June 2026 · Anthony Sogg

Salesforce Just Paid a Billion Dollars for the Boring Part

The headline says content platform. The receipt says something more useful: even the company selling you the agents knows they're only as good as what you hand them to read.

Salesforce acquired Contentful on June 1, and the announcement wants you to look at the agents. Look at the price instead.

The company signed a definitive agreement to buy Contentful, a composable content platform used by more than 4,800 brands, in a deal that The Information reported at $1 billion to $1.5 billion. It's expected to close in the third quarter of Salesforce's fiscal 2027. That same week, Salesforce rolled out a slate of marketing agents that qualify leads, build content, and run campaigns.

It's a familiar story, as old as time: A chicken in every pot an AI team in every marketer's hands, marketers as makers again.

But nobody pays a billion dollars for a content management system because they're feeling generous about your blog. They pay it because the thing they're building depends on that system to work at all.

A billion Benioff-buck insurance policy

Salesforce has spent two years telling everyone the future is agentic. Agentforce, Data 360, autonomous everything. And then it turned around and bought a structured content spine to sit underneath the whole thing, which you should read as the confession it is.

Content and data are fuel for the AI machine. Without them, you ain't goin' nowhere (thanks Bob Dylan).

The agents needed something to read, something clean and structured and governed and machine-legible, and there wasn't enough of it lying around, so Salesforce made sure that wouldn't be a problem moving forward. That's not really a content play. It's an insurance policy against your own agents having nothing trustworthy to act on.

And it's a loud build-versus-buy tell. If agents could just conjure brand-ready content and act on whatever was scattered across your systems, you wouldn't need to buy the company that makes content structured and addressable in the first place. The acquisition is the counterargument to the hype, and it's signed by the company doing the most hyping.

A couple of weeks back, I wrote a post arguing that agents read your data, not your hero film, and that the asset you actually need is a structured knowledge base both people and machines can act on. Well, Salesforce just put a number on that asset. About a billion dollars.

Running on empty

The quiet truth in every agentic rollout is that the agent inherits whatever's sitting beneath it. Give it a resolved, consented profile and clean structured content, and it can do genuinely impressive things. Give it four product feeds that disagree, a price that's right on the site and wrong in the markup, and inventory it can't verify, and it'll make fast, confident, wrong calls, then quietly route your customer to a competitor it can actually read.

The bellwether? Look no further than Salesforce's own Q3 filing. Agentforce and Data 360 ARR hit nearly $1.4 billion, up 114% year over year. Gartner, counting it differently, puts paying Data 360 customers up 141% over the same stretch. Take whichever number you trust more, because they're both pointing the same way, and neither one is agent adoption. It's the critical step before adoption.

This is, honestly, a very, very, very good sign. Did I mention it was a good sign?

This is, ostensibly and objectively, the right strategy. These companies are paying real money to secure their success by structuring their data into a state an agent can use, and use well.

The Contentful deal is that same move, just one floor up: first make the data legible, then make the content legible, and only then do you send in the clowns agents.

Gartner already named what is happening

Gartner's 2026 Magic Quadrant for customer data platforms split the market into two futures, and it's worth knowing which one you're standing in.

There's platformization and agentification. Platformization is the data platform becoming the foundation a whole application ecosystem runs on. Agentification is the platform as a thin front door with autonomous agents doing the work on top.

It's no secret that the Salesforce business strategy is platformization through adjacent acquisitions, stitching together a powerful suite even if it's not as connected as they'd like. See the 360 strategic push over the last few years.

These suites (warehouse, data layer, content layer, agents, all under one roof and all on one invoice) aren't unique to Salesforce. Adobe and others are following the same path. (Resistance is, apparently, expensive.)

Roll for initiative

Here's where it gets uncomfortable for everyone who isn't Salesforce, so let's make a game of it. Choose your own adventure. Go ahead, peek forward just like you used to do to see whether the dragon ate you or you saved the day. And yes, I know rolling for initiative is a D&D thing, but that'll weave in here shortly.

You're a marketing leader. A billion-dollar acquisition just lit up your feed. You've got a budget, a board asking what you're doing about AI, and a data layer held together with tape. Choose.

Path one: buy the platform to rescue the stack. You consolidate. You sign for the suite, the agents, and the shiny new content layer, and you hand over the keys. The demo was incredible. The savings pencil out this quarter. Turn to page 47.

Page 47 is two quarters from now. Retention's sagging, and nobody's connected it to the agent you dropped onto a foundation no one in the building actually owns. You never settled the disagreement your teams already had about the data, you just automated it, at machine speed, with less human attention on the result and a bigger invoice to show for it. Oh, and by the way, your IT team now hates you for demanding a specific technology instead of partnering with them on the business problem, which is where their expertise actually lives. Good job.

A capability you can't operate was never an asset; it was a liability with a renewal date. You rolled a natural one. The dragon ate you.

Path two: fix the substrate before you automate it. You spend the first dollars on the boring, unsexy part instead. You resolve the identities, structure the content, and sort out the consent flags. You prove one team can run a capability end to end. And then, only then, you hand that team an agent to amplify something that already works. Turn to page... and there's no shortcut page. That's the point.

It's slower, and it's unglamorous, and nobody's going to throw you a press release for cleaning up a product feed. But the agent you eventually switch on inherits something worth inheriting, and it compounds instead of decaying.

Here you've rolled a natural ten, but your multiplier (intent, strategy, foresight, commitment) means that when you reach the dragon, it's a one-hit knockout.

Both paths cost money. The fallacy is throwing the money at a thing because, in our collective hivemind, tangible feels better. The better play is clearing the land, setting the foundation, the piping, the utility connections, and the drainage before you drop a heavy house on loose soil.

Can you buy your way into success, the way Salesforce is doing? To an extent, yes. But the dollar amount you pay isn't the final cost. There are all the other costs, like trying to shave enough off a square-pegged acquisition to fit your round API inflow. Most companies can't afford that.

And you still can't buy clean, resolved, consented data and structured content that an agent can act on turnkey. Tools that help? Sure. But the hard part is yours. It's the whole constraint, and there's no page number that skips it.

Where I might have this backwards

Maybe I've got this backwards. Maybe generative content gets good enough that a structured CMS matters a little less every quarter, and Salesforce just overpaid to plug a competitive hole against Adobe rather than signal anything deep about data. That's a fair read, and worth holding onto.

But notice that even in that version of the story, the move is toward more structure, not less.

Adobe lashes its CDP to its content tools for the same reason. When two giants independently decide their agents need a governed content and data spine before anyone can trust them, that's not a coincidence. That's the shape of the problem showing through.

So watch the price, not the agents. The agents are the part they want you to see. The billion dollars went to the part they'd rather you ignore: the boring, structured, consented substrate that quietly decides whether any of this works at all.

Salesforce went and bought theirs. The only question left for the rest of us is whether we're willing to do the slow, unglamorous work to build ours, because that bill's coming due either way, whether or not the de-clothed emperors around you admit it.

Anthony

← More writing · If this was useful, or wrong in an interesting way, tell me: hireme@crmjedi.com.