8–9 points
Ready for a bounded pilot
Your process has enough structure to justify a small test. Freeze the scope, use the smallest safe sample, define exceptions and acceptance criteria, then test before expanding.
Free automation self-check
Is this workflow ready to automate?
Before buying software or a build, check whether one real process is stable enough for a small controlled pilot. This checklist is useful even if you never contact PERSIK.
Score one workflow, not your whole business
Give yourself one point for each statement that is already true. If you are unsure, count it as zero and use the remediation note below. This is a practical screening tool, not a guarantee that automation will work or save a particular amount.
- Repetition: the same process happens often enough to describe step by step.
- Consistent inputs: the key fields, files or messages follow recognisable patterns.
- Clear output: you can show what a correct finished result looks like.
- Permitted access: the systems provide an API, export/import, email/file route or another permitted way to move the required data.
- Manageable exceptions: unusual cases can be identified and handed to a person instead of guessed.
- Named owner: one person can approve the rules, access and acceptance criteria.
- Minimum data: the first test can use public, synthetic, redacted or tightly minimised data rather than a full production dataset.
- Measurable baseline: you can compare the pilot with the current process using a relevant measure such as time, rework, missed items or turnaround.
- Fallback and small scope: the manual process can continue if the pilot stops, and the first test can be limited to one trigger, one workflow and one defined output.
6–7 points
Prepare first
Do not buy a large build yet. Fix the missing conditions below, then rescore the same workflow. A short process map or cleaner sample may be all that is needed.
0–5 points
Not ready yet is a useful result
If the rules change every time, access is unavailable, no owner can approve the process, or the only test requires broad sensitive-data access, automation is likely to create more risk than value. Improve the process first and revisit automation later.
How to fix a weak score
- Low repetition: wait and collect several real examples before designing automation.
- Inconsistent inputs: standardise a form, template, filename or required fields.
- Unclear output: save one accepted example and define what makes it correct.
- No usable access: check native exports, APIs or approved integrations before adding browser automation.
- Too many exceptions: list exception types and keep them in a human review queue.
- No owner: assign one person to approve rules and changes.
- Too much data: reduce to the exact rows, fields or redacted sample needed to test the rule.
- No baseline: measure the current process for a short period before building.
- No fallback: keep the existing manual path until the pilot is independently accepted.
Stop conditions
Do not send passwords, MFA secrets, API keys or private keys through an enquiry form or email. Do not use a small pilot to collect a full database, unrestricted mailbox or higher-risk personal data when a public, synthetic, redacted or smaller sample can prove the same rule.
Keep legal, medical, financial, employment and other material decisions with appropriate human review. Do not bypass CAPTCHAs, access controls or platform restrictions to make a workflow possible.
A sensible first pilot has five written lines
- Trigger: what starts the workflow?
- Minimum input: what exact data or file is required?
- Output: what should be produced?
- Exception: when should a person take over?
- Acceptance: what observable result will count as success?
If those five lines cannot be written without major uncertainty, the next step is process clarification, not a purchase.
If your workflow scored 8–9
Use the fit check as your brief
Send only a plain-language description of the repetitive task, how often it happens, the systems involved and the finished result you want. Do not send credentials or unnecessary sensitive data. PERSIK can then assess whether one bounded Automation Studio pilot is sensible.