TIME uses AI agents to draft ad proposals, but keeps sales relationships human

TIME is using two Kana AI agents to prepare advertising proposals and recommend fixes for live campaigns, putting software into work previously tied up in manual handoffs. But the publisher is keeping the advertiser relationship with its salespeople, chief operating officer Mark Howard told AdExchanger in a November 9 report.

The bottleneck began when an advertiser sent a request for proposal, or RFP: an invitation to explain what TIME could offer. A sales representative would enter the details in Salesforce. Slack alerts, Google Meet discussions with product experts and emailed notes followed as teams worked out which advertising inventory to include.

That process could leave the sales team spending days coordinating across platforms to assemble a response. Howard’s complaint was not that salespeople lacked tasks. It was that coordinating those tasks consumed time that could go toward the personal, relationship-driven work he considers central to premium media sales.

Campaign data moves into the pitch

Kana’s sales agent reads incoming RFPs and drafts proposals. It draws on what Kana co-founder Vivek Vaidya calls a knowledge graph—a shared pool of context assembled from TIME’s Salesforce customer records, order management system and ad server. The connections cover both directly sold advertising and programmatic advertising, which is bought through automated systems.

The agent also brings live-campaign data into that pool. Vaidya described this as unusual for a tool used before a sale: campaign results typically sit with advertising operations after a deal closes, helping teams assess past performance rather than prepare the next approach to an advertiser.

The intended use goes beyond answering an incoming brief. Vaidya said the system researches historical performance and similar advertisers, giving sellers material for pitches to prospective customers. The agent supplies that research; the salesperson uses it to develop a proposal.

After the sale, recommendations replace another handoff

Once a deal closes, Kana’s second agent works with advertising operations. Existing dashboards can flag campaigns that are delivering too few ads or falling behind schedule. Vaidya’s distinction is that those alerts do not necessarily tell the team how to correct the problem.

The agent recommends changes to ad setup and delivery, the campaign’s spending or delivery schedule, or its budget. The aim is to address a shortfall before TIME owes the advertiser a make-good—replacement advertising to compensate for an unmet commitment. Its described role is recommending fixes, not independently conducting the commercial relationship.

TIME’s revenue operations team is also using the tool. Howard said he had seen a significant increase in what the team could accomplish in a day or week, without quantifying that gain. He also described budget-planning uses extending beyond the year-end outlook and first quarter to modeling for 2027.

A possible connection TIME does not want

Vaidya said Kana could connect its proposal agent to an advertiser’s agents through an MCP server, a standard connection between AI agents and other software. That was a conditional description of what Kana could build, not a capability TIME said it had deployed.

Howard rejected agent-to-agent dealmaking as TIME’s current direction. He said the publisher was not looking to automate its sales relationships, while qualifying that position as the plan for now. The boundary is therefore a stated business choice, not a claim that the technology cannot connect the two sides.