Back to all articlesHuman and AI collaboration

Resolve Conflicting Instructions From Two Managers

Alactic Editorial10 min read

Prepared with AI assistance by Alactic Editorial. Examples identified as hypothetical are illustrative, not customer results.

Consider this hypothetical scenario: a sales manager tells an AI-assisted team to release a customer offer. An operations manager tells the same team to hold it until capacity is confirmed. Both manage part of the work, both expect compliance, and the second message arrives just before the scheduled release.

Do not treat the latest instruction as the winning instruction. Identify the exact business decision in dispute, check who owns that decision, and send the unresolved conflict to a designated tie-breaker. While that happens, continue only the work that is authorised and does not settle the disputed question by accident.

An AI assistant can help compare instructions, prepare an escalation, and maintain a decision record. It should not infer a new management hierarchy from job titles, message order, confidence, or urgency. The worksheet below makes that boundary practical.

1. Separate operational seniority from decision ownership

Operational seniority describes who coordinates people, schedules work, or supervises a team. Decision ownership describes who may decide a particular issue. Those responsibilities can overlap without being identical.

A manager might supervise the employee preparing an offer but lack authority to commit delivery capacity. Another might own fulfilment readiness but not set commercial terms. Neither fact alone establishes who can approve the combined commitment.

Before choosing between instructions, look for an existing authority record: a project charter, an approval matrix, a named business owner, or a documented delegation. Check its scope as well as its existence. Authority over drafting does not necessarily include authority over release.

  • Decision owner: May decide the specific business question within a defined scope.
  • Consulted manager: Supplies constraints or recommendations without necessarily holding final approval.
  • Execution owner: Carries out the authorised decision.
  • Tie-breaker: Has an established mandate to resolve disagreement or clarify ownership.

These are recommended roles for the workflow, not a claim that every business already uses them. If no tie-breaker has been designated, record that gap. Ask the organisation's established governance channel to confirm who can decide; do not nominate someone merely because their title sounds more senior.

2. Describe the conflict as one observable decision

Broad descriptions such as sales and operations disagree make escalation harder. They invite a debate about relationships rather than a decision about work.

Rewrite the conflict in terms of an action, its scope, its timing, and any conditions. Ask whether both instructions can be satisfied at once. If one manager wants a draft today and another prohibits external release, there may be no conflict: the team can draft without sending. If one requires release today and another prohibits release today, the incompatibility is real.

A useful conflict statement follows this structure:

We cannot both perform action A for this scope by this time and comply with instruction B, because B prohibits that same action until a condition is met.

Do not reduce a conditional instruction to an unconditional veto. Release after capacity confirmation is different from do not release. Preserve exceptions, deadlines, and the original scope when summarising.

Also distinguish disagreement from missing information. If both managers agree that release requires capacity confirmation, but nobody knows whether confirmation exists, the immediate task is verification. If they disagree about whether confirmation is required, the issue is authority over the release condition.

3. Filled authority-conflict worksheet: a fictional worked example

The following scenario, names, times, quantities, and outcome are hypothetical. It illustrates a recommended operating process, not an Alactic customer result or a product capability.

A team is preparing a service-package offer for a selected customer group. Maya, the sales manager, owns campaign preparation. Leon, the operations manager, owns service scheduling. Their existing launch charter names Priya, the programme sponsor, as the tie-breaker for disputes involving both commercial release and delivery readiness.

Authority-conflict worksheet — hypothetical completed record

  • Record: Service-package release conflict, Tuesday.
  • Affected decision: Whether to release the current offer to the selected customer group at 14:00 with next-week service availability stated in the offer.
  • Competing instruction A: At 09:10, Maya directs the team to release the current offer at 14:00. Her stated objective is to preserve the planned commercial launch date.
  • Competing instruction B: At 10:05, Leon directs the team to withhold release until operations confirms next-week service capacity. His stated concern is committing capacity that has not been confirmed.
  • Exact incompatibility: Releasing at 14:00 before capacity confirmation would satisfy A but violate B. Holding beyond 14:00 would satisfy B but miss A's release instruction.
  • Known authority: Maya owns campaign preparation; Leon owns service scheduling. Neither role description independently settles this combined release decision.
  • Authority evidence: The launch charter designates Priya to resolve cross-functional launch disputes. The worksheet records where the team can find that charter.
  • Interim boundary: Continue internal proofreading and capacity checking. Do not release the offer, promise availability to customers, or mark release approval as complete. These limits remain within the existing preparation-only authorisation.
  • Tie-breaker: Priya, acting under the launch charter, not merely because of seniority.
  • Decision requested: Authorise release as drafted, hold release pending capacity confirmation, or approve a revised offer with a different commitment.
  • Requested response and fallback: Seek a decision by 12:30. If none arrives, retain the preparation-only boundary and report the release as blocked, not approved.
  • Recorded resolution: At 12:05, Priya directs the team to hold the current offer. Leon will provide a capacity status by 15:00. Maya may prepare a revised version without the next-week commitment, but that version requires explicit release approval.
  • Execution owner and status: Maya owns the revised draft; Leon owns the capacity update. External release remains on hold. The execution lead records the decision in the shared work record and requests acknowledgement from both managers.
  • Reopening trigger: Confirmed capacity or a materially revised offer returns to the named approver. Neither event automatically authorises release.

The worksheet records more than a winner. It preserves the conditions under which the decision applies, who must act next, and what remains unauthorised. That prevents a later reader from interpreting permission to prepare as permission to send. Reuse its field labels for the next conflict, replacing the fictional entries with verified instructions and authority evidence.

4. Choose an interim boundary that does not decide the dispute

A useful interim boundary protects the disputed decision while allowing independent work to continue. It is narrower than stopping everything and more cautious than continuing until someone objects.

Divide pending work into three groups:

  1. Uncontested and already authorised: Continue it. In the hypothetical example, internal proofreading can proceed without committing capacity.
  2. Dependent on the disputed decision: Hold it. Release and customer commitments belong here.
  3. Unclear: Ask for clarification before proceeding. An apparently internal status change may trigger downstream work or be interpreted as approval.

Reversibility helps assess an action, but it does not create authority. A draft edit might be easy to undo and still fall outside the team's remit. Conversely, holding work has costs: a missed launch slot, idle preparation, or delayed customer communication. Record those costs for the tie-breaker rather than using them as permission to proceed.

If an existing authorised contingency covers the situation, follow it within its stated scope. Do not invent an emergency exception because the deadline is close. Where no contingency exists, make the blocked state visible and request an explicit decision.

5. Send one decision-ready escalation

The tie-breaker should not have to reconstruct the conflict from a long conversation. Send both managers the same neutral summary, attach or locate the worksheet, and ask for a specific decision.

Hypothetical escalation message:

The team has incompatible directions on today's offer release. Maya requests release at 14:00; Leon requires capacity confirmation before release. The unresolved question is whether we may issue the current offer with next-week availability before that confirmation.

The launch charter names Priya as tie-breaker for this cross-functional decision. We are continuing internal preparation but withholding release and customer commitments. Please confirm whether to release as drafted, hold for capacity confirmation, or approve a revised commitment. We need a response by 12:30 to preserve the planned release time. Without a decision, the release remains blocked under the preparation-only boundary.

The response time is a coordination request, not a mechanism that transfers authority when time expires. Silence is not approval unless an applicable, previously authorised process explicitly establishes that rule.

If ownership itself is disputed, separate that question from the business choice. First ask the established governance channel to confirm who can decide. Then send the business question to that person. Asking several executives independently until one agrees creates another conflict rather than resolving the first.

6. Give the AI assistant a bounded coordination role

For conflicting manager instructions, an AI assistant is most useful as a careful coordinator of the record. It can help identify incompatible actions and draft a concise escalation, subject to human review and the information it is authorised to access.

It should not decide that one manager outranks another because their message is newer, more forceful, or more detailed. It should also avoid blending the directions into a compromise that nobody approved. Reducing the audience or changing the promised service date may create a new business decision, not resolve the original one.

A reusable operating instruction is:

When human directions appear incompatible, identify the affected action, scope, timing, and conditions. Check the available authority record. Do not infer decision ownership from title or message order. Prepare a conflict worksheet for human review, identify the documented tie-breaker or flag that none is confirmed, and distinguish authorised preparation from disputed execution. Do not execute the disputed action without the required authorisation.

This is a recommended workflow, not evidence that any particular assistant enforces it. If implementing it in software, verify the controls, permissions, and approval behaviour available in the workspace and current plan. A written instruction alone should not be mistaken for an execution safeguard.

7. Recover when someone has already acted

If the team followed the latest message and released the offer, the next step is not to rewrite the record as though approval existed. Record what happened, when it happened, and which instruction the executor relied on.

Separate containment from the original decision. Identify any further actions that would repeat or enlarge the disputed commitment, and seek an authorised boundary for them. Do not send a corrective customer message, cancel commitments, or alter records merely to make the conflict disappear; those actions may require their own approval.

Give the tie-breaker a factual recovery brief:

  • What action occurred and its known scope.
  • What remains pending or can still be controlled.
  • Which consequences are confirmed and which are uncertain.
  • What containment or correction requires a decision.
  • Who will carry out and verify the authorised recovery.

Other failures are less visible. A private resolution may never reach the execution team. A revised draft may inherit an approval that applied only to an earlier version. A manager may acknowledge receiving the worksheet without approving its contents. Use explicit decision language, identify the affected version or scope, and distinguish receipt from authorisation.

8. Close the record and test the next handoff

A conflict is not operationally closed merely because someone has spoken. Record the chosen action, its conditions, the decision-maker's authority, the execution owner, and the next review trigger. Mark which earlier directions are superseded for this decision, without treating the resolution as a permanent transfer of either manager's responsibilities.

For the next live conflict, assign one coordinator to complete the worksheet before disputed execution. Check three observable outcomes: whether the exact incompatibility was recorded, whether the tie-breaker's mandate was verified, and whether the resulting action stayed within the recorded approval.

Also record time spent awaiting ownership clarification separately from time awaiting the business decision. If ownership repeatedly takes longer to establish than the decision takes to make, the practical improvement is an agreed authority map—not faster guessing by staff or an AI assistant.

Human and AI collaborationDecision ownershipTeam operationsApproval workflows

Put the next conversation to work.

Explore Alactic Communications for WhatsApp conversations, campaigns and AI-assisted workflows. Choose the product and plan that fit your team.

Explore Communications