How we use large language models (LLM) to personalise customer communications

Personalised customer experiences set intelligent debt collection apart from the rest of the industry. To take personalisation even further, we’ve developed an AI copywriter within our product which uses large language models (LLM). 

What does an AI copywriter do?

Our internally developed solution writes outgoing customer collections communications for emails and SMS. It’s based on large language model (LLM) technology which is a form of artificial intelligence that understands and generates text in a human-like fashion, similar to Chat GPT.

How does it work?

To teach the AI copywriter how to write for our customers in the best way possible, we fed it our highest performing email and SMS collections communications which have seen the highest open rates, click through, and payment conversions.

The LLM technology analysed these communications to understand the characteristics of high performing messages, such as tone, language, and structure. We also input compliance requirements into the copywriter, to ensure any communications it writes are aligned to local regulations.  

Once it understood the components of great communication for customers in debt, the AI copywriter got to work. All AI copywriter-generated messages go through a second layer of compliance safeguarding and are reviewed by our specialist Compliance team. From there, they’re free to go live within our Collect product. The language model now produces templates on an ongoing basis for different segments of our customer base, at all stages of their collections journey.

What’s the projected impact of the AI copywriter?

As we continue to use the AI copywriter to produce more communications, there’s a wide range of benefits:

  • We can trial a broader range of communications with different styles of writing at scale to further personalise every customers’ collections experience. 
  • The LLM will be responsive and continue to learn in real time which communications customers prefer. 
  • All this translates to even higher engagement rates, payment conversions, and collections performance.
  • In time, this will also enable us to support more businesses in other countries with new languages faster, keeping up with increasing demand to work with our global clients and their customers around the world.

The use of LLM brings new levels of empathy, intuition and personalisation into how people in debt are communicated with, combining compliance and technology with unprecedented efficiency.

Learn more
View the Collect customer experience for yourself
We make debt recovery empowering, creating engaged customers who manage their repayments faster and more sustainably. Check out how we engage customers with their debt through omnichannel communications, machine learning, and our self-serve customer portal.
Play with the interactive tour

Other resources