Why do we give the agents male names and the assistants female ones? I asked that question out loud at Learning Technologies yesterday, and nobody had a good answer.

We’ve been here before. Siri. Alexa. Cortana. The pattern is established: give the assistant a woman’s name, make it compliant, make it warm, make it endlessly patient with people who would never treat a human colleague that way.

Now we’re in the ‘agent’ era. The tools that reason, decide, and act autonomously. The products being sold to CFOs and CTOs at enterprise price points have quietly shifted in their naming.

Harvey is your legal advisor. Devin is your engineering co-pilot. Jasper creates your content. One investor who tracks AI pitches found that among early-stage teams who gave their AI agent a human name, 70% chose male names, 20% female, and 10% neutral.

The research is not subtle: Digital assistants are programmed female, and digital advisors in legal, financial, and medical contexts are programmed male. The rule is consistent: service sounds female, expertise sounds male.

This is not a quirk of naming conventions but a design decision. When you name an AI agent, you are not just picking a label. You are creating a persona, and that persona shapes how people perceive competence and authority.

Every design choice, from naming to voice, is a sociotechnical decision that carries the biases of its designers. Most of those designers are men. Most of the enterprise buyers they are pitching to are men. The persona choices are not accidental: they are conversion tactics.

So when the assistant coordinating the agents has a woman’s name, and the agents do not, that is not a neutral organisational choice. It is a rehearsal of a very old idea about who assists and who decides.

We are building the infrastructure of future work right now. The defaults we set will outlast the conversations we are avoiding.