Have you invested in an AI customer service agent that isn’t living up to the hype? Do you have disappointingly low resolution rates? Are your customers getting frustated by vague, unhelpful answers and getting stuck in doom loops.

The problem isn’t the tech, it’s the content you’re feeding it. Your policy documents, journey maps and tone of voice guidelines were never designed for a machine. You know it needs better knowledge and conversations, but where do you start?

The answer is a short, 5-Step kickstart programme that gets customer service AI answering real questions accurately and consistently.

Research shows that organisations that follow these steps see resolution rates jumping from 20-30% to 70% or more. It also upskills your team to continually optimise the content for consistently superior customer service.

the 5 step kickstart programme

    • One-to-one meetings with key staff members to identify fears, challenges and barriers. Also uncover boring repetitive tasks they would prefer not to do

    • Board report on findings

    • Awareness & Desirability – board advice to dispel fears, communicate benefits, describe the vision and encourage buy in

    • Analysis of non-resolved issues to identify key blockers

    • Assessment of your existing content (policy manuals, FAQs, help centre and macros) and to identify gaps, inconsistencies and ambiguities

    • Analysis results from test data and conversation history to uncover missing data

    • Prioritised list of content and knowledge fixes based on impact

  • A facilitated workshop designed to:

    • Share gaps from Step 2 and explore other commonly asked questions

    • Map customer journey trees

    • Discuss key handoff points

    • Creating structured content that AI can use reliably

    • Examples and guidelines to upskill the in-house team

    • Roadmap for future knowledge creation

    • Defining the brand voice

    • Contextual tone alterations

    • Tone intent to guage handoffs to humans

    • Guardrails to ensure consistency

The programme is delivered in a single sprint lasting 3-4 weeks. The AI engine is then tested before being deployed. The key metric is resolution rates. This process is designed to kickstart low resolution rates (25%-30%) to overcome barriers to achieving high resolution rates (75%-80%, depending on industry sector)

GET IN TOUCH