Use this checklist to prepare a useful conversation with your development partner. Ask for a recent example or a demonstration where practical. The purpose is to understand the work and responsibilities together.
For each section, record the answer, the evidence, an unresolved question, and the person who owns the next action. A missing answer is a reason to investigate; it is not, by itself, proof of poor work.
1. The work and its purpose
- Can you connect the current priorities to a business or customer objective?
- Who can change scope, and how are those changes reflected in estimates and dates?
- What must be decided before implementation can proceed?
- Which recent feature spent the most time waiting or being reworked, and why?
Ask to see: A current roadmap and one recent feature followed from request through release.
2. Estimates and tradeoffs
- What assumptions does the estimate depend on?
- What is included in discovery, implementation, testing, deployment, and support?
- Which unknowns could change the estimate, and how will they be resolved?
- Is there a smaller useful scope the team has considered?
Ask to see: One proposal with its assumptions, exclusions, dependencies, and acceptance criteria.
3. Ownership and access
- Which organization controls the repositories, cloud accounts, domain, and important provider accounts?
- Who can release and troubleshoot the application?
- What access would need to change if a person or supplier left?
- Where are system knowledge and operating instructions maintained?
Ask to see: An account and responsibility inventory. Keep credentials out of the review document.
4. Quality and release confidence
- How are changes reviewed and tested before production?
- Which important business workflows have repeatable checks?
- How does the team respond if a release creates a problem?
- What work repeatedly escapes review or returns for correction?
Ask to see: A recent change with its review, relevant checks, and deployment record.
5. Operating the product
- Who notices a failed background job or a broken integration?
- How are urgent incidents distinguished from normal support requests?
- What evidence supports the current performance or scaling concern?
- When was a recovery procedure last exercised, and what did it demonstrate?
Ask to see: A recent incident or failure followed through detection, investigation, recovery, and follow-up.
6. Decide what to do next
Choose a short list of actions with owners and a review date. Separate an unanswered business decision from an architecture problem, a delivery practice that needs improvement, and a genuine capacity gap.
If the relationship needs an independent technical counterpart, Technical Advisory can help you evaluate the evidence and work through the next decisions with your existing agency.