Designing Actions in Coaching: Framework and Templates

Designing actions in coaching means partnering with the client to choose a step that serves their purpose, fits their circumstances and can be reviewed honestly. An action might complete a responsibility, build a skill, test an uncertainty or obtain information. Its value comes from that job, not from producing homework at the end of every session.

A useful framework preserves the client’s choice while checking authority, resources, receiving commitments and evidence. It also explains what happens if circumstances change. A revised plan can be responsible without being counted as completion of the original action.

Coachful is our #1 recommendation for connecting the chosen action with preparation, client tasks, goals and subsequent coaching review. This is an editorial workflow recommendation, not an independent market ranking or a guarantee that recording an action will produce its desired result.

Place action design within the current coaching competency

The 2025 ICF Core Competencies place this work within Facilitates Client Growth. Its emphasis includes client autonomy, integrating learning, designing goals and accountability, and considering resources, support and barriers. Action planning sits within a partnership rather than a coach prescribing the same task to everyone.

The five action types and two templates below are original practical aids. They are not an official ICF sequence, assessment or substitute for professional judgment. Use only the fields that help the client make and review an informed choice.

Choose which of five jobs the action needs to do

A commitment fulfills an appropriate responsibility

A commitment concerns something that needs to be done to an applicable standard. A required report must contain the right information and reach its recipient on time. Calling it a learning opportunity does not remove those obligations.

Clarify what the client owns and what others must decide. They may be responsible for providing technical evidence without controlling the committee’s eventual decision. Review their contribution accurately rather than judging them only by whether the committee adopted their preference.

A practice develops a specific skill

A practice gives the client an appropriate opportunity to rehearse or use a skill. Specify what they want to improve: a concise question, a clear request or a pause before replying. Choose feedback that concerns that skill.

Role-play can help prepare, but it does not establish how a real colleague will respond. If the client rehearses with the coach and never encounters the live opportunity, record completed rehearsal and untested live practice separately. Avoid making colleagues unwitting participants in an intrusive performance exercise.

A learning experiment examines a stated uncertainty

An experiment needs a prediction or question that could be informed by the proposed action. The client might predict that asking for evidence before a decision will lead the chair to dismiss their contribution. The observation needs to concern that situation, not merely whether the client felt nervous.

Agree what would count as relevant evidence and what alternative explanations remain. One friendly reply does not establish that every future challenge will be welcome. A canceled meeting provides no direct evidence about the predicted live exchange. A useful experiment can remain inconclusive.

An information request clarifies what is not known

Sometimes the responsible next action is to obtain a fact, a current version or a decision route. The client does not need another confidence exercise if they cannot yet determine which figures the proposal uses.

Ask who can provide the information and what scope they will accept. Existing aggregate records may be sufficient; a new analysis may require a separate resource decision. An unanswered request remains unanswered even if the client sent it punctually.

A support request obtains an actual contribution

A colleague’s expertise is not the same as an accepted allocation of their time. Specify the contribution requested, when it is needed and what the receiver can change or decline. Build the reply into the action design.

If support is unavailable, the next step may need a smaller scope, different date or another route. Do not allocate a colleague’s work through the client’s coaching task list. A request for help is an action the client can own; the other person’s agreement remains theirs.

Apply five checks before turning insight into a plan

1. Purpose and ownership. What goal or learning does this serve? Does the client actually choose it? If you suggest a step, name it as a suggestion and allow a real decline.
2. Authority and effects. Who may make the decision? What information can be shared, and who will be affected? Keep required responsibilities and protections intact.
3. Capacity and dependencies. Count preparation, action, receiving work and review. Confirm the contributions needed from others rather than treating them as free resources.
4. Opportunity and fallback. Identify a recognizable occasion to act and a responsible response if it disappears. Preserve the purpose while recording any change in what was actually attempted.
5. Evidence and review. Separate completion, quality, learning and external outcome. Agree when to review and which questions the evidence could reasonably answer.

For example, “speak up more” does not identify a purpose or appropriate occasion. “Provide the capacity question before the committee approves a schedule” is more usable, but still needs an accepted agenda route, checked evidence and a fallback if the item is removed.

A task can also be too small to serve its purpose. Opening a document is not equivalent to preparing a required briefing. Smaller scope helps when it preserves a meaningful contribution, not when it merely makes a completion mark easy.

Choose evidence at four different levels

Completion concerns whether the chosen step occurred. The client sent the request, rehearsed the question or attended the meeting. Define the unit before counting it. Preparing a question is completed preparation, not a completed live contribution.

Quality concerns whether the step met its relevant standard. A brief can be sent on time yet contain the wrong figures. A request can be polite yet omit what the receiver is being asked to accept. Identify the material criteria rather than making speed or volume the only test.

Learning concerns what the action allows the client to understand. The record may show that a particular recipient wanted clarification, that a skill needs practice or that a proposed test was not feasible. Preserve uncertainty where the observation cannot distinguish several explanations.

External outcome concerns a further result such as an authorized decision, an accepted agreement or a changed process. It may require other people and more time. Do not infer it from the client’s completion mark or an encouraging coach comment.

In a contrasting hypothetical, an internal equipment report is required by 16:00. Submitting it at 17:00 with accurate figures can satisfy the content requirement while missing the deadline. A useful review records both. Calling the late report an experiment would not change the original responsibility. Conversely, a well-designed learning test may be completed correctly and remain inconclusive. Choose the evidence standard that fits the actual job.

Agree what information is sufficient and who needs it. A short factual note can support a coaching review without recording a colleague’s private material. More data is useful only when it helps answer the client’s question responsibly.

Use a cue to support action without treating it as a guarantee

Gollwitzer and Sheeran’s implementation-intentions overview, hosted by the US National Cancer Institute, distinguishes a goal intention from a plan linking a recognizable situation to an intended response. It also discusses the importance of viable plans and an underlying goal the person actually intends to pursue.

The same authors’ 2025 review, Psychology of Planning, discusses if-then planning and moderators involving people, goals and plan features. That is a reason to examine fit, not evidence that a single cue resolves missing authority, time or skill. The NCI-hosted overview is not being presented as the original 2006 meta-analysis, and no pooled effect is used as a forecast for this coaching case.

An original cue might be: “When the chair invites comments on the capacity item, I will ask the prepared evidence question before approval.” The cue identifies an opportunity. It does not authorize interrupting an unrelated emergency item or force a meeting to occur.

A fallback should name a different action honestly. If the item disappears, the client can request an appropriate written route. That may meet the responsibility to contribute evidence while leaving the planned live practice and its prediction untested.

Work through a complete fictional committee case

All people, data and outcomes in this example are fictional. Asha is an archive operations lead, Pavel chairs the planning committee, Ivo maintains the operational figures and Owen is Asha’s coach. The committee is considering digitizing 160 boxes in ten operating days. This is a fictional archive project, not a Coachful product capability.

Asha has relevant technical responsibility. She often waits until a decision is over before raising concerns, then worries that colleagues see her as obstructive. In the 18 February 2039 coaching session, she chooses to contribute the capacity question before the proposed schedule is approved.

She also wants to examine a prediction: if she asks the chair for evidence during the live discussion, he will dismiss her contribution. Owen checks that she is not proposing to withhold necessary information in order to test him. She will provide the relevant facts regardless; the learning question concerns an appropriate way of contributing, not whether to fulfill her responsibility.

Prepare the evidence and receiving arrangement

The original committee item is scheduled for 21 February. Asha asks Pavel for permission to contribute a short technical question when he opens comments. He reads the request and accepts that route, with up to thirty active minutes for this contribution and its handling. She also asks Ivo to check the existing four-day box counts rather than produce a new forecast.

Ivo declines a fresh capacity study but accepts ten minutes to read the request, verify those existing records and return his answer. He confirms four full operating days of the same recorded duration with counts of 8, 10, 6 and 12 boxes. The total is 36, or an average of nine per day. The records do not establish that the mix of boxes matches the proposed project.

Asha calculates that 160 boxes in ten days requires sixteen per day. Nine is a description of those four observed days, not a validated prediction for every future day. She does not turn 160 divided by nine into an approved alternative schedule. The useful question concerns the basis for sixteen and the comparability of the work.

Asha uses twenty minutes to check the confirmed figures, prepare her short brief and complete the action card below. Her fifteen-minute request-handling allocation covers the original Pavel and Ivo messages and their replies. Ivo’s actual verification stays inside the accepted ten minutes.

Rehearse a bounded question and complete the action card

The initial forty-minute coaching session includes eight minutes of rehearsal. Asha first frames the question as “Why are we ignoring the real capacity?” Owen notes that this assumes the figures have been ignored, which she has not established. She chooses a revised question and practices saying it aloud:

The proposed 160 boxes in ten days requires sixteen per day. The four recorded days total 36, averaging nine, but their box mix may differ. What evidence or additional resources support sixteen per day for this project, and what assumptions should we record before approving the schedule?

The rehearsal is completed. It does not establish Pavel’s likely reaction. Asha’s completed original action card is:

This card is a completed plan, not evidence of completion. Its value is that the client and coach can later compare the intended opportunity, accepted route and actual events.

Use the fallback when the original opportunity disappears

On 20 February, Pavel circulates a revised agenda that removes the capacity item. No schedule approval will be taken at the 21 February meeting. Asha does not force her question into another item to protect a coaching completion score.

She asks Pavel whether he will accept the written technical question before the issue returns. Pavel reads and agrees to receive it. She sends the prepared question with its four confirmed counts and the comparability caveat. Her ten-minute fallback period includes reading the agenda, requesting and reading permission, and sending the existing brief. Pavel spends five minutes on this changed-route request and reply.

The original live practice has not occurred. The written contribution is a different completed action with an accepted receiving route. Asha can use the fallback while leaving the live experiment open rather than calling every possible outcome proof of success.

Complete the actual question, answer and acknowledgment

On 22 February, Pavel reads the brief and replies: “Are you recommending that we delay the project? I need to distinguish a capacity question from a proposed decision.” His ten-minute allocation covers reading the note and asking that clarification.

Asha answers:

I am asking for the basis of the proposed capacity before we commit. The four days are not a controlled comparison with the planned box mix, so they do not establish a replacement completion date. I am not asking you to record this note as an approved delay. Please keep the sixteen-per-day assumption visible and decide what evidence or resources are needed through the normal project process.

Pavel reads and acknowledges on 23 February: he understands it as technical input requiring clarification, not an approved delay recommendation. No project date is approved in this exchange. The next possible committee discussion on 28 February remains future at the coaching review cutoff.

Asha spends ten minutes reading Pavel’s clarification and writing her substantive answer. Pavel spends a further five reading that answer and sending the acknowledgment; Asha reads it within her same handling allocation. Pavel’s earlier five-minute original agenda-request handling is separate. His actual total is twenty-five minutes, within the thirty he agreed for this contribution and its handling.

Use the review template without changing the original result

At the 24 February twenty-minute review, Asha says she feels less certain about what “dismissal” would look like. Pavel asked a real clarifying question and acknowledged the answer. That matters, but it happened in writing. She does not treat it as a direct test of the prediction about a live committee exchange.

Owen later sends an actual private client-conversation comment: “You kept the responsibility to provide evidence while changing the route. What remains untested about your original prediction?” Asha reads and replies that Pavel’s live response is still unknown. She wants to prepare for the normal project process without treating every meeting as a personal experiment.

Her five-minute reading-and-reply allocation is separate from the live review. Owen’s ten-minute written-feedback allowance includes preparing and sending the comment and reading her reply. Asha remains somewhat apprehensive. The work ends without inventing complete confidence or an additional booked session.

Count the whole action, including its receiving work

The whole action-design case uses 260 person-minutes. The eight rehearsal minutes are already within each person’s forty-minute coaching row and are not added again. Message waiting, routine committee work unrelated to this contribution and the future 28 February discussion are outside this finite account.

Turn the chosen plan into a useful Coachful workflow

1. Use an optional action-design worksheet. The action-card fields can support a short reflection. Check the authorized audience and omit private archive or staff details that are unnecessary to the coaching purpose.
2. Record the client’s actual task. Asha can own preparing or sending her question. Do not create a completed “asked at committee” task when the agenda item disappeared, or assign new analysis to Ivo after he declined it.
3. Link the action to the right goal. Native goals can show linked tasks, habits and milestones. A milestone has its own title and context; This is the finish line marks a milestone that actually completes that goal. Sending one question should not silently close a broader leadership goal.
4. Review the changed status. Keep original live practice distinct from completed written contribution. A task completion mark is not Pavel’s approval or a validated result for Asha’s prediction.
5. Return the promised feedback. Worksheet Coach review is private to the authorized coach team. Use the agreed private client conversation or a real session for feedback delivery, with time for the client’s reading and reply. Saving an internal review does not send Owen’s comment.

The images illustrate preparation and linked goal activity, not the fictional archive brief or a native scientific experiment tool. Our goal-setting worksheet guide helps define the intended difference. The active listening guide supports reviewing actual replies, and Coachful features describes the wider workspace.

Agree accountability without turning review into surveillance

Ask how the client wants the plan brought back into conversation. An agreed review date, a private note or one chosen reminder may be enough. Repeated notifications and public completion records are separate choices, not the default price of receiving coaching.

When a step has not happened, begin with its actual status. Was it attempted, intentionally revised, blocked or forgotten? Those accounts call for different questions. A lost opportunity may require another route; a misunderstood responsibility may require clarification; a task chosen mainly to satisfy the coach may need to be withdrawn.

The client can own an adjustment without claiming the original commitment was fulfilled. Keep any effect on other people visible. If someone was waiting for an answer, the next action may include communicating the changed status rather than silently replacing the task in a private plan.

Close the review with an actual choice: continue, revise, obtain support, stop or bring a different issue. The coach’s role is to support informed ownership and useful learning, not maintain a perfect completion streak. Preserve legitimate professional and organizational responsibilities alongside that autonomy.

Troubleshoot an action that does not fit

The client agrees to please the coach. Offer a real opportunity to revise or decline. Ask how the step serves their purpose and what they would choose without the expectation of homework.

The action depends on unaccepted work. Change the plan into a request or fact check before committing to implementation. Preserve the receiver’s answer and capacity limits.

The fallback is easier but tests something else. Use it if appropriate, then label the result accurately. A written exchange can contribute evidence without answering a prediction about speaking live.

The task is done but the desired result is absent. Check quality, assumptions and external decisions. A completed question can be useful while the underlying schedule remains undecided. Avoid counting activity as proof of impact.

New circumstances create a risk. Revisit the action and use the appropriate authority or specialist route. A coaching commitment never supplies permission to ignore relevant protections or responsibilities.

Frequently asked questions

Does every session need an action plan?

No. Choose a plan when it serves the client’s purpose. Reflection, a clearer decision or no immediate task may be appropriate.

Can the coach suggest an action?

Yes, transparently and within the role. Offer the client a real choice to accept, adapt or decline rather than hiding a prescription inside a question.

Should every action become an experiment?

No. A required responsibility, rehearsal, request and uncertainty test have different standards. Keep those differences visible.

What if the planned opportunity disappears?

Review a suitable alternative and any permissions it needs. Record the original action as unattempted or changed rather than claiming the alternative produced identical evidence.

How much tracking is useful?

Enough to establish what happened and what it means for the next choice. Avoid unnecessary private data or repeated reporting that adds burden without helping the client.