华尔街日报如何将 AI 落地于广告销售与媒介排期流程
Putting AI To Work At WSJ
华尔街日报广告销售团队引入多项内部 AI 工具,全面自动化资料检索、排期核对与媒介方案生成等流程。团队正在测试的 AI 媒介策划工具将 RFP 方案生成时间从四五小时缩短至约一小时,促使团队集中精力跟进百万美元级大单,带动中标率提升 5 个百分点。但在商业定制内容制作上,该报坚持由人工撰写,仅将 AI 作为辅助工具。

Wednesday, November 18th, 2026 – 1:00 am
SHARE:
The best AI use cases in ad sales aren’t flashy. Mostly, it’s about automating the tasks nobody wanted to do anyway.
That’s been the case at The Wall Street Journal, where AI is quietly taking over most of the busywork for the ad sales team, says Josh Stinchcomb, the Journal’s global CRO, on this week’s episode of AdExchanger Talks.
For example, what seller doesn’t know the frustration of trying to track down a stat buried somewhere in an old email thread or Slack channel when they’re trying to get ready for a meeting?
But now, the Journal’s sales team has access to an internal AI-powered search tool that does the digging for them. It can even recap whatever the seller last discussed with a particular prospect so they can walk into their meeting prepared and knowing exactly where the conversation left off.
“It’s really helpful for pulling from the information we have available to ourselves but isn’t always at our fingertips,” Stinchcomb says. “That’s been a great time-saver.”
There’s also a tool for digital ops that checks line items against ad creative and IOs and also handles taking campaign screenshots, which Stinchcomb describes as a decidedly “laborious and really unfun” task. The result, he says, is fewer botched campaign launches and less inventory lost to make-goods.
But the bigger bet is media planning.
Right now, it takes a planner four or five hours to turn an RFP into a plan, Stinchcomb says, in part because they have to check inventory across multiple different systems. The Journal is testing a tool, however, that can read RFPs, look at past campaigns for similar audiences and build a recommended plan in about an hour.
With the team producing thousands of plans a year, that’s a lot of hours back in their lives. The time and effort saved is being redeployed toward larger million-dollar-plus RFPs, where Stinchcomb says the Journal’s win rate has gone up five percentage points. A wider rollout of the tool is planned for early next year.
“Bringing it all together is a great strategy,” he says. “It’s an efficient way to bring in chunks of revenue.”
But there’s one place where AI isn’t welcome, though, and that’s in the writing of branded content, which is a growing business for the Journal. WSJ Intelligence, its custom content studio, takes its cue from the Journal’s newsroom. The work might be “facilitated by these tools,” Stinchcomb says, “but not written by them.”
“AI is great at sort of rehashing what’s out there,” Stinchcomb says. “It’s not necessarily great at breaking news, as it were, or bringing a fresher perspective.”
Branded content has to be as interesting and engaging as anything else that’s out there competing for a reader’s attention.
And if it commits the sin of being boring or derivative, Stinchcomb says, then “[people] don’t read it, and I’m not getting paid.”
Also in this episode: Why less than 5% of the Journal’s digital ad revenue is programmatic, how WSJ Intelligence keeps its custom research credible, News Corp’s “woo and sue” approach to AI licensing deals and Stinchcomb’s past life as the singer in a “tragic emo-style” band in high school.
For more articles featuring Josh Stinchcomb, click here.
来源:AdExchanger · adexchanger.com