4 min read
Automation is often proposed as the answer to slow or manual work. Sometimes it is exactly the right answer.
It can also make an outdated operating model run faster without making it better.
A redundant approval that takes two days can be routed in two seconds. A bad handoff can move instantly. An unreliable data source can feed a dashboard continuously. The technology works, but the business problem remains.
Automation multiplies the design around it
Automating a step does not answer whether the step should exist.
If an approval no longer maps to a meaningful risk decision, faster routing does not create value. If a report requires reconciliation because upstream data ownership is unclear, automating report generation does not fix the information problem. If employees re-enter data because two systems represent the customer differently, an integration may only move the inconsistency faster.
The strategic question comes first.
What should the work look like before it is automated?
AI makes this easier to get wrong
Modern AI tools lower the cost of creating automation and internal software. That is useful, but it also means organizations can scale a weak process before anyone pauses to question it.
The better approach is to separate three decisions.
First, should the work exist in its current form?
Second, if the work is necessary, what is the simplest way it should operate?
Third, which parts should be automated, supported by AI, integrated, or left to people?
Those decisions are related, but they are not the same.
Most problems have more than one type of answer
A useful strategy may combine several moves.
Simplify. Remove unnecessary work, approvals, or exceptions.
Redesign. Change ownership, decision rights, or the flow of work.
Connect. Fix the way systems and data interact.
Automate. Remove genuinely repetitive manual work.
Equip. Give people better information or decision support.
Build. Create a capability that does not exist when the business case justifies owning it.
Automation earns its place inside that strategy. It should not define the strategy before the problem is understood.
The goal is not to automate more work.
The goal is to create a better system and automate the parts where technology materially improves it.