Part of our guide: AI transformation for COOs
You know the feeling of a business that is growing while the operations underneath it are not keeping up, and every new client, every new product and every new regulation adds another layer of manual work. You are hiring to keep pace, and the new hires simply absorb more of the same processes that are not working efficiently in the first place.
Here are five signs that your operating model needs to change, and what to do about it.
Sign 1: Your team spends more time moving data than using it
If people are exporting data from one system, reformatting it in a spreadsheet and importing it into another, that is friction rather than work. In data-heavy operations teams, 20-30% of their effort can go into data movement and reconciliation rather than decision-making or service delivery.
This is one of the most straightforward opportunities for automation, and AI-powered data integration and document processing can reduce manual data handling by 40-70% for well-defined workflows. The gain comes from taking the repetitive extraction and transformation work off their plate (not from replacing people), leaving them to focus on the exceptions, decisions and customer interactions that need a human.
Sign 2: Your processes depend on specific people's knowledge
When your team says "ask Sarah, she knows how this works", that is a sign the process exists in someone's head rather than in a system. Key-person dependency is both an operational risk and a scaling bottleneck.
AI-assisted process documentation and knowledge capture can help here, although the deeper issue is often that the process was never properly designed in the first place. It evolved organically and it is now held together by institutional knowledge, so the fix here is more than technology. It means mapping the process properly, identifying where decisions are made and then working out the decisions that can be supported or automated.
The third sign is that quality depends on how busy you are, showing up as error rates that climb at month-end, at peak season or during staff holidays. Processes of that kind are not resilient, working when everyone is available and focused and breaking as soon as the pressure on them increases.
Consistency is where AI adds measurable value, and an AI-assisted process delivers the same quality at 10 transactions per day or 10,000. Claims triage, invoice matching, customer onboarding and compliance checking are all processes where AI-assisted automation does more than save time, improving quality by removing the variability that comes with manual work under pressure.
Sign 4: You're hiring to grow but costs are growing faster than revenue
If your headcount is growing at the same rate as (or faster than) your revenue, your operating model is not scaling. You are adding people to do the same kind of work at the same efficiency, and that is sustainable in the short term. It creates a cost structure that becomes progressively harder to manage, and the alternative is not replacing people with machines. It is a redesign of work so that people focus on the high-value activities (customer relationships, complex decisions and creative problem-solving) and technology handles the volume processing, data extraction and routine triage. Done well, this means you can grow revenue without proportional headcount growth, and that is the definition of operational leverage.
The fifth sign is that new regulation or new products always mean "more of the same." The test is whether your team responds to new FCA reporting requirements by creating new spreadsheets and new manual checks. The second test is whether the launch of a new product requires recruiting additional ops staff to handle the volume.
If the answer to either is yes, your operating model is additive rather than scalable and each new requirement adds linear cost. A properly designed operating model absorbs those new requirements by configuring capabilities that already exist rather than building new ones from scratch.
What to do about it
None of these are problems you can hire your way out of.
If you recognise two or more of these signs, the issue is the operating model underneath rather than any single process.
The first step is to map what your operations actually do, end to end, looking past the org chart and the systems diagram. The map covers the actual work: what comes in, what happens to it, what goes out and where the manual effort lives. Each process is then scored against the potential for improvement: what could be automated, what could be AI-assisted and what still needs a human throughout.
This gives you a prioritised view of where to invest, stating "these specific processes, in this order, with this expected improvement and this business case" rather than "we need AI" as a general statement.
Our Breathe engagement maps your capabilities, scores your priority processes for AI opportunity and gives you a prioritised plan. Get in touch to talk through what is not scaling.