Claude Certified Engineers logoCCE

CCAR-P · Study Guide

← Domain 6: Stakeholder Communication & Lifecycle Management

6.3 · Lesson 3 of 5

Manage stakeholder feedback loops and expectation alignment (including SLAs)

What You Need to Know

Stakeholder feedback loops keep expectations and delivery aligned — including SLAs for latency, quality, cost, and escalation. Publish what stakeholders actually agreed to. Review quality, cost, and SLA adherence on a cadence with evidence. When measured p95 or quality drifts from the agreement, escalate early and renegotiate or redesign — do not hope silently.

Feedback must change the system: backlog, prompts, retrieval, or scope. Emoji reactions and private eng targets are not alignment. When the original SLA was unrealistic, reopen discovery notes rather than pretending the architecture meets a fantasy number.

Alignment loop

  • Agree and publish SLAs with owners
  • Measure and review on a fixed cadence
  • Escalate drift early with evidence
  • Feed actions into backlog and evals
  • Renegotiate when reality and expectation cannot meet

Decision rules

  • Publish SLAs stakeholders actually agreed to.
  • Review quality, cost, and SLA adherence on a cadence.
  • Escalate expectation drift early — silent drift fails.
  • Convert feedback into backlog and eval changes.
  • Reopen discovery when the original SLA was never feasible.

Why silent drift fails

Stakeholders remember the number they heard in the kickoff. If the system runs at 2s while they still believe 200ms, trust collapses even if engineers “knew” the SLA was aspirational. Architects own the reopen: evidence, options, and an explicit new agreement.

Review checklist

  • Are SLAs written, agreed, and visible?
  • Is there a recurring review with quality/cost/SLA charts?
  • Is there a drift threshold and escalation owner?
  • Do review actions become tickets and fixtures?
  • Would stakeholders restate the current SLA correctly?

Exam stems with mismatched expectations reward naming SLA drift and reopening alignment — not blaming ethics, ADRs, or informal vibes.

Exam application

Prefer answers that publish SLAs, schedule evidence-based reviews, and escalate drift. Distractors: perfect-alignment denial, emoji-only feedback, or ignoring measured gaps.

Exam traps

  • Silent SLA drift

    Internal reality diverges from stakeholder expectation with no reopen. Exam items punish ignoring the gap.

  • Unpublished “targets”

    SLAs must be stakeholder-agreed and visible — not silent eng hopes.

  • Emoji-only feedback

    Cadence reviews should track quality, cost, and SLA adherence — not reactions alone.

  • Waiting for a crisis to realign

    Escalate expectation drift early with evidence; do not wait for executive rage.

Practice scenario

Stakeholders still expect 200ms responses; measured p95 is 2s and nobody updated the agreement. What is the issue?

Choose one answer

Build exercise

Install an SLA feedback loop for a support copilot

35 minutes

What you'll learn

  • Publish agreed SLA packs
  • Run cadence reviews with evidence
  • Escalate expectation drift early
  • Feed feedback into backlog and evals

  1. Step 1

    Publish the agreed SLA pack

    Write latency, quality, cost, and escalation SLAs stakeholders signed. Store them where product and ops can see them.

    Why: Unpublished targets create silent drift. Agreement must be shared.

    You should see: SLA one-pager with owners and measurement sources.

  2. Step 2

    Set a review cadence with evidence

    Schedule recurring reviews of quality, cost, and SLA adherence. Bring charts and failing cases — not anecdotes only.

    Why: Feedback loops need scheduled forums and shared measures.

    You should see: Calendar invite + dashboard links + agenda template.

  3. Step 3

    Detect and escalate expectation drift

    When measured performance diverges from the agreement, escalate early. Reopen discovery notes if the original SLA was unrealistic.

    Why: Drift unaddressed becomes a trust failure. Early escalation is the professional move.

    You should see: Drift playbook: threshold → owner → stakeholder message template.

  4. Step 4

    Feed feedback into backlog and evals

    Convert stakeholder complaints and review actions into backlog items and eval fixtures. Close the loop visibly.

    Why: Feedback without change is theater.

    You should see: Three feedback items traced to tickets and eval case ids.

Sources