What should our AI personalise for each user?

A user's immediate task, constraints and objectives are more useful guides to relevance than a fixed persona alone. The same person seeking help with marketing performance may need different answers and proof points when considering brand identity. Their triggers, objections, alternatives and fellow decision-makers can change with the situation, even when their job title stays the same.

Build audience understanding around the conversations people need to have: what prompted their search, what could stop them buying, which alternatives they are considering and what would make an answer credible. Segmentation can remain useful, but a persona should not be the finished answer. For an AI interaction, use the user's stated problem, organisational circumstances and follow-up questions to determine which information deserves attention, rather than assuming everyone in an audience category needs the same response.

Let individual challenges and responses influence both the questions an AI experience asks and the recommendations it provides. UST faced a relevance problem: prospects struggled to identify the most useful resources within its extensive library of AI expertise. Traditional assessments, with predefined data and fixed outputs, offered limited scope for genuinely personalised guidance.

NavigatorAI addressed that problem by using generative AI to identify user-specific challenges, guide prospects through tailored questions and generate personalised recommendations informed by UST's AI playbook and expertise. Publicly available business and audience data also helped shape each journey. Human input remained central through prompt design, optimisation and supporting content. The practical model is to connect user-specific needs with relevant business expertise, allowing the interaction to shape the route to an answer rather than sending every user towards fixed outcomes.