Conexions

Our work / Case study

What we learned putting generative AI on the front line

Several production AI capabilities in a customer-service team in six months, and the lesson that shaped the AI programmes that followed.

  • Automation & AI

The problem

An established customer-service operation wanted to put generative AI in the hands of its frontline teams, inside the tools they already used every day.

Our approach

We led the implementation of several AI capabilities directly within the existing customer-service environment:

  • AI service replies: suggested responses to customer enquiries, based on the context of each case.
  • AI case summaries: long customer and case histories condensed into something readable at a glance.
  • Sentiment analysis: extra insight into how each customer interaction was going.
  • AI recommendations: suggested next actions to support the person handling the case.

Several production AI capabilities went live within six months.

DEPLOYED INTO THE SERVICE DESK SERVICE REPLIESsuggested responses CASE SUMMARIEShistories at a glance SENTIMENThow it's going RECOMMENDATIONSsuggested next actions THE LESSON: ADOPTION IS THE MEASURE, NOT DEPLOYMENT A feature only succeeds if it removes enough work that people want to use it
Shipping features isn't the goal; removing work is.

The lesson

The technology worked, but take-up of some capabilities was lower than expected.

That turned out to be the most valuable part of the project. It showed that successful AI transformation isn’t about how many AI features you can deploy. It’s about whether each one removes enough work that people actually want to use it.

That lesson directly shaped the approach on the AI programmes that followed: start with the work people want gone.

Work covered: generative AI, CRM and customer operations.

Want AI your team will actually use? Book a call and we’ll start with the work, not the features.

← All work