BotBot | Your AI Has a Default Personality. Did You Choose It?<br>Get early access

BotBot<br>Back to all posts<br>You've spent years honing how your brand feels: the look, the voice, the way a customer gets greeted at the door.

There's a brand book somewhere that argues about one shade of red. Someone rewrote the greeting script four times. You mystery-shop your own locations to make sure the feeling survives contact with a Tuesday afternoon.

That identity is one of the most expensive assets you own. Nobody knows your customers better than you do.

Then an AI agent picks up the conversation, and you sound like every other Claude on the internet.

This is no longer a side channel

Voice AI is taking one in five inbound contact-center conversations this year, up from one in twenty two years ago — orders, bookings, support, the moments where the sale actually happens.

The surface where your brand meets customers most often is increasingly the one surface nobody designed.

The providers know exactly what's at stake

OpenAI, Anthropic, and Google have spent nine months proving it:

November — GPT-5.1 shipped personality presets.

December — OpenAI added warmth sliders.

Anthropic published Claude's constitution, then research showing post-training doesn't create a personality so much as select a character from a cast the model already contains.

This month — Google is testing energy and formality sliders for Gemini.

Every provider now agrees behavior comes in adjustable traits.

But all of those knobs tune a consumer's own assistant

Your drive-thru, your support line, your booking flow: the feel is whatever the defaults add up to.

A character selected by a vendor, for everyone at once.

Your logo didn't come from a template, but the personality your customers actually talk to shipped without a single decision from anyone accountable for it.

That's an outsourced identity.

The other side of the alarm

This is the most open design space in your business.

Adobe found 42 percent of organizations planning distinctive AI personalities for different audiences — almost nobody has claimed their own feel yet.

The brands that treated the web as a design surface early owned a decade of it.

So the question: what would you change about your AI's personality if changing it were a design decision instead of an engineering ticket?

(I made the longer argument — the palette, the modes, the trigger map — in Behavior is a design material now.)

About the author<br>Michael Mayernick<br>Michael is CEO and Co-Founder of BotBot, the operational control layer for AI agents. He’s an entrepreneurial builder with deep expertise in machine learning, analytics, and enterprise AI systems. Previously, he led engineering for IBM Watson Orders, an edge AI platform deployed across McDonald’s restaurants, where he built teams and helped scale a nationwide rollout. Before IBM, Michael was CTO and Co-Founder of Spinnakr, an analytics startup backed by Andreessen Horowitz and 500 Startups. He has also chaired Sandbox Network, a global community of entrepreneurs and creatives, and has testified on AI and data analytics before Congress, the Department of Commerce, and the European Union Parliament.

Related Articles

AI's Missing Design Layer, Six Months Later<br>In the spring I argued AI stacks were missing an experience layer. Palettes, modes, and trigger maps were a proposal then. Now they're how some of the world's most forward-thinking brands differentiate their AI experiences.

© 2026 BotBot. All rights reserved.<br>Operate AI experiences at scale.