How to automate your company’s processes: where to start

In almost every company there are people copying data from one place to another: from email to the ERP, from a PDF to a spreadsheet, from the spreadsheet to the CRM. It is necessary work, but it adds no value, it is easy to get wrong and nobody enjoys it. Automating it well frees up hours every week.

How to find what to automate

• Ask your team which task they repeat most each week and which one seems most pointless. They usually know by heart.
• Look for the points where information changes tool: every time someone copies and pastes between two programs there is a candidate.
• Focus on what has clear rules: if a new person can do it by following a list of steps, a machine can too.
• Look at where errors and delays pile up: orders keyed in wrong, unrecorded invoices, notifications that don’t go out.

Work out whether it pays off

The basic sum is simple: hours per week the team spends × cost per hour × weeks per year. Compare it with what it costs to automate and to maintain the automation. Add what doesn’t show up in that sum: fewer errors, faster replies to customers and people working on tasks that do need judgement. Start with the processes where the saving is clear and the risk is low.

The tools, from simplest to most powerful

• Features your software already has: email rules, CRM workflows, scheduled imports. The first thing to check, because sometimes it is already paid for.
• Connectors between applications: services that link two tools with rules like “when this happens, do that”. Quick to set up for simple flows.
• Custom integrations: when connectors aren’t enough, volumes are high or the process has its own logic.
• Applied AI: agents that read documents, sort emails, extract data from invoices or draft replies. Useful where a person used to be needed because the information didn’t come structured.

Mistakes worth avoiding

• Automating a badly designed process: simplify first, then automate.
• Not planning for what happens when something fails. Every automation needs alerts and a manual fallback.
• Leaving it without an owner. Someone on the team has to know what it does and whom to tell if something changes.
• Trying to automate everything at once. A well-measured pilot is more convincing than a big project.

How we do it

We start with a two-week assessment to map processes and estimate the saving of each initiative, prioritise by impact and effort, and get the first measured pilots running between weeks 4 and 8. The initial 45-minute session is free.

Work out what manual work costs you today