The handoff test for any client automation: Can the next person see the failure path without asking the builder? If not, the workflow is not production-ready. It is founder memory with a nicer UI.

You build the automation. Hand it over. Client does nothing for three days. The workflow works fine. The client does not know that. The moment they cannot see the logic is the moment the automation starts dying.

A property manager showed me their Saturday call log. 23 calls. 6 flagged urgent by the answering service. Monday morning they had three emergency plumber dispatches and two furious tenants. The service routed everything the same way. The urgent calls looked identical to the routine ones until someone listened to the voicemail backlog Monday morning. The fix: keyword routing that sends "flood," "no heat," "main break" straight to the on-call phone, everything else to voicemail. One day to build. No more Monday fires. What's the one call type that always ends up on your Monday list?

A founder told me support got calmer after they made ownership visible before asking AI to answer faster. They added one row above every ticket: customer value SLA clock refund risk owner next action proof it happened If a £420/month account is 2 hours from SLA breach, the workflow should put Sarah on the ticket before the customer chases. The reply draft can wait. Save this for the next time a client asks for an AI support agent. What field would you add before shipping it?

The easiest automation to sell is the one where the buyer can count the leak before the call. Failed payments work because the math is already there: 37 failed charges 11 over £250 6 active customers 1 owner before 3pm That is a workflow offer, not an AI ops retainer.

A SaaS onboarding workflow has one job in week one: get the customer to the first useful action before they forget why they signed up. The activation version tracks four things: 1. signup source 2. promised outcome 3. first action completed 4. owner if nothing happens in 24 hours If the workflow only sends welcome emails, it is admin. If it moves a new customer from "interested" to "I got the thing I paid for", it is revenue protection. Save this check for the next onboarding build: no first useful action = no activation workflow

Wade Foster's best Zapier lesson maps perfectly to AI workflows: boring setup wins. A sellable automation offer needs 5 lines before anyone opens n8n: 1. customer-visible leak 2. trigger event 3. value threshold 4. owner rule 5. proof that money moved Example: failed payments. Leak: 37 failed charges this month Trigger: Stripe failure Threshold: anything over £250 gets a person Owner: finance lead before 3pm Proof: recovered, written off, or escalated by Friday That is easier to sell than "AI billing agent." It gives the buyer a number, a boundary, and a reason to trust the workflow. Save this before pitching the next boring ops automation. If the offer cannot name the leak, owner, and proof, it is still a tool demo. What boring workflow would you package this way?

Most builders test until the workflow runs clean. Then they ship. Then they cross their fingers. The four checks before calling an automation done: 1. Silent failure If this breaks and nobody gets pinged, the monitoring is not working. Not "the test passed". Not "it ran in staging". Who gets a notification when it fails in production? 2. Owner Who fixes this at 2am on a Friday? Not "ops". Not "the team". A named person with a phone number. 3. Detection speed If the answer is "when the client complains", the monitoring is not working. The test is not whether you have a monitoring dashboard. It is whether the right person knows within a set time. 4. What the customer sees A stalled quote. A wrong invoice. A booking that never arrived. That gap is a churn risk nobody is accounting for. The workflow that keeps the client in the loop is not a nice-to-have. It is the retention mechanism nobody is building. The useful test before launch: Can you name a specific human with a phone number for each of those four cases? If not, you have a demo with a nicer interface.