Automation project discovery questions that expose hidden work.
Good discovery is not a generic call about budget and deadlines. It should uncover the technical, operational, and support conditions that can change effort, risk, and price.
Updated 30 August 2026Use the free calculator →
How to use this checklist
Start with the desired outcome and current process. Then work through systems, data, exceptions, testing, ownership, monitoring, and support. Record unknowns instead of silently making optimistic assumptions.
You do not need every answer before the first conversation ends. You do need a commercial response for important unknowns: clarify them, add a justified contingency, narrow the scope, or propose a paid discovery phase.
Outcome and process
- What business result should the automation produce?
- What triggers the process, and what marks a successful finish?
- Which steps, decisions, or approvals still require a person?
- How is the process handled today when something goes wrong?
Systems and access
- Which systems, accounts, workspaces, and environments are involved?
- Are the required credentials, API scopes, and test accounts available?
- Are any connections using custom APIs, webhooks, or undocumented endpoints?
- What rate limits, usage limits, or vendor restrictions apply?
Data and exceptions
- What fields are required, optional, or frequently missing?
- Who owns data quality and corrects incomplete records?
- How many records or events are expected on a normal day and at peak times?
- Where should failed, duplicate, late, or unexpected items go?
Testing and acceptance
- What test cases must pass before launch?
- Who supplies representative test data?
- Who gives final approval, and how quickly can they review?
- What measurable conditions mean the project is accepted?
Monitoring and support
- Who receives failure alerts and investigates them?
- What logs or audit history must be retained?
- How long is post-launch support included?
- Which future changes are maintenance, and which require a new quote?
Turn answers into scope and pricing decisions
- Missing access can delay delivery or justify a readiness requirement before the schedule starts.
- Unclear data quality can add validation, cleanup, exception handling, and manual review.
- High failure impact can require stronger testing, monitoring, fallback procedures, and support.
- Frequent process changes can change the delivery model from a one-off build to ongoing maintenance.
- Undefined acceptance can create endless revision and should be resolved before a fixed price is promised.
Get questions matched to a real project.
The free review selects follow-up questions from the risk in your answers and shows how those answers affect effort and your pricing floor.
Review an automation project →