Before Automating a Broken Workflow: What Should the Founder Decide?

Before automating a broken workflow, decide what result it should produce, who owns the decisions, and what happens when a case falls outside the ordinary rules. Test the revised process manually on real examples before paying for a build. If the team can apply the same rules independently and the remaining work is repetitive, a small automation pilot may be worth scoping.

Find out where the work actually stops

Pick a recent case that stalled and trace it from request to outcome. Look for the first point at which someone needed information, an owner or an approval they could not find. The Lean Enterprise Institute’s guide to value stream mapping distinguishes documenting the current flow from designing a future one. For a lean startup, a short map of one case can be enough to start that discussion.

Imagine enquiries arrive through a form and a shared inbox. A coordinator forwards both to sales, but neither message says who should respond to an existing customer. Automating the forwarding step could speed up the handoff while leaving the ownership question open. Decide how to identify an existing customer and who owns that case before asking a provider to build routing.

Make three decisions before requesting a quote

Decision What to agree Evidence to give a provider
What counts as complete? A named owner can see the request, source and next action in the agreed system. One example of a completed record, with sensitive details removed.
Who can decide? The process owner approves routine routing rules and names the people who resolve exceptions. A short decision table with the owner for each case.
What happens when the rule cannot run? Send missing or conflicting inputs to a named review queue and specify how the team continues during an outage. A test case with incomplete data and a documented manual route.

For a broader inventory of inputs, handoffs and approvals, use the Web Morphix guide to reviewing startup systems before buying automation. For this decision, narrow the inventory to one process and write down the rule that caused the most recent delay.

Run the repaired workflow by hand

Ask two teammates to apply the proposed rules to the same small set of representative cases. Include an ordinary enquiry, an existing customer and a request with missing information. Compare the intended owner, next action and reason for each decision. If teammates disagree, find out whether the rule is unclear or the required information is unavailable. Resolve that gap or keep a deliberate review step.

Look for repeated copying, matching or notification work left after the process is repaired. Those steps are good candidates for a limited pilot when the inputs are reliable. Keep unusual approvals and unclear cases visible to a person. If the process rarely runs or the rules change every week, continued manual handling can be the better use of the team’s time.

Buy a pilot with a clear finish line

Ask for one trigger, a defined set of outcomes, tests for the ordinary and exception cases, and a named person who can pause the workflow. Agree how to check the downstream record, how to spot a duplicate action, and who approves wider use. Price the rule design separately if key decisions are still open.

Bring the decision table and redacted examples to Web Morphix’s CRM and automation support team if you want help assessing a workflow pilot. Ask what the team can test with the tools you already use and where it expects a decision from you.

Frequently asked questions

Can I automate part of a workflow while the rest is being fixed?

Yes. Choose a stable step with a clear input and result. Hold ambiguous cases for review, and confirm how the pilot interacts with the manual process.

What if the rules are clear but the team still misses handoffs?

Check whether people have the information and time to act. If the issue is repetitive transfer or notification, test that step. If work is waiting on authority or capacity, address that first.

Who should sign off on the pilot?

Name the person responsible for the real operating outcome. Have them review both successful runs and exceptions before approving wider use.

Share This Article

Want to Get Latest Update?