The 80/20 of business automation: where to start
Most companies over-engineer their first automation. After mapping hundreds of workflows, here's how we identify the highest-leverage starting point every time.
Most businesses come to us wanting to automate everything at once. They've seen what automation can do - reduce headcount, eliminate errors, run 24/7 without a salary - and they want all of it. The problem is that starting too broad is how automation projects die. They sprawl, get complex, lose executive support, and get shelved.
After mapping hundreds of business workflows, we've developed a simple filter that identifies the highest-leverage automation in any business within a single workshop.
The four questions that find your first automation
- What tasks does your team do more than 5 times a day that follow the same steps every time?
- Where does data get manually copied from one system to another?
- What triggers an action in your business that someone has to manually watch for?
- What reports or summaries does someone build manually on a schedule?
The answers to these questions - not the answers to 'where would AI help' or 'what could we modernise' - reveal the processes with the highest automation ROI. They're repetitive, rule-based, high-frequency, and well-understood. They don't require AI. They require reliable execution.
The three automations that pay for everything else
1. The data sync
Almost every business has at least one place where a human manually copies data between two systems that should talk to each other. CRM to ERP. Form submission to spreadsheet. Order management to accounting. These data syncs are high-frequency, low-complexity, and automation-perfect. A reliable sync automation typically frees 5-15 hours per week from day one.
2. The notification and escalation workflow
Someone in your business is watching for things. A payment going overdue. A support ticket sitting unresponded for too long. An inventory level hitting a threshold. These watches-and-notifies are perfect automation candidates - they're time-sensitive, rule-based, and currently dependent on a human remembering to check.
3. The recurring report
Weekly sales reports, monthly reconciliations, daily pipeline snapshots - these are typically built by someone pulling data from multiple sources, formatting it in a spreadsheet, and emailing it. Full automation candidate. We've never encountered a recurring report that couldn't be automated, and the average time saving is 3-6 hours per report per week.
What not to automate first
Don't start with processes that require judgement, exceptions, or significant human reasoning - even if they consume the most time. These automations are harder to build, have lower reliability, and will generate edge cases that require ongoing attention. Start with boring, reliable, high-frequency wins. Build confidence in automation within your organisation before tackling the complex stuff.
The question isn't 'can we automate this?' - almost anything can be automated. The question is 'what's the ratio of implementation cost to ongoing time saved?' A 2-week build that saves 10 hours per week pays back in 2 weeks and generates ROI indefinitely.
Measuring what you built
Before automating anything, document the baseline: how long does this take, how often does it happen, what's the error rate. Measure the same metrics 30 days after automation. This baseline-vs-actual comparison is what builds internal support for more automation investment - and it's what shows leadership that the project was worth doing.
Find out which of your workflows to automate first
15 questions about your current manual processes. Instant score showing which workflows are costing you the most time, and which to tackle first.
Take the workflow auditFrequently asked questions
What types of tasks should I automate first?
Start with the ones that are repetitive, rule-based, and high-frequency. Data syncs between two systems, recurring reports, and notification triggers are the three categories that pay off fastest. Avoid starting with processes that require human judgement or frequent exceptions. Those automations are harder to build, break more often, and deliver less return in the short term.
How long does a typical business automation take to build?
For a focused, well-scoped automation like a data sync or a recurring report, anywhere from a few days to two to three weeks depending on the tools involved and the state of the data. The bigger risk is not the build time but the scoping. Teams that try to automate too broadly in one go consistently run over schedule. Pick one process, build it properly, then expand.
Do I need a technical team to automate my business processes?
Not always. Tools like Zapier and Make let non-technical people automate simple, linear workflows without writing code. For anything complex or conditional, or where data has to stay inside your own infrastructure, a developer or automation specialist will save you significant time and frustration. The right approach depends on the complexity of the process and who will maintain it once it is live.
How do I measure whether an automation was worth building?
Document the baseline before you build: how long the task takes, how often it runs, what the error rate is. Then measure the same metrics 30 days after the automation is live. That comparison is what makes the return on investment clear and concrete, and it is what gets leadership on board for automating the next process.
