Trading desks may never look the same. Ad tech companies like MIQ and Viant and agencies like Horizon Media are putting AI trading agents to work, using tools built on large language models to collapse hours of media planning into seconds. At MiQ, Sigma has already clawed back 6,000 labor hours and led to a stunning 132% jump in conversion rates. One exec prompted Sigma to rank campaigns, compare DSP spend, and even draft emails to clients—all in one chat. The bots don’t sleep, and apparently, they don’t miss much either.
But behind the curtain of automation lies a looming identity crisis for ad tech. As AI takes on tasks from bidding to performance analysis, agencies are reckoning with how much control they’re handing over to algorithmic co-pilots. With new language-based tools like Horizon’s Blu platform or Viant’s AI-powered planning and bidding system, the “middleman” conversations (between account execs, solutions architects, and data analysts) are becoming increasingly unnecessary. Meanwhile, terms like “agentic” and Model Context Protocol (MCP) are creeping into vocabularies—MCP being the infrastructure fueling agent-to-agent communication and giving AI tools real traction.
Whether it’s Viant processing 5 million ad requests per second or MiQ visualizing customer personas from 700 trillion data points, it’s clear that AI isn’t just assisting—it’s helping agencies rethink who’s actually steering the ship. Sigma can not only recommend bid changes but execute them with one message. That’s turning campaign optimization into a task as easy as texting. Still, the AI rollout is being paced carefully, with leaders acknowledging that not all agencies are ready to onboard full automation—or explain how exactly it works.
Read more at AdAge.