the short answer
Start by listing the tasks that eat the most hours, pick one repetitive, rule-based workflow, measure how it performs today, automate it with a human reviewing results, then compare the numbers after two to four weeks. Scale only what clearly saves time or wins revenue.
Week 1: find the bottleneck
- Ask each team member which tasks they repeat every day.
- Estimate hours per week for each task.
- Circle the ones that are high-volume and follow clear rules.
- Pick one. Just one.
Week 2: measure the baseline
Before changing anything, record how the workflow performs: time spent, response times, error rate, leads lost. Without a baseline you can't prove the AI helped — and you can't decide what to do next.
Week 3: automate with a human in the loop
Launch the automation in supervised mode: the AI drafts or acts, a person approves or spot-checks. Collect the cases it gets wrong and fix the instructions, data or rules behind them.
Week 4: compare and decide
- Compare against the baseline.
- Reduce supervision where accuracy is proven.
- Write down what you learned.
- Choose the next workflow.
Common mistakes
- Starting with a vague goal like “use AI more”.
- Automating a process nobody has agreed on.
- Skipping the baseline measurement.
- Giving AI access to everything on day one.
Want a partner for it?
Our digital transformation engagements follow exactly this rhythm — one workflow, measured, then the next. Book a discovery call and we'll help you pick the first one.