Picture an ordinary week in a busy business.
Someone copies customer details from one system into another. Someone else spends an hour every Friday pulling together the same report. An approval sits in an inbox until a colleague chases it. And a spreadsheet that half the business depends on needs updating by hand before anyone can move forward.
None of these tasks looks like a big deal on its own. Added together, they consume hours every week, create opportunities for mistakes and limit what your team can achieve.
In my previous article, I looked at the gap between using AI to complete individual tasks and embedding it into the way a business actually works. This article takes the next step: how to recognise the processes in your own business that could benefit from automation, and how to approach them responsibly.
If people regularly copy details from one application to another, reconcile records by hand or re-enter information that already exists elsewhere, you have what's often called "swivel-chair" work. It's slow, it's tedious, and it's where errors creep in. Connecting your systems so information flows between them can remove much of that manual effort.
Think about how a new customer is onboarded. Often it involves several people exchanging emails, requesting documents and chasing updates. A structured workflow can track progress, automatically trigger the next step, and notify the right person when action is needed, so nothing relies on someone remembering to follow up.
Spreadsheets aren't a problem in themselves. Many are exactly the right tool. The warning sign is when one becomes critical: several people update it, information is duplicated, reporting depends on it, and a single mistake can disrupt other work. Depending on the underlying problem, a purpose-built application or automated workflow may be a more reliable home for that process.
Purchase requests, document sign-offs and internal authorisations often follow a predictable path. Where they do, automation can route requests, send reminders and keep a record of decisions. The decision itself should stay with a person wherever it needs judgment or accountability. The aim is to remove the chasing, not the oversight.
This is the sign that ties the others together. Your team might already use AI to write emails faster or produce reports more quickly, yet still follow the same inefficient process every day. Sometimes the bigger opportunity is to redesign the process itself, rather than simply speed up each step.
- One important point: not all of these problems need AI. Some are best solved with conventional automation, better integration between systems or a simpler process. The right answer depends on the task, and starting with the problem rather than the technology is what makes the difference.
The goal isn't to automate the most complicated process or the one that attracts the most attention. It's to find a worthwhile problem where the benefits outweigh the cost of building, maintaining and managing the solution.
A process is a stronger candidate when it scores well on most of these:
A simple exercise to try this week: ask each person in your team to name one task they repeat several times a week that they think could be simplified or automated. Then look for patterns across the answers. The processes that come up more than once are usually the best place to start.
Automation can deliver real value, but there are three traps worth avoiding.
Not every repetitive task should be automated. A process that changes often, relies heavily on judgement or runs on poor-quality data may need fixing before it's automated. Automating a broken process just makes it break faster.
A prototype isn't a business system. Something that works in a demonstration still needs proper testing, access controls, monitoring, maintenance, and support before you can depend on it day-to-day.
Unmanaged AI use creates risk. When staff paste customer information or commercially sensitive data into AI tools that the business hasn't approved, that data can end up somewhere you can't see or control. This is often called Shadow AI, and it's growing fast. Check Point's 2026 Cloud Security Report found that 77% of organisations have changed their security strategy in response to AI, but only 26% say they have the architecture to enforce it.
The answer isn't to ban AI. It's about having a sensible approach to which tools are used, what data goes into them, and how they're secured. The Information Commissioner's Office's guidance on AI and data protection is a useful place to start.
Once you've found a promising opportunity, a structured approach makes the difference between something that works once and something your business can rely on:
Over the past year, we've been exploring automation inside Netitude before extending it to our clients.
Now we'd like to show you what we've learned.
Join us for an evening exploring practical AI and automation for businesses:
If any of these five signs sound familiar, come along and explore what could be done about them. You'll see real examples, hear practical advice and leave with ideas you can put to work straight away.
For the wider background on why the gap between using AI and changing how a business works matters, read my previous article on the AI adoption gap.
What is business process automation? Business process automation uses technology to carry out repetitive, rules-based tasks, such as moving data between systems, routing approvals or sending reminders, so people can spend more time on work that needs judgment.
How do I know if my business is ready for AI and automation? Common signs include staff moving information between systems by hand, processes that rely on long email chains, critical spreadsheets that need constant updating, time spent chasing approvals, and teams that are busy but never improve the underlying process.
Does automation always involve AI? No. Many processes are better suited to conventional automation, system integration, or simpler processes. AI is most useful where a task involves understanding or generating information, rather than following fixed rules.
What are the risks of using AI in business? The main risks include automating a process that isn't ready, relying on untested prototypes, and staff using unapproved AI tools with sensitive data. A clear approach to approved tools, data handling and security reduces these risks.