Most automation projects fail for the same reason: they start with the most interesting process rather than the most expensive one. These five are ordered by how much time they give back against how little work they take to set up.
A test before you automate anything
Automate a process that is broken and you get a broken process that runs faster and is now harder to see into. Before anything else, write down the steps as they actually happen, including the exceptions people handle without mentioning them.
If you cannot write the rule down, it is a judgement call, and judgement calls should stay with people. Automate the lookup, the routing and the copying. Leave the deciding.
1. Lead routing
An enquiry arrives, someone reads it, decides who should handle it, and forwards it. That is a lookup dressed as a decision, and it usually adds hours to the response time for no benefit.
Route on what the form already captures — service, region, budget band, urgency — straight to the right person or queue, with an acknowledgement to the sender. Response time is one of the few things a small business can genuinely beat larger competitors on, and this is the cheapest way to win it.
2. Invoicing and payment chasing
Invoices go out late because someone has to remember, and get chased late for the same reason plus the awkwardness. Both are pure scheduling.
Generate invoices from whatever already records the work, send them on a schedule, and let the reminders go automatically at seven, fourteen and thirty days. The reminders being automatic is most of the value — nobody has to decide to send them, so nobody has to feel bad about it.
3. Client onboarding
The gap between someone agreeing to work with you and work starting is where a surprising amount of goodwill is lost. It is usually a sequence: welcome, contract, deposit, kick-off booking, information request, access credentials.
The whole sequence is triggered by one event. Automate it and onboarding becomes same-day rather than same-week, which changes how the relationship starts. It also stops the intake form arriving three days later with an apology attached.
4. Follow-ups
Most quotes are never followed up. Not from a decision — from the sheer administrative weight of tracking who was sent what and when.
A quote sent triggers a follow-up in three days, another in ten, a last one at thirty. Written once, personal in tone, easy to stop the moment someone replies. This is generally the first automation that pays for the entire project, because the leads were already yours.
5. Reporting
Someone spends the first morning of every month assembling the same numbers from the same three places into the same spreadsheet. It is the most automatable task in most businesses and usually the last one anyone touches, because it is only painful once a month.
Pull the numbers on a schedule into one dashboard. The time saved is real but secondary — the actual gain is that the numbers become current instead of monthly, so a bad trend is visible in week two rather than after it has finished happening.
What we would not automate yet
Anything customer-facing that requires tone. Automated replies that try to sound personal read worse than a slower human one, and people can tell.
Anything with a genuine exception rate above roughly one in five — you will spend more time handling the exceptions than the automation saves.
Anything you have not done manually enough times to know the edge cases. The first version will encode your assumptions, and if those are wrong the automation makes them permanent.
Keep Reading
Common Questions
Lead routing or quote follow-ups. Both touch revenue directly, both take the least work to set up, and both prove the case internally before you commit to anything larger.
Got A Question About Your Own Site?
If "5 Business Processes You Should Automate First" raised a question about your own setup, send it over. We answer these properly whether or not it turns into work — and if this is not actually your bottleneck, we will say so.