Our take: automate the boring work first, the clever work later
Why the automation projects that pay off in small businesses are the dull ones, and why the impressive ones usually stall.
The short answer
When businesses think about automation now, they think about the impressive version: an assistant that advises customers, a model that predicts demand, an agent that handles a process end to end. Those are real and some of them will be worth building. They are not where to start. The work that pays off first is boring: retyping from email into systems, routing requests to the right person, sending reminders and confirmations, chasing documents, reconciling records, assembling reports, updating statuses. It is where the hours actually go, its rules are clear enough to write down, and automating it succeeds quickly and measurably, often without any AI. It also builds what the clever work needs: clean, connected data; defined processes; and a team that trusts automation because it has already made their day better. Businesses that start with the impressive project usually stall on data and adoption. Businesses that start boring finish the impressive project a year later on solid ground. Our take is to sequence by hours saved and risk, not by novelty, and to let each automation earn the next.
Boring first, clever later
| Aspect | Boring automation | Clever automation |
|---|---|---|
| Examples | Email parsing, routing, reminders, reconciliation, reporting, status updates | Advisory assistants, forecasting, autonomous agents, personalisation |
| Where the hours are | Here, in most small businesses | Rarely; the clever work saves judgement time, which is scarcer but smaller |
| Rules | Clear; a person could write them down | Fuzzy; learned from data or delegated to a model |
| Data needed | What the systems already hold | Clean, connected, historical, labelled |
| Time to value | Weeks | Months to a year, if it arrives |
| Risk | Low; errors are visible and reversible | Higher; errors are plausible and reach customers |
| Adoption | Easy; it removes a chore | Hard; it changes how people decide |
| What it builds | Data flows, defined processes, trust, ownership habits | Depends on all of those existing |
| Typical failure | Over-alerting; ownerless automations | Stalled at data; ignored by users; unmeasured |
Sequencing a programme
- Measure where the hours go for two weeks, by task, across the team.
- Pick the top boring task with clear rules and a visible result, and automate it end to end with an owner and a number.
- Repeat through the boring list, connecting systems and cleaning data as you go.
- Notice what the boring automations reveal: which data is now reliable, which processes are now defined, which people now ask for more.
- Choose the first clever project where those foundations exist, with a person who wants it.
- Build it small, with review, measurement and a switch.
- Keep going on both tracks, boring and clever, each earning the next.
What the boring work delivers
Hours back within weeks. Data that flows between systems instead of through people. Processes that are written down because automating them required it. A team that has experienced automation as help rather than threat. Ownership habits, logs and switches that will carry the clever work safely. And a business that, a year in, is genuinely ready for the impressive project and knows exactly which one to build.
What this means for you
Automate the boring work first: the retyping, routing, reminding, reconciling and reporting where the hours go and the rules are clear. It pays back in weeks, builds the clean data, defined processes and trust that the clever work needs, and tells you which clever project to build. Sequence by hours and risk rather than novelty, let each automation earn the next, and the impressive project arrives a year later on ground that holds.
Frequently asked questions
Is this not just a lack of ambition?
It is the sequence that reaches the ambitious outcome. The impressive AI project needs data that is clean and connected, processes that are defined, and people who trust automation because it has already made their day easier. None of that exists in a business that has not automated the basics, which is why the impressive project stalls at the data stage or is quietly ignored by the people it was meant to help. The boring work builds the foundation the clever work stands on.
What counts as boring work?
Anything a person does repeatedly by following rules they could write down: retyping from email into systems, routing requests, sending reminders and confirmations, chasing documents, reconciling records, assembling reports, updating status. It is dull, it is where the hours go, and it is exactly the work that automation does better than people, without any AI at all in most cases.
When is the clever work ready?
When the data it needs is flowing cleanly through the boring automations, when the process it improves is defined well enough to measure, when the team has seen automation help them and trusts it, and when there is a person who will own it. That is usually a year into a well-sequenced programme, and the clever project then succeeds because everything it depends on already works.