AI is brilliant at specific types of work, and terrible at others. The difference between a transformative investment and wasted budget often comes down to one thing — did you pick the right problem to solve?
That's where the 3Rs Framework comes in.
What Are the 3Rs?
Three characteristics that make a task a good candidate for AI automation:
- Repetitive. Does this task happen again and again with a similar pattern? Insurance claim intake. Customer enquiry triage. Quote generation.
- Rules-based. Is there clear logic? "If claim value is under £5,000, assess immediately. If over £15,000, send to expert review." That's where AI shines.
- Resource-intensive. Does this task consume time, people, or money in a way that hurts? Not just any repetitive task — the ones that pull skilled people away from higher-value work.
Pick a task that's all three, and you've found your opportunity.
Real Examples
- Motor Claims Intake. 200 times a week, 15–20 minutes each. Same questions, same format. Classic 3Rs opportunity.
- Insurance Broker Renewal Processing. Same workflow, same documents, same decision logic. Admin staff handling it, delaying strategic work.
- Credit Hire Enquiry Handling. Driver details, vehicle requirements, availability, quotes. Repetitive, rules-based, resource-intensive.
- Customer Service Escalation Triage. Which tickets need human attention? High volume, clear logic, expensive when slow.
How to Apply the Framework
- List your biggest operational processes.
- Score each against the 3Rs. 1–3 scale. Total above 7 is a priority candidate.
- Talk to the people doing the work. They'll tell you instantly.
- Map the actual process. Don't guess. This step stops 80% of failed AI projects.
What Comes Next?
Once you've identified a 3Rs opportunity, the next question isn't "which AI tool should we buy?" It's "what would success look like if we actually solved this?"
Try It Yourself
Spend 30 minutes this week. List three processes that frustrate your team. Score each against the 3Rs. Which scores highest? That's where you start. Not with vendor demos. With a clear picture of the problem.
Related articles
Why Most AI Projects Fail (And How to Avoid It)
Over 80% of AI pilots never make it to production. The real reasons projects stall, and a practical guide to successful adoption.
The CEO's Guide to AI: Questions to Ask Before You Invest
7 critical questions every CEO should ask before approving AI budget. A practical checklist for smarter investment decisions.
WEF 2026: AI Advantage Comes From Adoption, Not Invention
Key takeaways from Davos 2026: the AI winners won't be those building the smartest models, but those who use AI at scale.
