Find the people the project depends on
A stakeholder may influence a decision, provide something the project needs or be materially affected by the outcome. Start with the people you know, then ask:
- Who can approve, delay or stop the project?
- Who owns the budget or key resources?
- Who knows something the project needs to understand?
- Who will do the work?
- Who will have to use, support or live with the outcome?
- Who controls an important dependency?
Check the list with people who see different parts of the work. A sponsor may know the decision-makers but miss the service team that will inherit the result. Users may point you to an experienced colleague whose opinion carries more weight than their job title suggests.
You can record a group where its members need similar involvement. Split it when views or responsibilities differ. “Operations” is too broad if the service owner can accept the handover but frontline staff are worried about an unworkable process.
Use power and interest to focus attention
A Power/Interest grid, also called a power interest matrix, is a simple way to map stakeholders. Power means their ability to affect this project. It includes formal authority, control of a scarce resource and practical influence over others. Interest means how much the work matters to them or needs their attention now. A person arguing against the proposal may have very high interest.
The workbook uses High and Low, with Not assessed where you need to investigate. Write the reason for a judgement. “Can refuse operational acceptance” is more useful than giving someone a high rating because they are senior.
High power · Low interest
Keep satisfied
Agree which decisions need them and when to raise a concern. Avoid burying the request in routine reporting.High power · High interest
Work closely
Involve them in the decisions they affect. Understand their concerns and follow up on commitments.Low power · Low interest
Keep under review
Provide a way to hear about changes and raise questions. Reassess when the work starts to affect them.Low power · High interest
Involve and listen
Give people affected by the work a useful way to contribute, and explain what happened to their feedback.Keep impact separate: what changes for this person if the project goes ahead? A team may have little authority but face a substantial change to its work. That is a reason to involve it even if the grid suggests limited attention. Record impact and needs in the register rather than trying to add a third numerical axis.
The grid will not tell you whether someone supports the project, what their objection is or what to do about it. Those questions need a conversation and an engagement assessment.
Record their position and the involvement you need
For each relevant stakeholder, distinguish the current position from the desired position. Base the current entry on what they have said or done. Then consider what involvement their role actually requires.
- Not assessed
- You do not yet have enough evidence to judge their position.
- Unaware
- They have not yet been made aware of the project or its relevance.
- Resistant
- They oppose the current proposal or are acting to prevent it. Record the objection; it may be justified.
- Cautious
- They are willing to consider the work, with questions or conditions still unresolved.
- Neutral
- They are aware, with no expressed support or objection. Silence alone is not enough evidence.
- Supportive
- They support the agreed direction and are willing to contribute within their role.
- Leading
- They actively carry a relevant part of the change, for example by resolving obstacles or committing resources.
These are working descriptions, not a score. Cautious is useful because a person asking reasonable questions needs a different response from someone opposing the proposal. An independent reviewer may appropriately remain cautious while still giving a timely decision. Describe the behaviour you need in the engagement objective; a change of label is not the objective.
Use short evidence-based notes. “Asked for a named support owner before agreeing handover” is useful. “Difficult and negative” is not. Avoid turning the register into a personality assessment.
Plan engagement around outcomes
Instead of asking how often you should communicate with someone, ask what needs to happen. Do you need a decision, advice, support, early warning, acceptance of a change or simply awareness?
That answer usually tells you whether you need a meeting, a short written update, a workshop, a one-to-one conversation or no contact at all for the moment. Give the action to the person best placed to carry it out, name the owner and agree a date that leaves time to act on the answer.
Record the feedback or result after the contact. Did you obtain a decision, change the plan, discover a new concern or agree that something remains unresolved?
Stakeholder engagement includes listening, negotiation and participation in decisions. Communication planning organises the information that supports that work. Once the objective is clear, use the communication-planning guide to agree the audience, format, timing and sender.
When support is missing or objectives conflict
Find out what the disagreement is about before trying to persuade anyone. Someone may question the evidence, lack time to contribute, disagree with the outcome or carry a cost the project has overlooked. These need different responses. A well-founded objection may require a change to the project.
If a sponsor wants an earlier release and operations needs more time to prepare, put the actual choice in front of them: what could be released, what remains unsupported and what each option means for cost, timing and service. Establish who has authority to decide, record the decision and any conditions, then check that affected people understand what was agreed.
For an absent sponsor, make the request specific. State the decision needed, when it is needed and what will happen if it is delayed. Agree a delegate or escalation route if they cannot give it attention. More status reporting rarely resolves an ownership gap.
Worked example: introducing an AI coding assistant
Entirely fictional. A medium-sized wholesale business is testing a coding assistant with two development teams on an internal ordering service. The eight-week trial will examine accepted changes, review effort and rework before any wider rollout. There is no approved headcount saving.
The sponsor expects faster delivery, developers want clarity about responsibility for generated code, and operations will inherit the releases. The same project looks different from each position.
Developer representatives
High power, high interest. Cautious → Supportive. They have no budget authority, but their knowledge and collective participation can determine whether the trial works. They care about accountability and how performance data will be used.
The engineering lead should work through those questions with them before the trial. The objective is agreement on a fair test they can challenge, with unanswered questions recorded.
Information security reviewer
High power, low interest. Cautious → Cautious. This reviewer can withhold approval but needs limited day-to-day involvement while the trial stays within its agreed boundaries.
The technical lead supplies the proposed data flow and supplier settings before source code is submitted. The desired outcome is an independent decision with clear conditions. Enthusiasm for the tool is unnecessary.
Order-processing users
Low power, high interest. Neutral → Supportive. These users cannot approve the trial, but order errors would disrupt their work. Their impact would be missed if attention followed seniority alone.
The product owner asks two users to choose representative order journeys and review the results. They need evidence that the service still works for them, along with a route for reporting faults.
The workbook develops seven stakeholder entries and shows how an assessment changes after feedback. The ratings belong to this fictional situation; they are not default ratings for those jobs.
For an AI-enabled project, use the existing AI project governance checklist to establish who must answer questions about use, evidence and human responsibility. The stakeholder plan records how you will involve those people.
Revisit the map when the work changes
Stakeholders change as the project moves along. Review the map when the project enters a new phase, before an approval or handover, after an important engagement action, or when someone changes role or raises a new concern.
Warning signs include a decision repeatedly returning with new objections, users hearing about a change through rumour, the same concern appearing in several meetings without an answer, and a support team discovering commitments it never accepted. Check the relationship, the evidence and the plan rather than simply increasing contact frequency.
Keep the record proportionate. For a small project, a short list of people, needs, owners and next actions may be enough. Use the full workbook where the map or engagement comparison adds something.
Using AI to help with stakeholder notes
An approved AI service may help organise permitted workshop notes, draft a communication or list questions that still need an answer. Check the output against the source material. A plausible summary can omit a condition or invent agreement.
Do not ask it to infer motives or assign engagement labels from messages. Decide what information may be used under your organisation's rules, and keep the assessment and resulting decisions with the people responsible for the project.
Next practical step
Turn stakeholder needs into a communication plan
Further reading
For background on stakeholder mapping and engagement, see APM's stakeholder engagement guidance, the ONS stakeholder-mapping guidance and PMI's practitioner paper on stakeholder management.
For the AI example and note-handling section, see CIPD's discussion of responsible AI adoption and the NIST Generative AI Profile.
Use your organisation's governance, confidentiality and communication requirements where they apply.