It connects two things that are often confused: the maturity of the practice around the work, and the level of authority given to AI within a particular task.
Raise delegation only when the workflow has earned it.
Move from prompts to measurable outcomes.
The five stages describe increasing structure around the work. You do not need every task to reach the final stage. The aim is to build only as much infrastructure as a valuable use case requires.
- Stage 1
Ad hoc AI
Individual promptsPeople use general tools independently. Value is personal and difficult to repeat, assure or measure.
- Stage 2
Structured AI
Approved contextTeams share instructions, trusted sources and expected outputs for a defined task.
- Stage 3
Specialist agents
Defined workflowsBounded agents use authorised context and tools, with clear handovers to people.
- Stage 4
Governed delivery
Human decision gatesEvaluation, traceability and named approvals govern work before it advances.
- Stage 5
Outcome system
Measured learningConnected workflows improve from evidence about quality, capacity and commercial outcomes.
Set the human role task by task.
Practice maturity tells you whether the foundations are strong enough. Delegation level tells you what AI may actually do. The scale below separates assistance from execution and policy-bounded automation.
Not delegated
The task stays fully human.AI is not applied because judgement, sensitivity or weak controls make delegation unhelpful.
AI assists, human leads
A person performs and owns the work.AI retrieves, compares, drafts or suggests. The person shapes the output and remains responsible for it.
AI executes, human reviews
AI prepares the defined output.A person checks the evidence, resolves exceptions and approves the work before it moves on.
AI executes within policy
The workflow runs inside agreed controls.People retain the decision gate while routine steps proceed within explicit policy and thresholds.
L3 is not human-free. The workflow runs within policy and stops at a named human decision gate.
See what teams can begin testing now.
The matrix applies the delegation scale across the bid lifecycle. It is a starting hypothesis, not permission to automate. Your target depends on context quality, controls, consequences and decision rights.
Scroll horizontally to explore the lifecycle.
| Capture | Qualification | Strategy & planning | Proposal management | Quality assurance | Handover | Lessons |
|---|---|---|---|---|---|---|
| Opportunity signals | Structured Go / No-Go | Requirements extraction | Storyboard / outline | Compliance review | Commitment register | Debrief capture |
| Client / market intelligence | Evidence-based PWin | Win strategy | SME work packages | Pink review | Mobilisation pack | Outcome classification |
| Relationship mapping | Evidence completeness | Value proposition | Draft assembly | Red review | Risk / assumption transfer | Pattern detection |
| Competitor monitoring | Strategic-fit challenge | Solution design challenge | Narrative consistency | Cross-model challenge | Evidence / contract index | Lessons register update |
| Pipeline synthesis | Capacity challenge | Bid plan / work packages | Traceability checks | Residual-risk summary | Owner / action allocation | PWin recalibration |
| Precedent retrieval | Recommendation pack | Risk / opportunity register | Formatting compliance | Final readiness pack | Delivery acceptance | Improvement decisions |
The colour is not the conclusion. Validate each task against your evidence, policy, risk and accountable decision owner before changing its delegation level.
Choose one credible next step.
Start with one recurring point of delay, rework, weak evidence or lost capacity. Name its current delegation level, choose a realistic next level and identify the controls needed to move safely.
Name the pressure
Choose a recurring point of delay, rework, weak evidence or lost capacity.
Set the levels
Define what AI does today, what it should do next and who remains accountable.
Build the controls
Strengthen context, evaluation, traceability and decision gates before raising delegation.
Turn one use case into a governed pilot.
The companion guide helps you define the outcome, context, agent role, tests and human decision gates.
Explore the practical guide →
About the author
Yorik Tisseau (MBA) is the founder of Ignis Leadership. He brings more than 15 years of international experience leading complex bids and building stronger bid practices, from estimator to Bid Director.
Today, he helps organisations build and transform AI-augmented bid practices, connecting bid leadership with the workflows, controls and adoption needed to make human-agent collaboration work.
