People Management
Manage an Expert Who Knows More Than You About the Work
Manage deeper technical expertise with clear outcomes, review responsibilities and resource decisions. A complete fictional working agreement and two uncertainty cases show how to respect specialist judgment, obtain qualified review and keep unapproved technical conclusions open.
Manage an expert by making the outcome, constraints and decision responsibilities clear, then asking for the reasoning you need to support a sound decision. You do not need to outperform the person in their specialty. You do need to distinguish their technical recommendation from your resource decisions and any approval that belongs to another owner.
This LandOffer.ai guide contains an original fictional working agreement and two complete management cases involving a specialist in document rendering. The records and technical observations are authored examples, not results from a real rendering engine. No employee action, production change or software test was performed.

Admit the knowledge gap without withdrawing from responsibility
A specialist may know more about a particular area, such as typography, than their manager. That asymmetry does not settle who controls scope, budget or release approval. Pretending to understand every detail can make discussion less accurate. Withdrawing from the discussion entirely creates a different problem: the expert is left to infer priorities, negotiate resources and resolve conflicts outside their authority.
In a first-person LeadDev account, Andrea Posadino describes the limitations of concealing a knowledge gap and the value of supplying wider business context. It is a practitioner's experience, not a controlled study or a universal rule that managers should avoid technical questions. The useful distinction is between asking to understand a decision and trying to establish authority by sounding equally expert. Leading advanced technical expertise.
Tell the person what you can contribute. That may include clarifying the required outcome, obtaining another team's input, protecting investigation time or making a trade-off visible to leadership. Ask what they need from you instead of assuming that expertise removes their need for management support.
Keep learning enough of the domain to ask relevant questions. GitLab's engineering-management handbook expects technical credibility while distinguishing management and technical leadership. It also cautions managers against taking critical-path implementation work. These are GitLab's practices, not a requirement to apply its organization model elsewhere. GitLab engineering management.
Ask for a decision explanation, not an introductory course
Start with the work decision and its consequence. Ask which options are feasible, what evidence supports the recommendation, what remains uncertain and what would change the recommendation. The level of detail should match the consequence and the audience, rather than the manager's desire to understand every mechanism immediately.
For a low-impact reversible choice, a brief explanation and an existing review route may be enough. A change with contractual, security or broad operational implications may need a qualified reviewer and a separate approval. If you cannot assess a consequential claim, obtain the appropriate qualified review before treating it as established.
Agree on an explanation format with the expert. A short written comparison can let them preserve technical nuance while helping you explain a resource decision. Avoid repeatedly asking for the same beginner-level lesson during a deadline; reserve separate learning time if you need foundational context.
GitLab's DRI guidance illustrates another useful boundary: a decision owner should consult relevant expertise, and some circumstances require additional approvals. Its internal DRI rules are employer-specific. In your setting, verify the actual authority rather than assuming either the manager or the most knowledgeable person has the final say on everything. GitLab ownership guidance.
Worked asset one: an expert–manager decision agreement
The following team and arrangements are fictional. Manager Basma leads six engineers working on an internal document-generation service. Specialist Corin has deeper knowledge of font handling and text layout. Product owner Odelia sets the supported document requirements. Qualified reviewer Teo provides independent technical review, and procurement owner Vesta handles font-license questions through the organization's actual process within the case.
They agree that a proposed font replacement should be evaluated against the required document set before a release decision. Basma is not the typography approver merely because she manages Corin. Corin is not responsible for deciding the employer's acceptable contractual exposure merely because he understands the font files.
| Area | Filled responsibility agreement |
|---|---|
| Outcome and scope | Odelia defines required document types and supported scripts; Basma resolves delivery priorities with her |
| Technical recommendation | Corin compares feasible approaches, identifies uncertainty and records the basis for the recommendation |
| Technical review | Teo reviews the evidence against the agreed criteria; a missing required case remains a gap |
| Capacity | Basma reserves six hours of Corin's existing weekly capacity and two hours of Teo's; she defers eight hours of nonurgent layout cleanup with its work owner |
| Procurement boundary | Vesta obtains the appropriate license determination; neither Basma nor Corin invents permission from a file's availability |
| Release boundary | The existing authorized release owner retains the release decision; this agreement does not bypass that route |
| Explanation format | One page: recommendation, alternatives, evidence covered, missing cases, consequences and next decision needed |
| Escalation | An unresolved required case, disputed review or missing approval returns to Basma for routing; no silent default to release |
| Review point | Revisit the allocation after one week; continued investigation requires a new priority decision |
The eight-hour displacement matches the six specialist hours plus two reviewer hours. It does not include Basma's coordination or Vesta's work, which need separate planning. It is also reserved capacity in a fictional agreement, not a timesheet proving that technical validation has occurred.
Their completed opening conversation
Basma: “I cannot independently judge the shaping details. I can make sure the required documents are explicit, protect the review time and get the license question to Vesta. Please explain which decision you need from me and what remains a specialist judgment.”
Corin: “I recommend comparing the replacement against the current font on the approved document set. I can explain the gaps and alternatives. I do not want a new approval from you on every implementation detail, but I do need an owner for conflicts between the deadline and missing coverage.”
Basma: “Bring that conflict to me. I will take the scope and date options to Odelia and the release owner. I will not describe an incomplete review as your technical approval.”
This conversation preserves expertise while making management work concrete. It does not turn Corin into the sole reviewer of his own proposal or give Basma authority to settle a license question. The agreement gives both people a usable route when a deadline or unresolved claim makes the discussion uncomfortable.
Worked asset two: two uncertainty cases and their recorded decisions
These cases continue the fictional arrangement. Their technical observations are supplied scenario facts, not measurements or recommendations about any named product, font or script. The point is how the manager responds to incomplete knowledge.
Case one: the favored option has incomplete coverage
Corin's authored comparison E-17 covers nine of twelve required document scenarios. Eight meet the agreed checks; one has an unresolved layout difference. Three required scenarios have not yet been reviewed. A stakeholder calls the result “almost done” and asks Basma to confirm the release date.
Basma does not convert eight passing scenarios into an overall safety judgment. She asks Corin and Teo to distinguish the known difference from the unreviewed cases. Corin explains the difference and recommends keeping the current font for the affected document type until review is complete. Teo agrees that the three missing cases prevent a complete technical conclusion.
The completed decision record reads:
E-17 has eight checked scenarios meeting criteria, one checked scenario with an unresolved difference and three unchecked scenarios. No full-coverage approval exists. Basma and Odelia retain the current supported font for this release rather than reduce the required document scope without approval. Corin's replacement remains a proposal for later review.
Basma will report the deferred replacement to the release owner and stakeholders. Corin owns documenting the remaining evidence; Teo owns its independent review. Continuing beyond the reserved week requires a new capacity decision. The current font's existing release controls remain in place; this record is not new proof that every current behavior is correct.
The case produces a bounded outcome: a proposed replacement is deferred. It does not show that the specialist was wrong to suggest it or that the manager found a superior technical solution. The manager made the evidence gap and scope trade-off visible instead of assigning the expert an unearned approval.
Nine of twelve cases reviewed is 75% coverage of this defined checklist. Eight of nine reviewed cases meeting criteria is a different fraction. Neither number is a probability that production output will be correct. These are descriptive fractions of a supplied fictional checklist, not test results from a rendering engine. The unreviewed cases are not passes, and the one unresolved difference should not disappear inside an average score.
Case two: two qualified people disagree about an approach
A later fictional review concerns how to isolate a layout change for further investigation. Corin favors approach A because it keeps the comparison narrow. Teo favors approach B because it exposes a different interaction he considers relevant. Both can explain their reasoning, but neither has evidence resolving the disagreement yet.
Basma asks each to state the assumption behind the preference and the observation that would change it. She does not choose the argument with more unfamiliar terminology or ask the wider team to vote on technical truth. They agree on a small evidence plan before any production change.
| Decision field | Filled case record E-23 |
|---|---|
| Question | Which approach gives adequate evidence for the agreed layout-review question? |
| Current position | Corin recommends A; Teo recommends B; assumptions remain different |
| Investigation | Corin prepares two comparison artifacts within two hours; Teo reviews them for one hour |
| Scope limit | Approved nonproduction examples only; no production font or access changes |
| Management trade-off | Basma defers a three-hour nonurgent formatting investigation with its owner to make room |
| What happens afterward | Record the evidence and disagreement; Basma asks architecture owner Niko to handle the established technical escalation if it remains unresolved |
| Current outcome | Evidence plan agreed; artifacts, review and technical choice are not yet completed |
A bounded investigation is a management decision that can be complete while the technical answer remains open. Basma owns making room for the work and obtaining the right escalation route. She does not claim the plan itself confirms either approach.
If the time limit expires with uncertainty intact, the next step is another explicit scope, resource or escalation decision. “Keep looking until everyone agrees” would leave Corin and Teo with an unlimited assignment and no owner for its cost.

Give useful feedback even when you cannot check every technical claim
You can assess whether the expert makes assumptions visible, engages the required reviewers, records uncertainty and explains consequences to the appropriate audience. Be specific about behavior you observed. Do not infer low technical quality from an explanation you found difficult without checking what the intended audience needed.
Ask for qualified input when evaluating domain depth. Explain what the reviewer is being asked to assess and how that input will be used. A peer disagreement about one option should not become a broad negative judgment about someone's capability. Confidence and reputation are not substitutes for evidence. A reviewer should identify the specific gap rather than demand another explanation simply because a manager remains unfamiliar with the topic.
Keep career support on the agenda. Ask what kind of work the specialist wants to deepen, where their influence is blocked and what responsibilities have accumulated around them. Someone who is good at answering questions can become the default helper for everyone; expertise is not unlimited capacity.
When the expert's contribution supports a shared result, credit it accurately. Do not present their recommendation as your technical discovery, and do not make them responsible for a delivery commitment you gave without their evidence. Management accountability includes correcting that kind of mismatch.
Lead through the agreement, then revisit it
For the next substantial decision, confirm the desired outcome, the expert recommendation, required review and actual approval owner. Ask for the minimum explanation that makes the trade-off reviewable. Supply the context and resources you control, then leave implementation choices within the agreed technical responsibility.
Use the LandOffer.ai workplace guides alongside your organization's actual authority and review processes. Managing deeper expertise is ongoing work: you should become better informed, and the agreement should evolve as risk, scope and people change. Revisit any unclear authority or workload boundary before the next decision, so neither person has to infer it from seniority.
Sources and Further Reading
- Andrea Posadino, LeadDev: How to lead individuals with advanced technical expertise. May 2, 2024. First-person practitioner account; checked October 27, 2026.
- GitLab: Engineering Management. Official employer handbook, last modified March 19, 2026. Employer-specific practices; checked October 27, 2026.
- GitLab: Directly Responsible Individuals. Official employer handbook, last modified April 15, 2026. Employer-specific authority model; checked October 27, 2026.



