How Much Personality Should a Business AI Tool Actually Have?

This one's speculative — a thought experiment about design principles, not a report on something I've built.

Personality is a design choice with real costs

Giving an AI tool a warm, chatty, personable voice feels like an unambiguous improvement — friendlier is better, surely. But personality isn't free: it adds tokens, it adds room for tone to land wrong on a bad day, and it can obscure the actual information underneath a layer of conversational filler that a user has to read through to get to the point.

Match the tone to the stakes of the task

A casual, friendly tone fits a low-stakes recommendation tool well and fits a system reporting a financial discrepancy or a compliance issue badly — the mismatch between a chipper voice and a serious message actively undermines trust in the message itself, because it reads as not taking the situation as seriously as the content warrants.

Consistency matters more than warmth

A tool's personality doesn't need to be maximally likable to be effective. It needs to be predictable — the same tone, the same level of directness, every time, for the same category of interaction, so a user builds an accurate mental model of what to expect from it, rather than being pleasantly surprised sometimes and jarred at others by an inconsistent voice.

Plain and clear is an underrated choice

For a business tool that people rely on to get through actual work, a plain, direct, low-personality voice is very often the better choice, not a compromise made because personality was too hard to build. Personality is a legitimate design decision for the right context, not a default every AI product should reach for.

I'm Jesse Myers — Marine veteran, 32 years in enterprise IT, now building production AI systems. This site is where I write about what I've actually built, and occasionally about ideas I haven't built yet but think are worth taking seriously.