Considering this framework will help any support team to harness its time and focus its efforts, regardless of its size or maturity.

Handle

At the core of any support organisation is a group of talented people who are trusted to engage with customers and handle their issues. This is the most valuable yet least scalable part of their role. It’s the one duty that they own end-to-end. Put differently: if they’re not around, nobody else is going to be working on support tickets as they do.

Working through customer tickets remains the bread and butter of the job and, without it, we don’t learn how customers are using the product. We don’t know what the trends and behaviours are. We don’t understand what the quick wins or big hitters are. We don’t earn the right to be respected as product experts.

Guide

Armed with a plethora of context, titbits and know-how, a competent support engineer will find ways to share it all with the world. Traditionally, they have been tasked with maintaining documentation and knowledge base articles. Increasingly, support teams are asking customers to engage with AI mechanisms before speaking with a human. These teams are maintaining a repository of short-term gotchas (like incidents or known edge cases) which wouldn’t normally make their way into documentation, but AI can leverage. If your team aren’t leveraging AI for these situations, they should be – the ROI is guaranteed.

After every single support ticket, teams should be considering how and where they can guide customers who might encounter the same issue in the future.

Prevent

The best problems are the ones which never happen. Prevention sees confusing error messages rewritten, footguns removed from APIs and default parameters changed so that the sensible path is the easy path.

Driving prevention requires the volume, trends, customer names and revenue attached to the problem. Usually, experts within support teams will spend time probing closed cases and customer surveys before they sync with engineering and product teams to discuss product behaviours, the desired changes and their tangible impact. Now, support teams are discovering product insights by prompting AI and validating its output with their own experience. This forms a prioritised list, backed by volume and revenue, which can be taken to a product review. The human effort then shifts from identifying the right problem to debating the right solution.

Measures for the Framework

There are too many nuances between companies to offer a concrete and foolproof list, but some tried-and-tested, team-level measures include:

• First response time (lower is better and can show better handling)
• Customer wait time (lower is better and can show better handling)
• AI-derived QA score (higher is better and can show better handling)
• Cases per x customers (lower is better and can show better guiding)
• % of AI chats escalated (lower is better and can show better guiding)
• AI chats per x customers (lower is better and can show better prevention)

These measures can then be broken down by categories such as customer plan, product area or urgency.

Finding the Balance

The classic “teach a man to fish” proverb rings true here. Handling cases is like giving a customer a fish; guiding customers is like teaching them how to fish; and preventing issues is like making sure they never go hungry in the first place.

It’s tempting to treat this as a maturity ladder, where the goal is to spend as much time upstream as possible. It isn’t. A team that only handles is busy, reactive and quietly burning out, but a team that only prevents has lost touch with what customers actually experience and will spend cycles trying to fix things nobody is complaining about.

The balance will be unique to each organisation and its circumstances. It may need to be revised yearly, quarterly or even monthly. During a product launch, teams will need to handle more to build their expertise and absorb the influx of hungry customers. After a period of platform instability, uptime may become a new company SLO, and prevention a priority to stop more customers going hungry.

Support leadership should reflect on the organisation’s goals, and whether the team’s current balance of handling, guiding and preventing is best positioned to work towards them.