Automation & AIModule 2: Automation building blocksLesson 6 of 15
Course progress33%
15 min lesson · Updated August 2026
What is automation?
Automation is a designed system that performs repeatable work when defined events and rules occur; it may use ordinary software, AI or both, but still needs ownership, exceptions, monitoring and recovery.
What you will learn
By the end, you will understand:
Identify stable repeatable work suitable for automation
Separate deterministic rules from AI-assisted judgment
Calculate value using quality, risk and maintenance—not time saved alone
Visual explainer
See the idea clearly.
✓
A repeated manual process is mapped, standardized and split into deterministic automated steps, human decisions and exception paths before a monitored workflow runs.
Automation executes a process
A process becomes automatable when inputs, rules, actions and outcomes can be described. A scheduler, spreadsheet formula, email rule or integration can automate without AI.
If the human process is unclear or contradictory, automation often makes the confusion faster and harder to see. Map and simplify first.
Three broad kinds
Kind
Example
Deterministic
When a valid paid invoice arrives, mark its matching record paid. Same valid input/rule should produce the same result.
AI-assisted
Classify a message or draft a summary, then apply validation/human review because output can vary.
Human decision with automation support
System gathers evidence and recommends options; an authorized person decides.
Good candidates have stable structure
Repeated often enough
Clear start/end
Known inputs
Explicit rules
Manageable exceptions
Reliable systems/API
Low or controlled consequence
Measurable current baseline
Named owner
Manual fallback
Map before building
01
Observe current process
02
Remove unnecessary step
03
Define valid input
04
Write rules/decisions
05
List exceptions
06
Assign human ownership
07
Choose system of record
08
Automate smallest safe path
09
Test failures
10
Monitor and improve
Value is more than hours saved
Benefit
Measure
Cost/risk
Speed
Cycle or response time with service quality.
Consistency
Error/rework rate and policy adherence.
Capacity
Useful volume handled without missed work.
Experience
Customer/staff task success and satisfaction context.
Build, licences, usage, maintenance, incidents and review time.
Operation
Automations become production systems
APIs change, credentials expire, employees leave, fields are renamed and edge cases appear. Assign an owner, version changes, monitor errors and test periodically.
A workflow that silently stopped months ago is worse than a visible manual queue. Design alerts and reconciliation from the beginning.
Real-world example
Example: invoice reminders without AI
Example
A finance system checks due dates daily and sends an approved reminder when an unpaid invoice reaches a threshold. Disputes, large accounts and failed deliveries route to a person. The value comes from reliable rules, not an AI label.
Try this
Score one repeated task
Record frequency, average time, error rate, rules, exceptions, systems, consequence and fallback. Decide whether to keep manual, standardize, automate deterministically or add AI assistance.
Common questions
Questions beginners ask.
What is automation?
A designed system that performs repeatable actions when defined events and rules occur.
Does automation require AI?
No. Schedules, formulas, rules and integrations automate many tasks without machine learning.
What is deterministic automation?
A workflow whose explicit rules produce predictable outputs for valid inputs.
What should be automated first?
A stable, repeated, measurable process with clear inputs, rules, owner and manageable consequence.
Can automation remove every human step?
Not safely in many workflows. Exceptions, approvals, empathy and accountable decisions often remain human.
What is an exception path?
The route for inputs or failures that do not meet the normal automation rules.
How is automation value measured?
Compare time, quality, capacity and experience benefits with build, usage, review, maintenance and incident cost.
Who owns an automation?
A named business/technical owner should be accountable for purpose, access, monitoring and change.
