Conversational Intelligence Three Levels Matrix: A Case

A conversational intelligence three levels matrix can help a leader separate an information question, a recommendation and an opportunity to develop a new option. The useful question is what the work needs now. A direct factual answer may be more helpful than another exploratory question; a new possibility may be needed when the current options do not address the problem.

This guide briefly attributes Judith E. Glaser's framework, supplies an original planning matrix and completes an internal-guide case. In that case, the participants conduct a bounded meeting, change a recommendation, test two fictional examples, repair a failed instruction and receive an actual publication refusal. None of those results establishes a hormone change, a person's intelligence or a restored relationship.

Coachful is our #1 recommendation for selected meeting preparation, substantive private coach feedback and review. This is an editorial workflow recommendation. The platform does not turn these conversational categories into a validated assessment or authorize the organizational decision.

Keep the named framework and the practical aid distinct

In her Mind Tools interview with Rachel Salaman, Glaser describes transactional, positional and transformational conversations, and explains that all three have a use. She explicitly rejects staying in the third level all the time. The page carries a 2024 publication label; that does not establish when the original interview was conducted.

Glaser's 2014 article Navigational Listening describes informing, persuading and co-creating as different conversational purposes. Her The DNA of Healthy Conversations, also from 2014, discusses useful and unhelpful expressions of the levels, including exploration that becomes talk without action. These are originator accounts of a practitioner framework, not controlled trials validating the matrix below.

The table, questions, agenda and case here are original coordination aids. They do not reproduce a licensed assessment, proprietary chart or training protocol. Use the names to orient attention, with clear attribution; do not score a person or claim certification from using the worksheet.

The practical evidence is ordinary and inspectable: what was asked, which information was checked, whether someone could challenge a recommendation, what work was accepted and what happened to the resulting draft. A label alone does not establish trust, safety or brain activity.

Three purposes a meeting may need

Transactional: establish the information

Clarify a fact, instruction, document version or deadline. State what is known and what remains uncertain. For example, “This is the current internal instruction; the exception decision remains with Jo.” The recipient can check or correct the statement.

The mismatch is using repeated updates to avoid a decision. A longer status report does not choose between competing priorities. Conversely, asking a group to discover a known date wastes time if someone can simply supply it.

The output is a confirmed fact or an identified information gap. It need not be exciting to be useful.

Positional: compare recommendations and their reasons

State the proposed choice and why you support it. Identify evidence, uncertainty and a condition that could change the recommendation. Invite another person's reasoning rather than asking questions only to steer them toward your preferred answer.

The mismatch is treating scrutiny as disloyalty or presenting a decision already made as an open debate. A firm recommendation can coexist with inquiry; the relevant test is whether information can actually affect the choice.

The output is a clear comparison, including legitimate disagreement. A compromise is an option, not an obligation.

Transformational: develop a possibility that is still open

Explore a problem that the existing proposals have not resolved. Ask which assumptions can change and which requirements remain fixed. The group may create another option or discover that the attractive idea is not currently feasible.

The mismatch is promising co-creation while withholding the actual boundaries, or continuing exploration without an owner or next decision. End by stating what was chosen, declined or left for further work.

The output is an option with a route for evaluation. Exploration is not permission to implement it.

Use an original matrix tied to the decision

Start with one agenda item. Write its present need and the contribution required. If participants are solving different problems, name the distinction: one person may be supplying a fact while another is asking for a decision. Do not infer that they occupy different ranks of conversational ability.

A meeting can return to information after exploring an option. It can also stop because the necessary authority is absent. Movement among purposes is useful when it serves the work, not when an observer maximizes a count of discovery questions.

Choose a pivot for the actual mismatch

When a fact is missing, ask for the fact. Suppose a colleague asks which version a reviewer examined, but the presenter answers with a vision for the project. A practical pivot is: “The direction matters, and we first need the version used for this finding.” Check the answer against the record. Do not turn an ordinary clarification into a test of whether someone is open to transformation.

When recommendations conflict, expose the condition. One person wants an earlier date and another wants an additional check. Ask each to state the reason and what could change the recommendation. The group might choose a later date, narrower scope or no change. It need not blend incompatible requirements into a halfway answer. Record the authority and rationale for whichever choice is actually made.

When the available options miss the problem, explore within the boundary. An instruction might be both difficult to read and necessary to retain. Ask which part is fixed and which part can be redesigned. A new layout, example or sequence is a possibility to evaluate, not a reason to waive the requirement. If nobody present can authorize the next evaluation, the closing output is a request to the responsible person.

At review, examine the consequence of the pivot. Did a relevant source become clear? Did a recommendation change or its limitation become explicit? Was the new option actually tested? A count of conversational categories cannot answer those questions. A concise decision record can retain a useful disagreement, a failed test or an honest hold.

Prepare the receiving contribution and meeting boundary

Before asking a colleague to join, include preparation, the meeting and any later check. A request for thirty-five minutes can otherwise conceal an hour of reading and reviewing. Ask what the receiving person can accept and who can authorize the next step.

Choose a closing output: an agreed preparation task, a declined proposal, a draft for review or an authorized decision. State what the meeting cannot approve. A coach can support the client's preparation without attending the work discussion or gaining access to company records.

An observer is optional. If used, agree the purpose, audience and method. Notes about the agenda item can be sufficient; recording other people's speech is not required. Do not classify colleagues' motives or physiology from how often they ask or tell.

For rehearsal, try three short contributions on the same issue: state the current fact, explain a recommendation with a condition, then ask about an alternative within the actual constraints. Check whether the person receiving each contribution understands its purpose. A practiced sentence remains a rehearsal until it is used in the real work.

Obtain an actual invitation and scope answer

This original fictional case runs from 3–17 June 2042. Carmen leads customer training, Jo owns the internal content check, Rafi controls publication and Dee coaches Carmen. Existing arrangements allow Carmen to share a selected coordination account with Dee; customer and internal source records remain in the company system.

The company has a standard check and an exception route for an internal guide. Carmen wants to remove a repeated instruction. Jo believes that an apparently normal standard result does not settle an exception. Their roles are established, but they have not yet compared the wording through the same examples.

Dee spends ten minutes preparing for a forty-minute coaching conversation on 3 June. Carmen calls Jo obstructive. That interpretation remains in the private session; it is not evidence about Jo's motive or a label sent to Jo. Carmen chooses to ask for a discussion about the actual instruction rather than seek the coach's support for bypassing the check.

Carmen uses twenty minutes on 4 June to prepare and send Jo a request for a sixty-five-minute workshop. Jo spends fifteen minutes on 5 June reading it and sending a counterproposal:

I cannot offer the sixty-five-minute workshop. I can give thirty-five minutes to compare the current instruction, our recommendations and another wording option. I can prepare for ten minutes, test the resulting draft against two fictional examples for ten minutes, and provide one retest of up to seven minutes if it needs correction. Please confirm the internal-only purpose and publication boundary with Rafi before treating this as agreed.

Carmen uses ten minutes to read that answer and send Rafi a scoped request. Rafi spends fifteen minutes on 6 June reading it and replying: the bounded discussion and two-example internal test may proceed; no customer-facing change is authorized. Rafi also agrees to read the final draft and results for the publication decision. Jo retains the content check.

Carmen spends ten minutes reading the answer and sending Jo the actual scope. Jo spends five minutes on 7 June reading it and confirming the complete receiving contribution. Carmen uses five minutes to read that confirmation and send Dee a selected account.

Dee spends fifteen minutes reading it and sending substantive feedback: keep the current instruction and interpretation separate, compare real recommendations and preserve Rafi's release boundary. Carmen spends five minutes reading the feedback without another reply. There is no extra meeting involving Dee and the company participants.

Conduct a thirty-five-minute mixed-purpose conversation

Carmen spends twenty minutes on 10 June preparing and sending a concise brief. Jo uses the accepted ten minutes immediately before their 11 June meeting to read it and the existing source instruction. Both then participate for thirty-five minutes.

The information phase establishes that the exception flag must be checked before choosing the route. It is not a claim that the fictional instruction is a legal or technical standard for other organizations. Their short exchange is:

Carmen: I have been treating a normal standard-check result as enough to continue. Is that what the current instruction permits?
Jo: Not when the exception flag is present. That flag needs the named owner's decision.
Carmen: Then removing the repeated sentence without showing that order could change the route, not just shorten the text.

During the recommendation phase, Carmen explains the concern about repeated steps and says that a clear distinction between routes would change the removal recommendation. Jo recommends retaining the current process until the new wording has been checked. Jo would support clearer wording that preserves the condition.

They then explore whether the guide can name the exception first and state the ordinary route separately. Carmen changes the initial proposal: retain the check and make the branch explicit. Jo supports testing that wording. This is an actual change in the recommendation, not a claim that the colleague has become more trusting.

The final five minutes identify the owner and limit. Carmen will prepare the draft, Jo will read two fictional examples and Rafi will decide whether anything may be published. The meeting does not assign work to an unnamed customer, technical team or reviewer.

Let the reader test change the draft

Carmen spends thirty minutes after the meeting preparing the prototype and two original fictional examples, then sending them to Jo. One example has no exception flag. The other has a flag alongside a normal standard-check result. Neither contains real customer information.

The first draft includes the instruction, “When the standard check is complete, continue. Ask the owner if the exception route applies.” On 12 June, Jo spends ten minutes reading the draft, applying it to both examples and sending the result: the ordinary example works, but the exception example can lead a reader to continue before checking the flag. One of two passes.

That is a substantive failure of the wording. Calling the earlier meeting transformational would not fix the sequence. Carmen uses fifteen minutes to read the finding, repair the draft and send the revised complete passage:

Check the current profile for the exception flag before choosing a route. If no flag is present, complete the standard check before proceeding. If the flag is present, stop and ask the named internal owner for the exception decision; a normal standard-check result does not override that flag. Record the route and any decision still pending. This is an internal test draft, not permission to change the customer-facing guide or process a live account.

Jo uses seven minutes on 13 June to read the revision, retest both examples and send a limited answer: both now produce the intended route. This acceptance concerns the two fictional examples and Jo's content condition; it is not an independent customer study or a release decision.

Carmen uses ten minutes reading that answer, recording the result and sending Rafi the final draft and results. Rafi spends ten minutes on 14 June reading them and replying: no customer publication yet, because accessibility and localization checks remain outstanding. The internal wording test is complete, while the proposed external change is held.

Carmen spends five minutes reading Rafi's answer and sending Dee the selected outcome. No new work is assigned to Jo, and the unperformed accessibility and localization checks are not treated as completed by the discussion.

Review the result without scoring the people

For the 17 June coaching review, Dee uses ten minutes reading the selected outcome and preparing. Carmen prepares separately for ten minutes. They meet for twenty minutes each.

Carmen chooses to retain the three-purpose preparation record and accepts the publication hold. Carmen declines a proposed ranking of colleagues by level and does not ask Dee to circulate the private “obstructive” interpretation. The useful evidence is a changed recommendation, a failed instruction, a repaired draft and an actual limit from the publication owner.

Dee spends ten minutes recording the review and sending this substantive private close:

You checked the source before defending the proposed removal. The meeting compared recommendations and produced another wording option. The first reader test failed the exception case; the revision passed both fictional examples. Rafi has not authorized publication, and the remaining checks are still open. These results describe this work. They do not rank Jo, establish a brain or hormone effect, or show that the wider relationship is repaired.

Carmen spends five minutes reading the close without another reply. The bounded account is:

The account totals 452 person-minutes against 535 allocated, leaving eighty-three within this bounded scope. Both people in each conversation are counted, along with the receiving checks and actual decisions. Earlier source-document production, ordinary job duties and future publication checks are outside the account, not free. There is no claimed customer time saving or financial outcome.

Keep the preparation and operational evidence in their proper places

A useful original reflection contains the agenda item, present need, known source, recommendation, open option and actual decision owner. Complete it with the client rather than converting it into a score of the other participants. A request to inspect evidence should remain possible at every stage.

In Coachful, coaches can open a client's one-to-one conversation from the message icon on that client's profile. Dee uses Carmen's private conversation for the selected scope, substantive feedback, outcome and close. Company source records, prototype and publication decision remain with the responsible organizational owners. A private coach note would not establish that Jo or Rafi received the work.

Use the Ladder of Conclusions guide, reframing, refocusing and redirecting and Coachful's features for related preparation. The existing product images are illustrative demo screens, not a recording of this meeting, a scored matrix or proof of the proposed release.

Frequently asked questions

Is the third level always the right choice?

No. Confirmed information and reasoned recommendations may be what the task needs. Glaser explicitly described a use for all three.

Can a firm decision follow discovery?

Yes, when the authority and available choices are clear. State the decision and limits rather than promising consensus the group cannot provide.

Does a pass on two examples prove the guide is ready?

No. It supports only the defined internal test. Other review requirements and the publication owner's decision still matter.

Can I score employees with this matrix?

This original table is a planning aid, not a validated assessment of people. Do not infer intelligence, physiology or trust from a category.

Does disagreement mean the conversation failed?

No. A real constraint or unresolved difference can be a useful result. Record what happened and who can decide the next step.