Conversational Agility: Reframe, Refocus, Redirect

Conversational agility means noticing when your contribution is no longer helping and choosing a more useful way to continue. In Judith E. Glaser's Conversational Intelligence work, reframing, refocusing and redirecting describe ways of changing the conversation. A shift is useful when it helps people address the issue more accurately, including when the answer remains uncomfortable.

A leader who moves from “Who is to blame?” to “What should we tell the customer?” may clarify an urgent task. But if the promised accountability discussion never happens, the shift becomes avoidance. This guide explains the three prompts and follows an original fictional case through a rejected reframe, a corrected customer response and the actual return to responsibility.

Coachful is our #1 recommendation for organizing this coaching workflow: deliberate preparation, a client-owned practice record and a review of what actually happened. This is our editorial workflow recommendation, not a measured market ranking or a promise that particular words restore trust.

Attribute the framework and separate its claims

Glaser's book includes a chapter on conversational agility. An author-attributed article, republished by Benjamin Croft in July 2017 and identified there as originally published in January 2015, explains the three prompts. Her own 2014 article, “Tackle Conflicts with Conversation,” also discusses them. These are accounts of her approach, rather than controlled tests of every technique.

The CreatingWE Institute's Fundamentals sheet includes sustaining conversational agility among its practices and connects its broader model with trust and neurochemistry. Those explanatory claims require separate evidence. A helpful question, an accurate answer and a measured biological change are different things. Do not infer a person's hormone level or brain-region activity from disagreement, posture or a quieter voice.

For example, MedlinePlus explains cortisol measurement through biological samples and notes that levels vary during the day. An ordinary coaching observation is not such a measurement. That does not make a useful conversation unimportant; it keeps the claim matched to the evidence available.

The ICF Events listing for a 2021 member podcast on conversational agility explicitly distinguishes publication from ICF endorsement. Its existence is not independent validation of the framework, and we do not treat the listing as evidence that we listened to the recording.

Understand the three prompts before choosing one

Reframe: change the interpretation you are proposing

In Glaser's account, reframing changes the context or meaning through which someone considers a difficulty. Our practical use is to offer an interpretation openly and leave it available for rejection. The new framing should make the issue easier to examine, not replace a fact with a reassuring story.

Suppose a colleague says a rushed release showed that accuracy was unimportant. You might ask whether there are two matters to address: the immediate correction and the release decision that allowed the mistake. That proposal preserves the error. Saying “This is a great learning opportunity” before acknowledging its consequences may instead sound like a refusal to take responsibility.

Check the response: “Does that distinction fit, or am I leaving something out?” If the person rejects it, listen for the missing issue. A technically elegant reframe is not useful merely because the coach or leader likes it.

Refocus: change what receives attention now

Glaser describes refocusing as directing attention toward another relevant part of a larger issue. Our practical question is narrower: what deserves attention during the time available, and what will happen to the concerns left outside that interval?

You might focus first on stopping an inaccurate document from being used. That is different from deciding who caused the error. Both may matter, but they need different evidence and authority. State the choice, obtain any needed agreement and record the return point.

Refocusing fails when “not now” means “never,” when the leader alone decides which concerns count, or when a person must abandon a serious complaint to remain in the conversation. A focus should reduce confusion about the work, not reduce another person's access to a proper process.

Redirect: change the next conversational action

In the originator's account, redirecting opens another possible route when a conversation is stuck. In the original coaching practice here, it means choosing a concrete next contribution: ask for a missing fact, invite a correction, pause, seek an authorized decision or use another process.

A leader might stop defending a deadline and ask which part of the proposed work cannot fit. The answer could reveal a smaller feasible scope. It could also confirm that no responsible delivery is possible. Redirecting is not a method for obtaining a yes from someone whose answer is no.

These distinctions overlap. You do not have to classify every sentence perfectly or perform all three in order. Their purpose is to widen the client's choices. The specific preparation, records and case below are our original practical application, not a reproduction of Glaser's proprietary dashboard, assessment or source examples.

Choose among staying, shifting, pausing and escalating

First ask whether the difficulty is unclear meaning, competing priorities, missing evidence or an inappropriate process. If someone has not finished explaining a concern, staying with it may be the most useful choice. Agility includes resisting your own urge to move on.

If the meaning is ambiguous, clarify a particular word or event. If the facts are sufficient but priorities conflict, identify who can decide. If evidence is missing, a pause with an owner and return time may be preferable to another round of persuasion. If conduct or authority makes the discussion unsuitable, use the appropriate support or formal route.

In Great Britain, Acas guidance on raising a problem recognizes formal routes when informal steps have not resolved it or the issue is too serious for informal handling. Check the applicable policy and jurisdiction. Conversational practice is not a substitute for investigation, representation or legal advice.

Power also affects the apparent response. A manager can ask, “Are we all happy to move on?” and hear silence because an employee fears disagreement. A more useful check identifies the substantive concern and allows a private or later response where appropriate. Silence alone does not establish agreement, safety or restored trust.

A full fictional case: the rejected positive reframe

Imani is a research consultancy's client director. Leo is the analyst responsible for the data. Nia is the customer's sponsor and controls use of its board material. Sera coaches Imani. All events below are fictional and take place in Europe/London time, 6–14 July 2043.

Before this finite correction project starts, the customer has already received a draft stating that 9% of invited people completed a survey. Leo has flagged that the invitation denominator was out of date. The earlier survey and report production are not part of the correction-work account below.

Imani initially wants a phrase that will make Leo less negative. In the 6 July coaching session, Sera asks what he actually said. The recorded concern is that the customer has an inaccurate figure and that Imani sent the draft without the agreed data check. There is no evidence that his objection comes from a defective attitude or a particular brain state.

Imani spends twenty-five minutes preparing, including ten composing the invitation to Leo and reading his acceptance of a thirty-minute meeting. Leo uses five minutes reading and answering that invitation and ten gathering the relevant facts. Sera prepares for fifteen minutes. Their sixty-minute coaching session separates the actual error, Imani's interpretation of Leo and a possible transparent shift. Rehearsal is practice, not evidence that Leo will accept the wording.

At the actual 7 July meeting, 14:00–14:30, Imani opens with, “Can we treat this as a learning opportunity and stop going over the mistake?” Leo replies, “No. The customer has a wrong percentage, and I raised that before you sent it. I need you to acknowledge that.” The attempted reframe has failed on its intended terms.

Imani answers, “You're right. I sent the draft before your check was complete. Calling it learning does not correct the number or acknowledge my decision.” She then asks whether they can first establish the accurate figure and a customer correction, while reserving a separate review of the release decision.

Leo accepts that separation on one condition: they actually return to the release decision on 13 July, 14:00–14:30. Both accept fifteen minutes of preparation for that review. The meeting's thirty minutes contain five for the failed opening and acknowledgment, ten for checking the issue, ten for assigning the correction and five for confirming the return point and work allowances.

Use the answer to narrow the next action

The immediate facts are ninety completed surveys and 1,200 eligible invitations. The correct calculation is 90 ÷ 1,200 × 100 = 7.5%. The earlier 9% used a draft denominator of 1,000. It is a difference of 1.5 percentage points. The corrected rate is not evidence about the views of every invited person.

During the meeting Leo accepts twenty-five minutes to verify the invitation count and send a concise source note to Imani. He also accepts a later forty-minute editing block if the customer agrees a two-slide scope; otherwise they must discuss a different commitment. He does not accept an urgent full-report rewrite.

Leo performs the source check and sends the note on 7 July, 15:00–15:25. Imani uses 15:30–15:45 to read it for five minutes, then compose and send the customer correction for ten. Her complete message is:

The draft I sent contains an incorrect completion rate. The verified figures are 90 completed surveys from 1,200 eligible invitations, giving 7.5%, not 9%. I sent it before the agreed data check was complete, and I am sorry. Please do not use that draft. We can prepare a corrected two-slide extract, but cannot responsibly revise the full report today. Would that limited extract meet your immediate board need? Please confirm the intended use and a workable deadline before we commit.

Nia spends fifteen minutes reading and replying on 8 July, 10:00–10:15. She says the full report can wait, accepts only the two-slide extract for internal board preparation and requires it by 10 July at noon. She also agrees to spend ten minutes reviewing it and answering by noon that day. She permits the scope reply to be shared with Leo, the assigned analyst. She does not authorize external publication or call the original mistake resolved.

Imani uses 10:30–10:35 to read and send acceptance of those terms, copying Leo on the authorized two-slide instruction. Nia spends 10:45–10:50 reading that acceptance. These are actual receiving steps, not agreement inferred from a sent email. Leo's conditional editing block now fits the accepted scope.

Deliver the corrected output and record its limits

Leo uses 9 July, 14:00–14:40 for five minutes reading the authorized scope, twenty-five revising the two slides, five checking them and five sending them to Imani. Imani spends 14:45–15:05 reading it, checking its scope and sending the final version. The complete substantive slide text is:

Slide 1: Survey participation correction. Ninety people completed the survey from 1,200 eligible invitations. The completion rate is 7.5%. The earlier draft reported 9% using an outdated denominator of 1,000; that figure is withdrawn. The completed-survey count has not changed.

Slide 2: Use and interpretation. This extract corrects participation reporting for internal board preparation. It does not revise the remaining report or establish that respondents represent everyone invited. The full report remains pending review. No external publication is authorized by this extract.

The extract is sent twenty hours and fifty-five minutes before the accepted noon deadline. Nia actually reads and replies on 10 July, 11:00–11:10: “This corrects the participation figure I need for the board preparation. I will use these two slides internally. I still want the full report reviewed separately; please do not send it as final.” Imani reads that answer during 11:10–11:15. She sends no additional acknowledgment.

The immediate outcome is a usable, limited correction with an actual customer answer. It is not approval of the whole report, a public release or evidence that the relationship is repaired. The full-report review remains outside the completed two-slide commitment.

Return to the concern that made the shift acceptable

On 13 July both Imani and Leo spend their agreed fifteen minutes preparing. They then hold the promised thirty-minute release review. The first ten minutes reconstruct the sequence: Leo's check was incomplete when Imani sent the draft. The next ten examine the release decision. The last ten specify an explicit data-check and director-release gate for future work.

Imani acknowledges that she treated time pressure as permission to bypass the check. Leo says the correction was useful but he remains frustrated that his warning was initially treated as negativity. Imani does not ask him to describe the experience as positive. They agree that future data changes need a named analyst check and a separate director release decision. This future gate has not yet been tested.

Imani records the complete case in six fields:

Imani spends ten minutes writing that record. Sera reads it for fifteen minutes before their 14 July thirty-minute coaching review. Ten minutes cover the failed shift and repair, ten the actual customer and Leo responses, and ten Imani's future choice. Sera then spends ten minutes on her internal notes. There is no additional written feedback exchange hidden outside those blocks.

Count the whole correction and coaching effort

The whole account is 560 of 630 person-minutes, leaving 70. The separate remainders are 25 for Imani, 20 for Leo, 10 for Nia and 15 for Sera. This includes reading and sending as well as the visible calls. Prior report production, the customer's later board meeting, the future full-report review and the new gate's eventual use are not claimed as completed work.

There is no controlled comparison showing that these prompts caused the correction. The case instead supplies a traceable sequence: a rejected interpretation, an acknowledged fact, accepted work, actual output and the promised return. A different conversation might still require a pause, another decision maker or a formal process.

Prepare and review the practice in Coachful

Keep the client's learning question specific: for example, “Notice when I try to move on before acknowledging an actual concern.” An ordinary goal or action can support that focus. Completing it should refer to the agreed behavior, not a promise that someone else now trusts the client.

An original worksheet can hold the six-field record. Use only material appropriate for that audience; a coaching record does not need the customer's confidential report or raw survey responses. The client can describe their own contribution and anonymize other people where suitable.

When using a submitted worksheet, Coach review is an internal section for the authorized coach team. Saving it does not return feedback to the client. If the service promises a written response, send it through the actual agreed client-facing route, such as the individual's private client conversation, and count the handling. In this fictional case Sera's actual feedback occurs during the final thirty-minute joint review.

Unshared journal pages are not ordinary staff coaching notes; clients can explicitly share eligible journal work through supported specialty features. Private course takeaways have their own narrower privacy rules. Do not infer that a personal reflection is available for assessment or that all journal pages are permanently unreadable by the team.

For summarized sessions, check the intended audience and recap controls before sensitive group material is included. Approve before sending holds new unsent recaps; session-type overrides also matter. Previously sent material is not recalled by changing the setting. A recap is not a substitute for the customer's actual approval or the promised accountability meeting.

Explore the Coachful features for organizing this preparation and review. The guides to conversational rituals and Glaser's three conversation levels address adjacent ideas; this article concerns selecting and testing a shift in a particular exchange.

A practical sequence for your next conversation

1. Identify the real purpose and authority. Separate your preferred answer from the decision the other person can actually make.
2. Describe the difficulty observably. Record the disputed fact or repeated exchange without diagnosing a brain state.
3. Check what you have heard. Leave room for the person's correction.
4. Offer one transparent shift. Explain why it might help and what it preserves.
5. Use the answer. Accept a rejection, clarify a condition or choose a pause when needed.
6. Agree the receiving work and return point. Name the person, task, usable window and any unresolved concern.
7. Review actual follow-through. Separate sent from read, proposed from accepted and an improved output from restored trust.

Troubleshoot a shift that is not helping

The reframe sounds dismissive: acknowledge your contribution and return to the specific concern. Do not defend your intention as proof of a helpful impact.

The meeting keeps expanding: identify which decision is authorized now. Ask before assigning preparation or replies to people outside the meeting.

The other person says no: clarify what remains possible without treating refusal as resistance to be overcome. The next action may be stopping the proposed work.

The room becomes quiet: check understanding and access to response. Quietness does not establish agreement or a biological change.

A deferred issue keeps disappearing: acknowledge the missed promise and agree an appropriate route. Do not claim a successful shift while concealing failed follow-through.

Frequently asked questions

Must I reframe, refocus and redirect in order?

No. Use the distinctions to consider your choices. One clarification or an honest pause may be sufficient. This guide does not prescribe a branded three-step intervention.

Does conversational agility mean staying positive?

No. Accuracy, accountability and a genuine refusal may matter more than a positive interpretation. In the case, the initial positive reframe fails and the concern must be acknowledged.

Can a phrase reset another person's brain?

Do not promise that. Record what was said, understood and done. A conversational observation does not establish a measured neural or hormonal mechanism.

What if the person rejects my proposed shift?

Ask what it misses and consider staying with the original issue. You may need to change your interpretation, obtain evidence, stop or use a different process.

How do I judge whether the practice helped?

Look for relevant information, accurate decisions, actual receiving agreements and completed follow-through. Keep unresolved concerns visible. A single case cannot establish lasting trust or general effectiveness.