Back to all articlesAgency customer operations

Set Review Depth According to the Risk of Client Work

Alactic Editorial9 min read

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

A routine report paragraph waits for the same client sign-off as a new customer promise. The first delay consumes account-management time; the second may still receive only a hurried glance. A single approval queue does not ensure that the right people examine the right decisions.

Instead, agree three review depths with the client: execution within previously authorised boundaries, targeted checks of a specific change, and full client review of consequential work. Assign the tier according to what the deliverable changes—not its length, format, or production effort.

Reduced review is not permission to expand scope. It is a negotiated way to carry out work the client has already authorised, while reserving closer scrutiny for decisions that need it.

1. Define the operating boundary before choosing a tier

Two questions often become tangled in an approval thread: has the client authorised this activity, and has this particular output passed the required review? Resolve the first before choosing the second.

An agency might be authorised to prepare monthly performance commentary using an agreed reporting method. That does not necessarily authorise it to change an offer, alter a customer-facing promise, or publish recommendations as accepted client decisions.

Record the operating boundary in practical terms:

  • Included work: the outputs and actions covered by the agreement.
  • Fixed conditions: the facts, approved claims, audiences, channels, and other constraints that must remain unchanged.
  • Release owner: the person who may send, publish, or otherwise use the checked output.
  • Exceptions: changes that require a fresh decision before proceeding.

A useful boundary is specific enough that two people can classify the same deliverable consistently. “Manage reporting” is too broad. “Prepare and release monthly commentary using the agreed source and metric definitions, without new causal claims or recommendations” gives reviewers something they can check.

2. Classify the consequence, not the format

Most report commentary may be routine, but a paragraph announcing a service commitment is not routine merely because it appears in a report. Start with the actual change.

  • What changes for the recipient? Are they receiving an explanation of existing facts, a different offer, or a new expectation about what the client will do?
  • What must be true? Does the output depend on verified facts, or on capacity, costs, or decisions that remain uncertain?
  • How contained is a mistake? Could it be corrected before external use, or might recipients already have acted on it?
  • Is the decision already covered? Can the agency identify a current authorisation that matches the work?

The highest applicable trigger determines the tier. Do not average a consequential change down because the rest of the deliverable is familiar. Conversely, where the substance is approved and editorial discretion is delegated, shortening a sentence need not reopen the underlying decision.

If only one section needs a higher tier, isolate it where practical. The rest can proceed only if it remains accurate and useful without the unresolved section.

3. Use this filled review-tier table as the starting agreement

Fictional worked example: This hypothetical agency supports a client with performance reporting and customer communications. All roles, offers, and commitments below are illustrative, not Alactic customer results. The table is a filled working artifact to adapt with a client, not a universal approval policy.

Deliverable and tierAuthorised boundaryReviewersRelease conditionEscalation triggers
Tier A: routine report commentary. Explain a decline in enquiries without claiming an unverified cause.Use the agreed reporting source, period, metric definitions, and neutral style. No new recommendations or commitments.Agency analyst checks figures and interpretation. Agency account lead checks the boundary.Checks are recorded; the account lead releases under the client's standing authorisation.Conflicting figures, changed definitions, unsupported causal claims, or recommendations that change client activity.
Tier B: changed offer. Replace a complimentary onboarding session with a complimentary account review.Draft within the existing campaign, audience, and unchanged eligibility conditions. Obtain targeted approval of replacement-offer details.Agency account lead checks campaign consistency. Client offer owner confirms unchanged eligibility and replacement-offer fulfilment details.Offer owner approves the exact changed wording; dependent assets are checked; the named campaign release owner authorises publication.Proposed eligibility changes require tier reassessment before release. Price or audience changes, uncertain capacity, or a new service promise also trigger reassessment.
Tier C: new customer-facing commitment. Promise a response to every new enquiry within two business hours.Drafting only until the client decides on the promise and supporting operating arrangements.Agency account lead reviews the complete output. Client service lead checks capacity and handling. Client sponsor makes the final decision within their authority.Questions are resolved; the sponsor approves the complete version; the named release owner confirms readiness.Coverage gaps, conflicting approvals, unclear ownership, or substantive changes after sign-off. Hold release for the relevant decision-maker.

Keep supporting evidence alongside each item: the source and standing authorisation for Tier A; before-and-after wording and focused approval for Tier B; the complete draft, operating assumptions, and final decision for Tier C.

The trade-off is deliberate. Tier A requires precise boundaries upfront. Tier B requires a decision that can be isolated. Tier C uses more client attention because the complete output and its operational implications need examination.

4. Work a targeted review from draft to release

In the hypothetical offer example, the agency proposes replacing the onboarding session with an account review. The campaign already exists, and its audience and eligibility conditions remain unchanged. The client offer owner must verify those unchanged conditions and confirm what the replacement includes, who will provide it, and whether the advertised availability is accurate.

This is different from asking the owner to approve new eligibility rules. If the owner proposes restricting the offer to a different customer group, the agency must reassess the tier before release. The original targeted question no longer describes the whole change.

The following filled packet can be reused by replacing its entries. Every detail is hypothetical.

Item and version
Account-review campaign, draft B.
Previous wording
Eligible account holders can request a complimentary onboarding session.
Proposed wording
Eligible account holders can request a complimentary account review covering their current setup and questions about existing services.
Decision requested
Confirm that existing eligibility conditions remain unchanged; the stated review scope is accurate; and the service team can fulfil requests during the planned campaign period.
Context included
The full campaign message, existing eligibility note, proposed offer description, and planned campaign period.
Dependent checks
Agency account lead checks the landing-page copy and follow-up message for references to the old onboarding session.
Release gate
Client offer owner explicitly approves draft B's changed fields. Agency completes dependent checks. Named campaign release owner authorises publication.
Hold condition
Any eligibility change, unresolved fulfilment question, or additional promise returns the item for tier reassessment.

Suppose the owner confirms the offer but says the proposed campaign period cannot be supported. The item remains held; approval of the offer concept is not approval of inaccurate availability. The agency revises the affected material and obtains confirmation of the revised version.

Targeted review is narrow, not blind. Include the surrounding headline or call to action if it changes the meaning of the offer. Move to full review when the decision cannot be isolated reliably.

5. Make approval specific and keep release separate

An approval record should identify the version, decision covered, approver, and any conditions. A positive reaction such as “looks good” may leave unresolved whether the reviewer checked fulfilment, wording, or only the overall direction.

Use distinct work states:

  • Prepared: the draft exists.
  • Checked: required factual and operational checks are complete.
  • Approved: the designated decision-maker has accepted the specified version and scope.
  • Released: the named owner has sent or published it.

These are recommended workflow states, not claims about a software feature. Track them in a shared record that reviewers can access rather than relying on an ambiguous conversation history.

Do not infer approval from a missed response date. Use an agreed substitute with the necessary authority, reschedule release, or hold the affected item. Decide substitutes when setting up the agreement, not during a last-minute deadline dispute.

A substantive post-approval edit reopens the affected decision. Correcting punctuation may remain an editorial check; changing what the offer includes requires renewed review. Record the replacement version and make clear which earlier approval it supersedes.

6. Give escalations an owner and a restart condition

An escalation needs an action, not just a warning label. Record what stops, who receives the question, and what evidence allows work to resume.

In the hypothetical report example, conflicting figures stop the disputed commentary. The analyst records the competing sources; the client reporting owner resolves which source applies; the account lead checks the corrected passage against Tier A. Unrelated sections need not stop unless the discrepancy undermines the wider report.

If a higher-tier change is discovered after release:

  1. Pause further use of the affected asset where the agency has authority to do so.
  2. Tell the client what was released, where it appeared, and which boundary was crossed.
  3. Ask the authorised client owner to decide on correction, withdrawal, or other follow-up.
  4. Check related versions and queued material for the same issue.
  5. Repair the rule or handoff that allowed the error.

Do not respond to every failure by permanently moving all work to full review. An unrecorded edit needs version control; a missed dependency needs a better dependency check. Broaden review where the incident shows that the original decision was too consequential to isolate.

7. Pilot the agreement and inspect mismatches

Negotiate the tiers with the client using recent deliverables. Before starting, agree named reviewers, substitutes, release owners, evidence requirements, and escalation routes. Leave unresolved categories awaiting a client decision rather than quietly treating them as routine.

Hypothetical pilot: Apply the agreement to the next ten deliverables in one service stream. Record each proposed tier and reason before work starts. At completion, capture review rounds, waiting time, escalations, and corrections after release. These are observations to collect, not promised improvements.

Look for specific mismatches:

  • Repeated Tier A escalations: the standing boundary may be too broad or the source unreliable.
  • Tier B reviewers repeatedly requesting the full campaign: the decision may not be separable from its context.
  • Full-review work waiting on unnamed owners: responsibility is unclear; adding another review step will not solve it.
  • Fast approvals followed by corrections: review speed may be masking incomplete checks.

At the next client operations meeting, compare one item from each tier with the table. Adjust the rule that failed, assign an owner to every unresolved boundary, and record when the revised agreement takes effect. The useful outcome is a repeatable decision about what can proceed, what needs a focused check, and what must wait.

Agency operationsClient approvalsReview workflowsQuality assurance

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