Find the right work to automate
Map a repeated task, choose a useful goal, and decide where AI belongs. This module has 3 lessons and takes about 1h 50m. Each lesson has a practice exercise that runs real Python in your browser and a short quiz.
Start module 01 in the Learning LabTriggers, steps, and outcomes
Return ["validate", "draft", "review"] for request.created, and an empty list for every other event.
An automation starts when an event happens. A form submission is a trigger. Checking required fields is a step. Creating a draft ticket is an outcome. Write these down before choosing a tool.
A Python dictionary stores named values. Read event["type"] to find the event type. A list keeps steps in order. This lesson turns a recognised event into a short plan.
A clear workflow includes a stop path. Unknown events should not quietly trigger a real action.
See it in action
def solve(event_type):
if event_type == "request.created":
return ["validate", "draft", "review"]
return []
print(solve("request.created"))
A clear workflow includes a stop path. Unknown events should not quietly trigger a real action.
Rules, AI, and human decisions
Return human review for an external effect, otherwise rule when exact_rule is true, otherwise AI draft.
Use a rule when the answer is exact, such as checking whether a field is empty. AI can help with messy language, such as suggesting the topic of a message. A person should own decisions that need judgement or approval.
An if statement selects a branch. Put the highest priority condition first. Here a task with an external effect needs review even if an exact rule is available.
AI adds uncertainty. Compare it with a simple rule before adding the cost and review work of a model.
See it in action
def solve(exact_rule, external_effect):
if external_effect:
return "human review"
if exact_rule:
return "rule"
return "AI draft"
print(solve(True, False))
AI adds uncertainty. Compare it with a simple rule before adding the cost and review work of a model.
Practical: Measure a useful starting point
Return weekly minutes saved as cases multiplied by manual minutes minus review minutes. Keep negative results.
A baseline records the current process. Count cases per week and minutes per case. Estimate the new review time as well as the time that automation saves.
For 20 cases taking 6 minutes each, with 2 minutes of review per automated case, the estimated saving is 20 * (6 - 2) = 80 minutes. A negative result means the new process takes longer.
Time saved is an estimate until you measure real work. Track errors and review effort alongside speed.
Your experiment plan
Use the browser editor or the notebook. All records and AI responses are fictional. No network calls, credentials, or real messages are used.
- Map a fictional workshop enquiry from arrival to a reviewed draft.
- Run the example and predict its result before changing it.
- Complete the function in Practice and run the checks. Add your own normal, empty, and failure cases.
- Double review time and compare the estimated saving.
- Save the notebook with your observations and explain one limitation of the simulation.
What to hand in
- A workflow map, a baseline table, and a measurable pilot goal.
- A test table with inputs, expected outputs, actual outputs, and one revision.
Reflect on your results
When would you keep the manual process?
The automatic exercise checks the core code behaviour. Review observations and written deliverables yourself or with a mentor.
See it in action
def solve(cases, manual_minutes, review_minutes):
return cases * (manual_minutes - review_minutes)
print(solve(20, 6, 2))
Time saved is an estimate until you measure real work. Track errors and review effort alongside speed.
Free, open and run by a nonprofit.
This page is part of AI Automation, a free course of 20 modules and 60 lessons in the engineering.fit( ) Learning Lab. The Learning Lab is built and run by engineering.fit( ), a nonprofit founded by Abdullateef Opeyemi Bakare that helps African engineering students gain research experience. You do not need an account or pay a fee to learn.