Most lists of intelligent automation examples describe the model and stop there. The ones that reach production share a structure, and the model is the smallest part of it.
We wrote up seven working examples across finance, customer operations, field work and sales in Intelligent Automation Use Cases. This post pulls out the engineering pattern they share.
The pipeline
Robotic process automation moves data along a fixed path. Intelligent automation adds a judgement step on input that varies. In code terms, the flow looks like this:
def handle(case):
fields = model.extract(case.document) # read
record = system_of_record.lookup(fields) # check
result = rules.compare(fields, record) # decide
if result.clean:
act(case, result) # post, pay, reply
else:
review_queue.push(case, reason=result.conflicts,
owner=assign_person(case))
audit_log.write(case, fields, result)
The else branch is the product. Clean cases are cheap. The share that lands in the review queue decides whether the project pays back.
Why the check step matters
Invoice capture works because the purchase order already exists. A wrong extraction fails the match and goes to a clerk. Nobody pays it.
Claims intake checks against policy cover. Identity checks compare against the application and screening lists.
Sales account research is checked against Companies House for UK firms. Every example that works has a record to check against. Projects without one tend to stall.
Checklist before you write any model code
- [ ] Name the record the decision is checked against.
- [ ] Count one week of real cases: volume, clean share, exception share, minutes per case.
- [ ] Design the review screen. Who sees the doubtful case, what they see, what one click does.
- [ ] Decide what gets logged: input, model output, human decision and who made it.
- [ ] Repeat the same count after launch and compare.
What this looked like on a real build
On an inspection platform for work at height anchor systems, readings arrive from connected test devices with timestamps and GPS. The system assembles a draft report and certificate. The engineer still checks and signs each one. The reporting cycle came down by 70 percent, and most of that gain came from capturing data in a form a machine could assemble.
The full set of seven examples, with the selection table, is in the original article on dev-station.tech.

Top comments (0)