Why plan a campaign yourself when an agent could do it for you?
On Thursday, digital ad platform Basis launched a new agentic campaign planning solution called Compass that has direct integrations with buying platforms, including its own DSP, of course.
But Basis is “not trying to replace humans,” Mike Hoyle, the company’s chief product officer, told AdExchanger, nor is it trying to develop a “one-shot strategy” that can launch campaigns in real time without any human input.
It’s just trying to make the planning process “easier to navigate,” he said.
Compass generates media strategies and helps plan omnichannel campaigns using a combination of an advertiser’s first-party data and broader market data, all of which is synthesized agentically.
Tim Slater, digital media director at ad agency SMZ, an early tester of the tool, describes himself as “skeptical by nature,” and said he was initially reticent to outsource media strategy to a platform that might favor its creator’s other offerings. Basis would have a “vested interest” to prioritize its own DSP, for example.
But SMZ hasn’t seen any bias, Slater said, which gave him “more trust,” including for budget distribution.
Vibe shift
SMZ has been a software partner of Basis since 2019 and has been using Compass since February to determine the best ways to shift away from direct buys toward programmatic.
AI is “supercharging” the role of the agency, said Basis CEO Shawn Riegsecker. In his view, brands will eventually have a unique agent for each step of the campaign-planning process: a finance agent, a sales agent and so forth. The most effective systems, he added, will be able to receive signals directly from the agents via Model Context Protocol and use them to make effective campaign decisions.
SMZ, for example, is shifting away from “vibes-based buying,” Slater said, a term he coined for choices that advertisers make based on intuition and personal observation, rather than specific data points.
Compass allows users to upload briefs along with historical data and brand messaging, which it synthesizes and uses to generate a suggested media strategy.
The goal is to reduce pre-activation time through automation.
The next version of Compass, which is being built this summer, will pull in all of the historical data from Basis to inform recommendations. Advertisers will be able to directly query the data, which will connect to the agent in Compass via a set of MCP endpoints, Hoyle said.
Make it brief
The first step to generating a campaign with Compass, said Slater, is uploading a client’s media brief along with all of its analytics reports so the tool has an understanding of the brand’s performance, delivery and scale.
Then, SMZ can have a “back and forth” conversation with the Compass agent, he said, hashing out details like budget constraints and the availability of media assets. At the end of the conversation, Compass generates a slideshow presentation, which SMZ tweaks to suit its preferred formatting and voice, Slater said, along with a proposed media mix, creative strategy and targeting capabilities.
The targeting includes recommended audiences and personas. The ability to activate those audiences within Compass will also be available this summer, said Hoyle.
Each audience persona gets a name of its own and a short profile. One of SMZ’s clients is a company in the higher education space, for which Basis suggested four different personas to target. Each included a distinct set of attributes, including age range, pain points and what sort of content tends to resonate with them. The recommendations were generated using “pre-built expertise” from past campaigns across verticals, said Hoyle.
One audience was called the “onliners,” said Slater, and was made up of full-time workers who care about balancing their family and career. Their pain points include time scarcity and skepticism about online school, but they’re motivated by the potential to advance their careers, increase their salaries and develop professional credibility.
Compass determined that this audience watches more streaming TV than broadcast and indexes highly on LinkedIn and Meta.
Eventually, Compass will include an orchestration layer that sits on top of incoming data, such as ad placements and investment information, said Riegsecker. This will allow the system to make its own “judgment calls” and recommendations, including AI optimization at the bid level, he said.
Basis also has its own data warehouse with cross-channel performance measurement, along with data on both programmatic and direct buys from previous campaigns and over 400 third-party platform integrations.
And as it introduces more MCP capabilities into Compass, more historical data will be implemented directly into the platform, said Hoyle, rather than having to be manually uploaded by the advertiser.
“This is a foundation,” he said, “a starting point.”
