Business process automation examples: what actually gets automated.

Most articles on this subject list capabilities. That is not much use when you are trying to work out whether any of it applies to you, because the hard part is never the technology. It is deciding which specific process in your business is worth changing first.

So each example below is written the same way: what the manual version looks like, what the automated version does instead, and — the part usually left out — the condition that has to hold before automating it makes commercial sense.

What counts as business process automation

Business process automation means using software to carry out a defined sequence of business steps that a person would otherwise do by hand: moving information between systems, applying rules, sending communications, and escalating anything that falls outside the rules.

It is distinct from two neighbours it gets confused with. Robotic process automation usually means software imitating a person clicking through screens, used where no proper interface exists. Workflow automation tends to describe the routing and approval layer specifically. In practice most real projects mix all three, and the label matters far less than whether the process was worth changing.

Front-office processes

1. Responding to new enquiries

Now: an enquiry arrives by form, phone or messaging. It waits in an inbox or a voicemail until somebody is free. Out of hours it waits until morning.

Automated: the enquiry is acknowledged immediately, the caller or sender is asked the qualifying questions the business would ask anyway, and the answers are written into the CRM with an owner assigned.

Worth it when: enquiries arrive outside the hours you can cover, or a meaningful share currently receive no reply at all. This is the most commonly automated process in small and mid-sized businesses, and the evidence behind it is routinely misquoted — worth reading before you build a business case on it.

2. Booking appointments

Now: a back-and-forth to find a slot, then manual entry into a calendar, then a reminder somebody has to remember to send.

Automated: live availability is checked against real constraints — travel time, skill, territory — the slot is booked and confirmed, and reminders are scheduled without anyone touching them.

Worth it when: the appointment is the commercial step that matters, and no-shows or scheduling admin are a visible cost.

3. Qualifying and routing leads

Now: every enquiry reaches the same inbox regardless of value, location or type, and somebody sorts them by hand.

Automated: enquiries are scored or classified against your criteria and routed to the right person, branch or queue, with anything ambiguous held for review rather than guessed at.

Worth it when: you have more than one destination for an enquiry and the sorting is currently costing someone real time, or getting it wrong is expensive.

4. Following up quotations

Now: follow-up depends on memory and workload, so the quiet weeks get chased and the busy weeks do not.

Automated: every open quotation is followed up on a defined schedule, with the timing and message consistent, and replies routed back to the person who owns the deal.

Worth it when: your quotation values are high enough that a small recovery rate pays for the work, and you can honestly say follow-up is currently inconsistent.

Finance and administration

5. Invoice processing and approval

Now: invoices arrive by email, are keyed into the finance system by hand, and chased around the business for approval.

Automated: the invoice is read, matched against a purchase order or expected value, routed to the right approver by amount and category, and posted once approved — with exceptions and mismatches queued for a person rather than pushed through.

Worth it when: volume is high enough that keying and chasing is a role rather than a task, and your approval rules can actually be written down. Set out in full in how an invoice approval workflow works.

6. Chasing payment

Now: overdue invoices are chased when somebody gets to it, in whatever tone that person uses, with no reliable record of what was said.

Automated: chasing runs to a defined ladder across email, message and call, escalating on schedule, recording every contact and stopping the moment payment arrives. Recoups is a product Megabite built to do exactly this, including handling the calls.

Worth it when: your debtor days are materially worse than your payment terms and the cause is inconsistent chasing rather than customers who cannot pay.

7. Producing recurring reports

Now: somebody exports from two or three systems into a spreadsheet every Monday, reconciles the differences, and formats the result.

Automated: figures are pulled directly from source systems on a schedule and presented in a live dashboard or a generated pack, with the reconciliation logic written once instead of repeated weekly.

Worth it when: the report is genuinely used for decisions. Automating a report nobody reads simply makes it arrive faster.

8. Handling documents and records

Now: certificates, contracts and photographs live across email, phones and shared drives, and finding one means asking whoever did the job.

Automated: documents are captured against the job or customer record at the point of creation, named and filed by rule, with missing items visible rather than discovered later.

Worth it when: you have a compliance, warranty or handover obligation that depends on producing the right document on request.

People and delivery

9. Onboarding a new customer

Now: a checklist someone follows from memory, with the customer chased for the same information more than once.

Automated: each step is triggered in sequence — welcome, information request, account setup, first appointment — with progress visible and stalled onboardings flagged.

Worth it when: onboarding has a defined shape and a slow or messy start visibly costs you customers.

10. Onboarding a new employee

Now: accounts, equipment, access and paperwork coordinated by email between whoever remembers.

Automated: one trigger raises every task against its owner with a due date, so the new starter has what they need on day one and nothing depends on one person's recall.

Worth it when: you hire often enough that the process repeats, and access or compliance failures carry a real cost.

11. Recruitment administration

Now: applications tracked in a spreadsheet, candidates chased manually, and rejections often never sent.

Automated: applications are acknowledged, screened against stated criteria, moved through defined stages, and every candidate receives an outcome.

Worth it when: application volume is high relative to the people handling it, and your employer reputation matters commercially.

12. Internal approvals

Now: requests for spend, time off or discount go by email to whoever is thought to be responsible, with no record of the decision.

Automated: the request goes to the right approver by rule and threshold, escalates if it stalls, and leaves an auditable record of who approved what and when.

Worth it when: approvals are a bottleneck, or you cannot currently answer who authorised something without going through an inbox.

Which of these applies to you

Recognising a process on this list is not the same as it being worth automating. The order we use is deliberately unglamorous:

• Frequency first. A process that runs a hundred times a week repays a build. One that runs monthly rarely does.
• Then the cost of getting it wrong. A missed enquiry and a mis-filed document have very different consequences.
• Then whether the rules exist. If two competent people would handle the same case differently, the process is not ready and no software will settle the argument.
• Then the data. Automation moves information between systems. If the information is not in a system, or is unreliable, that is the actual project.
• Then the exception path. Every automated process needs a defined answer for what happens when the rules do not fit. Without it, the exceptions quietly become somebody's full-time job.

We set this out in more depth in what a growing business should automate first.

What does not automate well

Being clear about this saves more money than any of the examples above.

• Processes nobody has agreed. Automating a disputed process makes the disagreement permanent and harder to change.
• Genuine judgement. Pricing a difficult job, handling a complaint, deciding whether to keep a customer. These can be supported by software; they should not be decided by it.
• Rare, high-consequence events. If it happens twice a year and matters enormously, a person should do it and a checklist should guide them.
• Anything resting on bad data. Automation applied to unreliable records produces confident, fast, wrong answers.
• Processes about to change. If the business is mid-restructure or mid-system-migration, wait. You will build it twice.

Common questions

What business processes can be automated?

In practice: enquiry response, appointment booking, lead routing, quotation follow-up, invoice processing, payment chasing, recurring reporting, document handling, customer and employee onboarding, recruitment administration and internal approvals. The common thread is that each is rule-governed, repetitive and already agreed.

What is the difference between business process automation and workflow automation?

Workflow automation usually refers to routing and approvals — moving an item between people and stages. Business process automation is the broader term covering the whole end-to-end sequence, including the system-to-system data movement. Most real projects involve both.

Where do most businesses start?

Usually with enquiry response, because it is high frequency, has an obvious commercial consequence when it fails, and the rules are easy to agree. It is also the easiest place to measure whether the change worked.

Do you need AI to automate a business process?

Usually not. Most of the examples here are deterministic rules and integrations, which are cheaper, more predictable and easier to audit. AI earns its place where the input varies in ways rules cannot capture — interpreting a message, holding a conversation, extracting data from inconsistent documents.

How long does a first automation take?

A single well-defined process is usually weeks rather than months. The time goes into agreeing the rules and connecting the systems, not the automation itself.

What usually goes wrong?

Three things, in order: the process was never agreed, so the automation encodes one person's version; there was no exception path, so anything unusual piles up unnoticed; or the underlying data was not good enough to act on.

Related reading: business and workflow automation services, systems integration, and what AI agents should and should not be allowed to automate. For enquiry response specifically, see the AI receptionist.