Which business processes should you automate first?

The biggest mistake with automation is not picking the wrong tool — it is starting in the wrong place. Automate the wrong thing and you save a little time for a lot of effort, disruption, and budget. Automate the right thing first and you save real hours, recover real money, and create the proof of concept that funds the next one. This guide covers how to tell the difference.

The four-criteria test

A process is a strong first candidate when it is all four of these at once:

• Repetitive — done the same way many times a week, not just occasionally
• Rule-based — the steps and decisions follow a pattern you could write down clearly
• High volume — done often enough that the time saved adds up to something meaningful
• Costly when it slips — a missed enquiry, a late invoice, a lead that goes cold, a booking that never got confirmed

The sweet spot is a job that hits all four. That is where automation pays back fastest, makes the strongest business case, and is technically the most straightforward to build. Processes that score on only two or three of the criteria are worth automating eventually, but they should not be the first one.

A simple scoring method

Take the three or four processes your team complains about most. Score each one out of four on the criteria above — one point for each criterion it clearly meets. The process with the highest score is almost always the right first automation. If two processes score equally, pick the one with the higher cost of slipping, since that is where the financial case is clearest.

Then ask one more question: can you describe exactly what happens, step by step, right now, with no ambiguity? If the answer is yes, it is ready to automate. If the steps are fuzzy or inconsistent, tidy the process first. Automation magnifies whatever is already there — a clean process becomes reliably fast, a messy one becomes a reliably fast mess.

Where it usually hides in a growing business

A few of the most common first automations, and what they connect to:

• Answering the same enquiries, by message, phone or email. Most growing businesses have a small set of questions that make up the majority of their inbound. See AI chat agents and AI voice agents.
• Following up leads and keeping the CRM current. The time from enquiry to first contact matters enormously for conversion. See AI lead qualification and CRM.
• Keying in invoices and orders by hand. If data entry touches accounting or ordering systems daily, it is almost always worth automating early. See document processing AI.
• Sorting and answering the inbox. Businesses with high email volume, especially client-facing ones, lose significant time triaging messages that follow patterns. See email automation.

Pick the one eating the most time for the least judgement. The less genuine human decision-making a task requires, the better it automates.

Real examples from businesses like yours

Three very different businesses, each starting in a different place:

Lily's Nursery and Daycare started with enquiry handling. Parent messages arrived faster than staff could answer them, especially evenings and weekends. The first automation was an AI chat agent that answers questions around the clock and books open day tours straight into the diary. Result: 83% of parent enquiries now handled without staff, 10 hours of admin freed each week, 40% more tours booked. Read the full story: how a nursery automated parent enquiries with AI.

Tru Tribe, a music management company, started with lead qualification and CRM. Deals lived in memory and email, and follow-ups happened whenever someone remembered rather than immediately. The first automation captured every inbound enquiry, qualified it, and moved it through the CRM automatically. Result: new leads responded to in under 5 minutes instead of over 24 hours, 6 hours of CRM admin freed each week, bookings confirmed the same day enquiries arrive.

Kenu Solutions, an e-commerce business, started with order processing. Orders arrived across channels in different formats and were keyed in by hand, which was slow and error-prone. The first automation reads, checks and enters every order automatically. Result: order processing time cut by 85%, 8 hours of manual data entry freed each week, near-zero errors since go-live.

Three different starting points, all chosen because they were high volume, rule-based, repetitive, and costly when they slipped.

Sector-specific starting points

Where the right first automation tends to be by sector:

• Childcare, education and training — parent and learner enquiries, tour bookings, registration follow-ups
• Professional services — lead follow-up, quote chasing, inbox triage, appointment booking
• E-commerce and retail — order processing, invoice handling, supplier data entry
• Healthcare and wellness — appointment booking, new patient enquiries, reminder sequences
• Hospitality and events — booking enquiries, availability answers, confirmation sequences
• Property and lettings — viewings, lead qualification, tenancy admin
• Creative and media agencies — brief intake, project kick-off sequences, invoice chasing

In almost every sector, the highest-scoring first candidate is a customer-facing enquiry or booking flow, because it is high volume, rule-based at the first response stage, and directly connected to revenue.

What to leave for later

Avoid starting with anything that is complex, judgement-heavy, done rarely, or poorly understood by your own team. Creative work, negotiations, sensitive conversations, escalation handling, anything that relies on gut feel or deep domain expertise — these automate poorly or not at all, and they are not where the time goes in most growing businesses anyway.

If you cannot describe a process step by step in under ten minutes, it is not ready to automate. Come back to it once you have a clearer definition. The first win creates the template for everything that follows.

The one workflow rule

Begin with one automation, get it live, and measure what it frees up. That measurement is not just useful data — it becomes the business case for the next one. Automation builds confidence and budget in parallel. Businesses that try to do everything at once usually end up with nothing properly working. Businesses that start with one thing and get it right almost always build further, because the first one pays for itself and then some.

How to know it worked

Decide before you go live what you are measuring. Hours per week on the task before and after. Response time before and after. Error rate before and after. Revenue outcome if it is customer-facing. Take a simple baseline measurement in the week before launch, then check the same metric four weeks later. The difference is the value of the first automation, and it is the number that makes the conversation about the next one very short.

A quick exercise

For one week, keep a simple tally: every time you or a team member does the same task more than once, write it down with an estimated time. At the end of the week, sort the list by total time spent. The item at the top, if it meets the four criteria above, is almost certainly your first automation. If it does not, move down the list until you find one that does.

This exercise is also exactly what we do on the free audit — map your jobs, score them, and tell you where to start. It takes about an hour and costs nothing.