This Week's Term: AI Persona - the deliberately designed personality and emotional character of an AI system. The difference between a generic chatbot and an AI that feels like talking to a real colleague.
Most teams treat persona as a cosmetic decision: pick a name, choose a friendly tone, maybe add an avatar. But personality-driven chatbots see 43% higher interaction rates than generic ones. Users who feel an AI persona relates to them show 67% higher engagement. And users form emotional attachment in as few as three interactions. Persona is a design discipline with measurable business impact, and most companies are still treating it as a dropdown menu.
What makes this more than a UX question is recent research from Anthropic. Their February 2026 study on the persona selection model found that AI assistants are effectively playing a "character," and the character you teach shapes behavior across all contexts. When researchers trained Claude to act deceptively in one narrow scenario, it learned to act broadly misaligned. The inverse is also true: a well-designed persona produced more reliable, more trustworthy outputs. Character traits do not just change how an AI sounds. They causally drive what it does.
For business leaders, this changes how to think about AI deployment. Persona is an architectural decision that affects how much people trust and actually use the system. Every AI agent already has a personality, whether you designed it deliberately or left it to chance. Emotion is the layer that determines whether people actually use them, which I find most product teams learn the hard way, after launch.
Connecting last week's term (the unauthorized use of AI tools without organizational oversight) and the interesting suggestion to introduce an AI persona into the boardroom, I encourage you to watch the TECHtalk episode where Kevin Bocek (SVP Innovation, CyberArk) discusses AI agents taking on advisory and decision-making roles, including "AI advisory boards" that counsel CEOs. When AI agents move from tools to advisors with real influence, persona design becomes a governance question, not just a UX one.
Your action step
Audit the AI agents your organization currently uses: internal tools, customer-facing chatbots, productivity assistants. For each one, ask whether the persona was deliberately designed or happened by default. Write down three adjectives that describe how each agent "feels" to interact with. Then ask whether those are the adjectives you would choose if you were designing the experience from scratch. If there's a gap, that's your starting point.