Back to all articlesCustomer onboarding

Decide what must be ready before customer onboarding starts

Alactic Editorial10 min read

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

The kickoff is tomorrow, but the implementation lead still cannot open the customer workspace. The customer has sent a sample file without explaining its fields, and the person who can approve the workflow is not attending. Keeping the meeting does not make the project ready.

Before onboarding implementation starts, require evidence that the team can complete the first agreed work package: the necessary access works, essential customer inputs are usable, and someone with decision authority is available. Start when those prerequisites are met. Start conditionally only when a smaller, useful work package can proceed without the missing items. Otherwise, defer implementation and use any retained meeting to resolve readiness—not to imply delivery has begun.

A customer onboarding readiness checklist should support that decision, not produce a reassuring completion percentage. The practical test is: What useful output can we finish with what is available now?

1. Define the work that the start gate protects

Readiness is relative to a specific piece of work. A team might be ready to map a workflow but not ready to configure it, test it with customer data, or release it to users. A single project-wide ready/not-ready label hides those distinctions.

Before collecting prerequisites, write a short start statement containing:

  • The first output: what the team will produce, not just which meetings it will hold.
  • The working boundary: which environment, process, users, and data are included.
  • The acceptance decision: who will decide whether the output is usable, and what they will check.
  • The exclusions: work that is explicitly not authorised to begin.

For example, a hypothetical first work package might be: “Produce a reviewed onboarding workflow and a draft field map using synthetic records. Do not import live records or activate customer-facing activity.” Its prerequisites differ from those for a live implementation.

Keep the boundary small enough to assess honestly, but large enough to produce something useful. Shrinking the first package to an introductory call merely to claim a start is not a meaningful readiness decision.

2. Separate blockers from items that can follow

An item is blocking when its absence prevents the agreed output, makes the work unauthorised, or creates an unacceptable likelihood of unusable work. An item is nonblocking when the team can complete the current package without it and has identified when it will become necessary.

Use these questions for each proposed prerequisite:

  1. Which task depends on this item?
  2. Can that task be completed and reviewed without it?
  3. Would an assumption be reversible, or would it shape substantial downstream work?
  4. What is the latest point at which the item must be available?

Access is not automatically a blocker for every activity. Missing production access may block a live import while leaving workflow review possible. Conversely, a missing decision-maker may block useful design work even when every account is active.

Separate labels for blocking now and blocking later prevent a common mistake: treating “not needed today” as “not important.” A later prerequisite still needs an owner, evidence standard, and dependency date.

Do not let a large number of completed minor items outweigh a missing critical one. Finished branding preferences cannot compensate for the absence of permission to use the intended data.

3. Require observable evidence and named responsibility

“Access requested” is evidence of a request, not evidence of usable access. “File sent” does not establish that the file contains the agreed fields. “Sponsor invited” does not mean an authorised reviewer is available.

Each checklist entry should record the requirement, its blocking status, the evidence that satisfies it, the person responsible for supplying it, and the person who checks it. When something is missing, add a specific next action and review point.

For access, evidence might be confirmation that the assigned implementer can perform the required task in the intended environment. Do not place passwords, tokens, or sensitive records in the checklist. Record the result and reference the approved location for supporting material.

For customer inputs, evidence should establish usability. A sample needs enough context to interpret it; a workflow description needs unresolved choices identified; an acceptance criterion needs someone empowered to apply it.

Assign one accountable person to each missing item. A department name or shared inbox can be a contact route, but it does not resolve who must act. The delivery lead should verify readiness for the planned work rather than relying solely on the supplier's status update.

4. Filled hypothetical readiness checklist

Fictional worked example: All people, timings, and project details below are hypothetical. A customer team is preparing to implement a customer-intake workflow. The original opening package included workflow design, field mapping, and a trial import. The readiness review finds that a trial import cannot yet proceed.

The proposed conditional package is narrower: prepare a reviewed workflow and draft field map using synthetic records only. Maya, the delivery lead, owns the readiness decision. Elena, the customer's operations lead, owns customer-side completion.

Checklist at the readiness review

  • Agreed first output — blocking now; ready. Evidence: Elena and Maya have confirmed a written scope covering workflow review and a draft field map, with live-data work excluded. Responsible: Maya prepares the scope; Elena accepts it. Verification: both confirmations are recorded in the project decision log.
  • Access to the design workspace — blocking now; ready. Evidence: Priya, the assigned implementer, has opened the correct workspace and confirmed she can create and edit the planned design documents. Responsible: Omar, the customer administrator. Verification: Priya records the task check, without sharing credentials.
  • Documented intake process — blocking now; ready. Evidence: Elena has supplied the current steps, routing rules, and a list of unresolved choices. Responsible: Elena. Verification: Priya confirms the material is sufficient to draft the workflow and flags the open routing decision.
  • Authorised workflow reviewer — blocking now; ready. Evidence: Elena is confirmed as the reviewer for intake routing and has accepted the review appointment. Responsible: Elena. Verification: Maya records the limits of Elena's authority; budget and production-release decisions remain outside it.
  • Representative field definitions — blocking now; ready for the reduced package. Evidence: the team has a field dictionary and synthetic records showing expected structures. Responsible: Jonah, the customer data owner. Verification: Priya confirms these support a draft map, but not validation against real records.
  • Usable live-data export — blocking for the original trial import; not ready. Evidence: no export has been accepted for implementation use. Responsible: Jonah. Next action: provide the agreed sample through the approved sharing route before the next readiness review. Verification: Priya will check completeness and interpretation before import work is considered.
  • Permission for live-data use — blocking for live-data work; not ready. Evidence: the customer's designated approver has not confirmed the proposed use. Responsible: Elena coordinates with that approver. Next action: obtain an explicit decision before any live records are used. Verification: Maya checks the recorded authorisation and its scope.
  • Training attendee list — nonblocking now; open. Evidence: Elena has named the team coordinator, but the attendee list is unfinished. Responsible: Elena. Dependency: the list must be ready before training is scheduled; it is not needed to review the workflow.
  • Final internal guide styling — nonblocking now; open. Evidence: a plain working format is accepted for design review. Responsible: Theo, the customer enablement lead. Dependency: styling is required before the guide is distributed, not before the draft workflow is reviewed.

This checklist contains no overall readiness score. It shows why the original package is blocked and why a reduced package may still be worthwhile.

5. Make the conditional-start decision explicit

In the hypothetical example, Maya and Elena choose start conditionally. The workflow and draft field map have standalone value, all prerequisites for that reduced package are satisfied, and the missing live-data items can remain outside the work boundary.

They record the decision as follows:

Proceed with workflow review and a draft field map using synthetic records only. Do not import live records, validate against live data, or activate the workflow. Maya owns this boundary. Jonah owns the missing export, and Elena owns obtaining the live-data decision. Review readiness at the end of the design package before authorising further work. The trial-import date remains unconfirmed.

The condition is not “start now and hope the customer catches up.” It is a bounded authorisation with a clear output and stopping point.

There is a trade-off. Synthetic records allow design to proceed, but real data may expose exceptions that change the map. The team should mark affected choices as provisional and avoid representing the draft as import-ready. If the expected redesign would consume most of the package's value, deferring would be more sensible.

Conditional starts are appropriate only when the useful work can be separated from the blocked work. If people cannot maintain that separation in practice, the written boundary is not enough.

6. Use three decisions, not a vague amber status

A status colour does not tell the team what it may do. End the readiness review with one of three operational decisions:

  • Start: all blocking prerequisites for the agreed package are verified. Open nonblocking items have owners and dependency points.
  • Start conditionally: a reduced package is useful and ready. Its exclusions, provisional assumptions, review point, and stop conditions are recorded and accepted.
  • Defer: no useful authorised package can proceed, or the risk of rework makes a reduced package poor value. Record the exact evidence needed to reconsider.

The delivery lead should make the implementation-readiness decision with the customer's accountable counterpart. Commercial urgency can explain why a date matters; it cannot substitute for missing evidence.

If implementation is deferred, the booked kickoff may still serve a purpose. Rename it as a readiness-resolution session, invite the people who can clear the blockers, and keep its output specific. Do not ask the delivery team to perform implementation work merely to make the meeting feel productive.

7. Recover when a prerequisite slips or proves false

Readiness evidence can become stale. An account may stop working, an accepted input may turn out to omit a critical case, or the confirmed reviewer may become unavailable.

When that happens, pause the affected work rather than treating the original start decision as permanent permission. Identify which outputs remain valid, which assumptions are now uncertain, and which downstream tasks depend on them. Then reapply the same start, conditional-start, or defer decision to the remaining package.

For the hypothetical project, if Jonah's export is still unavailable at the review point, the team can finish and accept the synthetic-data design package. It should not drift into an improvised import exercise. Maya records what is complete, leaves import work blocked, and agrees a new review trigger tied to evidence rather than another unsupported date.

If the real export later contradicts the field dictionary, reopen the affected mapping decisions. Distinguish new information from an execution mistake so the customer can understand why the plan changed. Explain the impact before committing to revised delivery dates.

8. Run the gate before the next kickoff

For the next scheduled onboarding, write the first work package in one paragraph and list its prerequisites. Ask the named owners for evidence before the kickoff, leaving enough time to change the meeting's purpose if necessary.

At the readiness review, record the decision, authorised scope, unresolved blockers, responsible people, and next review trigger. Send that record to everyone expected to begin work. A conditional boundary that only the project lead knows about will not reliably guide the team.

After the first package, measure whether the gate helped: count tasks stopped by missing prerequisites, identify work redone because an input was unusable, and compare the promised output with what was actually accepted. Keep these as operating observations, not a blended readiness score.

The immediate next action is concrete: before the next kickoff proceeds, make every blocking item for its agreed work package either evidence-backed or explicitly excluded through a reduced-scope decision. If neither is possible, defer implementation.

Customer onboardingImplementation readinessDelivery operationsChecklists

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