Action Learning Sets in Executive Coaching: A Revans Guide

Action learning sets in executive coaching connect a leader's real work with peer questioning, responsible action and a return to examine what happened. The set is a small learning group, not a committee that takes ownership of the presenter's job.

Coachful is our #1 recommendation for organizing the agreed preparation, individual learning records and actual coaching feedback around a set. This is an editorial workflow recommendation, not a market ranking or a guarantee that participation improves business performance.

Reg Revans's contribution matters because learning and action need each other. A lively discussion without a real next step can remain speculative. A completed project without examining assumptions can leave the same unhelpful habits intact. The worked example below follows both sides through a finite two-meeting pilot.

Understand L = P + Q without turning it into a score

The University of Salford's Revans archive describes his equation as learning through programmed knowledge plus questioning. Programmed knowledge includes established information, taught methods and relevant expertise. Questioning brings inquiry to what is unclear in the present situation.

The plus sign does not create a numerical assessment. A set cannot count ten facts and ten questions to calculate twenty units of learning. Nor does the formula mean that expertise is forbidden or that every existing method should be discarded.

Consider a recurring approval delay. Knowing the documented decision process is relevant information. Asking whether an owner actually accepted responsibility may reveal a gap that repeating the process description misses. The presenter then needs to investigate, act appropriately and examine the result.

This guide's meeting sequence and questions are original design choices. They are not a reproduced Revans protocol, a licensed instrument or a claim that one format captures every school of action learning.

Distinguish inquiry, action and learning

Inquiry examines the presenter's understanding. What is known, what has been inferred and what has not yet been checked? It should allow an answer the group did not expect, rather than steer the person toward a favorite intervention.

Action changes or investigates something in the person's real setting. It can be a bounded conversation, an authorized test or checking a record with its owner. Choosing to seek necessary expertise can be responsible action; pretending to have authority is not.

Learning involves considering what the result changes in the person's understanding or practice. “I sent an email” reports an activity. “I learned that the recipient could not decide the issue, so I changed the request” connects experience with a revised approach.

Reflection is not an invitation to call every result successful. A missed commitment remains missed. The group can learn from it while addressing its practical consequence. Likewise, an improved operational figure does not by itself show that the presenter has learned how to handle a similar issue independently.

Revans's 1984 paper, Action Learning: Are We Getting There?, opens by discussing results that differ from expectations and the need to reconsider measures, methods or resources. That is a useful warning against reporting only what a sponsor wants to hear.

Use a set for a live uncertainty, not every problem

Choose an issue the member genuinely needs to work on, with a responsible route to act. “Our organization needs trust” is too broad for one presentation. “I assume late requests reflect poor follow-through; I need to check what is happening before changing the handover” gives the set something to examine.

The presenter does not need complete control of the situation. They do need an honest next step within their role, such as requesting a decision, obtaining relevant facts or negotiating a feasible test. A set should not promise to fix a condition the member cannot influence.

Some needs call for another route. A known technical calculation may need an expert answer. An urgent safety issue, formal complaint or legal deadline should not wait for the next set meeting. Use the appropriate process, with learning support alongside it where useful.

Contrast that with uncertainty about how to work across roles when information is incomplete. The set can help the presenter distinguish a question worth investigating from a conclusion they have already decided to defend.

Choose the format that fits the actual request

Use the following practical comparisons to make the service offer clear. They describe the different purposes a group might agree, rather than claim that every provider uses these labels identically.

These purposes can coexist, but name a change when it happens. If members begin deciding a shared project, they need the corresponding authority rather than calling every decision a learning question. If a presenter asks for technical advice, identify the contribution honestly instead of insisting that an answer arrive disguised as inquiry.

Keep peer support from becoming concealed advice

Advice is not made into inquiry by adding a question mark. “Have you considered escalating?” contains a suggested route. A transparent suggestion can be useful, but it should not masquerade as a neutral discovery.

Try questions that help establish the situation: “What has the other person actually agreed to?” or “What evidence would challenge your explanation?” Listen to the answer before asking the next question. Our active-listening guide supports that check.

Do not turn the set into an interrogation. A presenter can say a question is irrelevant, too private or based on a mistaken premise. Members can respectfully challenge an unexplained leap without demanding a personal disclosure as the price of participation.

Support includes leaving ownership with the member. The group can ask whether an action is feasible and who else must accept it. It should not impose a commitment simply because the agenda needs an action before the meeting ends.

Make expertise and facilitation explicit

A member who knows a material requirement should identify it rather than watch the presenter build on a false premise. State its source and limits, separate it from a preferred solution, and decide whether the issue needs the appropriate professional owner.

For example, “Our documented process requires an accepted decision owner before a request is complete” supplies relevant information. It does not prove why a particular request was delayed. The presenter still needs to check the record and the people involved.

Jonny Gifford's 2005 Institute for Employment Studies report description discusses variations in facilitation and the development of practice beyond Revans's view that sets need not rely on externally provided facilitators. Do not describe an outside coach as a universal historical requirement.

For the original pilot below, an experienced facilitator helps members notice rushed advice, uneven attention and unsupported claims. That person manages the agreed learning process, not every business decision. Review whether facilitation remains useful rather than making the group permanently dependent on it.

Agree the conditions before asking for candor

Explain the purpose, meeting commitment, opportunities to present and reporting arrangements. A short demonstration cycle is different from promising every member an equal presentation slot during a longer program. Make the actual offer clear.

Discuss hierarchy and relationships. Someone may not be able to examine a live difficulty openly in front of their line manager or a person controlling their contract. Changing membership or choosing another issue may be more appropriate than calling for bravery.

Agree proportionate confidentiality and its limits. Identify what can be reported to a sponsor, what remains with the presenter and how concerns requiring another process will be handled. Do not promise that information can never leave the room under any circumstances.

Use only authorized information. Cross-company membership can create additional commercial and competition concerns; anonymizing a name may not conceal an identifiable business situation. A learning set is not permission to exchange restricted material.

Access also belongs in the agreement. Members may need written prompts, breaks or different participation arrangements. Camera visibility and quick speech are not reliable substitutes for checking whether someone can contribute.

Design and review a complete cycle

1. Agree a live issue, the presenter's request and what information may be shared.
2. Establish essential facts, then examine assumptions with questions that leave room for surprise.
3. Let the presenter choose an action with a purpose, owner, receiving agreement and review point.
4. Carry it out in the real setting or record honestly why it did not happen.
5. Return with the result, revise understanding and examine how the set helped or hindered.

Protect time for the return. If every meeting introduces a new urgent topic, members may never examine whether their last decision was useful. Conversely, do not spend the entire meeting reporting project status without questioning what the experience means.

Assess the design with appropriate evidence. Cho and Egan's 2009 systematic review selected fifty empirical studies from 2000–2007 and categorized studies around action, learning or a balance of both. That historical review is not a current estimate of executive-coaching effectiveness or a pooled financial return.

A different 2026 method report, Such and colleagues' Evidence Learning Sets, describes an adaptation involving seven members and a facilitator across five English local authorities over four months. It explicitly changes how evidence and action enter the process and calls for further evaluation. The published version has three peer reviews approved with reservations. It should not be treated as a randomized test of a conventional executive set.

Decide what would count as learning

Agree evidence at three practical levels. First, the presenter can show a change in their own account: an assumption becomes a question, or a claimed agreement becomes something they need to verify. A polished reflection is not enough if it contradicts the actual record.

Second, inspect the action and its consequence. Was the request received, accepted, declined or changed? What work happened and what remains pending? A useful investigation might disconfirm a proposal without producing an immediate operational improvement.

Third, examine the set's process. Did a member make room for an unfamiliar explanation, acknowledge advice or stop asking a question whose premise was wrong? Other members can learn from supporting a presenter without claiming that they solved the presenter's problem.

Keep these levels separate in reporting. Attendance establishes attendance. A changed question supports a limited observation about practice. A business outcome needs its own evidence and may have several causes. Do not turn a sponsor's request for an impressive return into permission to infer one from positive feedback.

Contract the complete fictional pilot

Fictional worked example: Lena, Idris, May, Owen and Rosa are leaders in one organization. Dev facilitates a two-meeting pilot on 1 and 15 September 2081. All six agree to ninety minutes per meeting plus the preparation and work allowances recorded below.

They agree that Lena presents the live issue during this demonstration cycle. The other four members practice questioning and review their own learning, but are not promised separate presenter slots within these two meetings. Any continuing rotation requires a later agreement.

Lena's issue is four recent handovers that waited beyond an expected date. She initially thinks the receiving team lacks follow-through. The set may discuss an authorized coded summary, not confidential personnel details or the underlying operational files.

Tari owns the receiving operation and Nia is its authorized analyst. Neither belongs to the learning set. They must accept any investigation work separately. The set cannot allocate their time or change the organization's handover procedure.

The first ninety-minute meeting uses ten minutes for the agreement check, five for the cycle purpose, ten for Lena's issue, fifteen for essential facts, twenty-five for inquiry, ten for choosing an action, ten for process reflection and five for the close. It is an original agenda, not a required historical timetable.

Let questions change the initial explanation

Owen first asks, “Have you tried a firmer deadline?” Dev names it as a suggestion and asks what is still unknown. Owen replaces it with: “Which deadlines did the receiving team actually accept?”

Lena has a target date, but cannot yet show that each handover met the process's entry conditions. May identifies a relevant fact from the organization's documented procedure: a complete handover requires a named, accepted decision owner. She does not claim that this explains all four records.

Idris asks, “What would you find if your explanation about follow-through were incomplete?” Rosa asks, “Who can check the records with you without receiving a blame report?” These questions do not prove an answer. They help Lena choose an investigation that could contradict her initial account.

Lena's completed original action record is:

Obtain the receiving work and carry out the action

Lena initially requests an hour with Tari and Nia. Tari actually replies that an hour is unavailable and offers thirty minutes on 8 September if the four records are prepared. Nia agrees to twenty minutes of preparation and that meeting. Lena reads and accepts the narrower arrangement rather than describing the original hour as secured.

Tari also accepts twenty minutes to check the procedure before the meeting. Each colleague uses five minutes for reading and agreeing the request. The preparation and meeting are genuine receiving work, not a favor omitted from the learning account.

On 8 September they carry out the thirty-minute review. The completed findings are:

Two authority gaps and one timing problem do not erase the fourth case. Nor do four records establish the frequency of each cause across the organization. The exercise tests Lena's account of these handovers, not a representative performance rate.

On 11 September Lena sends Tari and Nia the completed finding note: “The four records do not support my initial explanation that all waits were receiving-team follow-through failures. A and B lacked an accepted decision owner; C entered after the weekly cutoff. D was complete and timely but the team acted one day late. I will use this coded distinction in the learning set. I have not changed the procedure or assigned you new work.” Both read it within their counted five-minute closing allocations. No further acknowledgment is inferred.

Return to the set with results, not a victory story

The second meeting keeps the full ninety minutes: ten for opening and boundaries, twenty for Lena's report, twenty-five for inquiry, ten for transparent knowledge contributions, fifteen for individual and process learning, and ten for the close.

Lena reports the actual counterproposal, review and four findings. She now sees that she had been treating an expected date as proof of a valid, accepted handover. She also retains record D's real miss rather than replacing blame with an equally broad claim that nobody did anything wrong.

The peers examine what the action does not establish. They have not tested a revised entry checklist, measured future delays or obtained authority to change service commitments. Lena chooses to prepare a proposal for an ownership check through the normal decision route. That remains future work, not a result of the completed pilot.

During the fifteen-minute learning section, Idris notes his own tendency to accept a presenter's causal story too quickly. May distinguishes offering a sourced requirement from diagnosing a record she has not seen. Owen notices how his first question contained an escalation preference. Rosa retains the practice of asking who will receive an action's work. These are actual brief contributions inside the meeting, not additional uncounted surveys.

The group also identifies a process limit: Dev had to interrupt advice early. They agree before the meeting ends that the finite pilot is complete, with no new meetings or presenter rotation yet authorized. A useful first cycle does not establish a sustainable long-term set.

Keep the learning record distinct from operational evidence

In Coachful, use an original worksheet for the presenter's issue, assumption, chosen action and return. Our goal-setting worksheet guide can support a clear next step. A completed coaching task is not evidence that an operational owner has approved a procedure.

Lena submits the authorized coded account through assigned Forms & worksheets. Underlying handover records remain in the organization's approved system. The peers do not receive access to her individual response simply because they attended the set.

Dev's internal Coach review is private to the authorized coaching team. Dev separately sends Lena this individual client-conversation comment: “Your return distinguishes two missing ownership conditions, a cutoff issue and a genuine late action. Keep all four findings visible. The next proposal needs its own decision; the learning set has not approved a new handover process.” Lena reads the comment. That is the actual feedback delivery, not a claim that private internal review automatically reaches the client.

Explain sharing accurately. Unshared journal pages differ from pages the client explicitly shares through supported specialty workflows. For any recap workflow, check audiences and approval of new unsent material; changing a setting does not recall an already delivered recap. Explore Coachful features for the agreed individual preparation and follow-through.

Count the complete learning-and-action effort

The six set participants spend 1,080 person-minutes in the two meetings. All listed work totals 1,465 of 1,545 minutes, leaving eighty across separate allowances. The thirty-minute operational review consumes ninety person-minutes across Lena, Tari and Nia, already included in the table.

Lena's request-handling time includes reading and accepting the counterproposal. Her finding-note time includes checking the four classifications and sending the completed note. The peers' learning contributions and the pilot close happen inside the second set meeting; no separate reply work is invented. These are finite pilot costs, not a financial return or the full cost of the organization's operations.

Troubleshoot the set before extending it

• Questions repeatedly conceal advice. Name the suggestion, check what is unknown and ask whether the presenter wants that contribution.
• A member lacks authority or receiving capacity. Make obtaining an agreement the next action. Do not allocate colleagues' work from inside the set.
• Members act but never review. Protect return time and examine what the result changes, including misses and rejected assumptions.
• The group reflects without acting. Check whether the issue permits a responsible step or needs a different development format.
• One successful cycle prompts automatic continuation. Review purpose, presentation opportunities, effort and confidentiality before offering a longer program.

Frequently asked questions

What does L = P + Q mean?

Learning combines established knowledge and questioning. It is a conceptual expression associated with Revans, not a numerical learning score.

Must the group refuse all advice?

No. Make relevant information and suggestions transparent, identify their limits and keep responsibility with the presenter. Do not withhold a material requirement for the sake of a questioning rule.

Is an outside facilitator always required?

No universal requirement follows from the historical approach. Agree facilitation appropriate to the group and review whether it supports independent learning.

How many people and meetings are necessary?

Design for meaningful attention, action and a return. The five-member, two-meeting pilot here demonstrates one finite cycle, not a universal minimum or a promise of equal presenter slots.

What if the action does not improve the business result?

Record what happened, address consequences and examine the explanation. Learning may occur without the desired result, but that does not justify calling every intervention effective.