The deal went quiet. What did the buyer need?
AI could help publishers win more of the business they already pitch. The bigger prize is finding out what customers would pay for next.

Cartoon · AI-generated
“Can I prove it? Am I gonna be able to convince my FD?”
Raoul Monks, the founder of Flume, was playing the part of a buyer at the Flashes & Flames AI Forum at Stationers’ Hall. In his example, the salesperson had come off a good call. The buyer sounded keen. The CRM said the deal was healthy. Five weeks later, it was gone. The buyer had stopped replying. The explanation was budget. Flashes & Flames
Anyone who sells event sponsorship will recognise the scene. Your contact likes the event and knows the audience. Then they have to explain to finance why those attendees are worth paying to reach, how the company will judge the results and why it should commit now. Your proposal answered the questions your contact asked on the call. The questions that decide the deal may come later, in a room you’re never in.
That was the strongest idea in Monks’s demonstration of THEA, Flume’s AI sales platform. It reads calls, emails and CRM records to identify what the buyer may still need. Sometimes the salesperson should send an audience breakdown. Sometimes they should ask how finance will measure success. And sometimes the most valuable thing the system can do is flag that nobody has spoken to the person who controls the budget.
A better forecast is worth something. A signed deal is worth more.
In a Gartner survey of 632 B2B buyers, 74% of buying teams experienced what Gartner classified as unhealthy conflict. That gives Monks’s argument a solid starting point. Getting a group of people to agree on a purchase is part of the sale, and a publisher can make that job easier for its contact. Gartner
The evidence presented at Stationers’ Hall does not back up the claim that AI coaching raises win rates. Monks described a retrospective exercise involving 154 deals. That can help establish whether the system recognises a deal in trouble. The revenue claim depends on whether salespeople who act on its advice close more comparable deals.
For a publisher, the test is straightforward. Give one group AI-assisted guidance on live opportunities and compare it with a group selling similar packages under the same qualification rules. Count the deals that sign, the ones that end in no decision and the discounts it took to get them over the line. A sales director might happily pay for better forecasting. The case for revenue growth rests on the additional business the team actually wins.
What if the buyer never takes the call?
“One of our main issues really is to do with getting people on the phone or willing to talk to you in the first place,” an audience member said.
That question mattered because the demo started with a sales conversation. For some publishers, getting the right people into the room is the harder problem. AI makes it cheap to send another email. A specialist publisher’s advantage is knowing something that makes the email worth opening.
A change in regulation may be bringing a new group of buyers into the market. Audience data may show those buyers researching a problem a prospective sponsor can solve. That is a reason to call this week. AI can help find the connection and assemble the evidence. The commercial measure is how many qualified meetings it earns per 100 target accounts. Track that alongside the percentage of qualified deals that close and you can see where to spend your next pound.
What the deals you lose can tell you to build
Another attendee described listening to roughly 100 sales calls and finding that many customers struggled to articulate what they wanted. For a publisher, that may be the most valuable observation of the session. Calls that end without a sale can still tell you something about demand.
Suppose several sponsors ask how to reach a specialist audience between annual events. That is a reason to test year-round sponsorship. Suppose prospective subscribers keep asking what has changed in their market this week and what it means for decisions they’re already making. That is a reason to test a paid intelligence service. AI can help a small team spot the pattern across months of calls and emails. A buyer paying for a pilot tells you whether the pattern is a business.

There’s a precedent for getting existing sponsors to pay for a different format. In a supplier case study, Turtl reports that Informa Connect generated about $55,000 from 11 sponsorship packages across three eBooks. FutureWeek’s launch of AIMM Forum, a community for senior media and marketing leaders dealing with AI, is another case worth watching. The launch gives us a product to follow. Paid memberships and sponsorship commitments will tell us how much of a market there is. Turtl case study · VideoWeek
There is a third route to revenue, and it starts with a question publishers could barely ask a few years ago. Who is already using our content inside AI products, and can we charge for that use? The SPUR coalition, which includes the Guardian and the Financial Times, has published a standard for tracking how AI tools use publisher content. That gives publishers something more concrete to negotiate over. The commercial milestone is a customer signing a licence whose price or terms reflect measured usage. Digiday
The next time a sponsorship deal goes quiet, I’d look at what the buyer was trying to achieve. If they needed a better case for finance, give the next buyer the evidence to make it. If several buyers wanted something the package couldn’t deliver, ask them to pay for a pilot that can. That is where a talk about sales coaching becomes a more interesting conversation about the future of publishing.
Source: Tomorrow’s Publisher