A Value-First Guide to Surviving the SaaSpocalypseContents
Part II — Context
Chapter 6
Context as Competitive Moat
Here's a question every leadership team should be asking right now: if everyone has AI, what makes yours different?
Because everyone will have it. The models keep getting better and easier to reach. Whatever your competitors can't buy this year, they'll be able to buy next year.
Access to AI isn't an advantage. It's table stakes. So what is the advantage?
The commodity mistake
There's a mistake being made right now that will look obvious in hindsight. Organizations are treating AI as the strategy, when the strategy is what AI works on.
I don't think the model stays the advantage. I expect today's most capable model to be next year's baseline. What can't be made a commodity is what the model works on.
Context is the moat. Everything else is a drawbridge anyone can build.
What can't be copied
A competitor with the same model can't copy four things you have.
Your customer relationships. They can't copy your relationship history. Their data doesn't hold your patterns.
Your business patterns. The discovery that customers who come to one particular kind of session tend to adopt faster than everyone else. That's yours, and every interaction adds to the pattern library.
Your institutional knowledge. Why you priced the way you did. Which integration approach works for which kind of customer. The onboarding sequence that got people to value sooner.
Your proven practices. How you handle an escalation. How you approach a particular industry. The rhythm of what you publish that keeps people coming back. These came out of your context, and they work because of it.
The model makers went looking for context
If the model were the whole product, the companies that make the models wouldn't need to send engineers into your building. In May 2026, two of the best-known ones did something close to that.
On May 4, Anthropic announced a new company with three financial firms: "Anthropic, Blackstone, Hellman & Friedman, and Goldman Sachs announced the formation of a new AI services company." Its job is to bring Claude into the most important work of mid-sized companies, and the announcement is clear about how. "Applied AI engineers from Anthropic will work alongside the firm's engineering team to identify where Claude can have the most impact, build custom solutions, and support customers over the long term." It's equally clear about why: "Companies from community banks to mid-sized manufacturers and regional health systems stand to gain from AI, but lack the in-house resources to build and run frontier deployments." Anthropic's chief financial officer, Krishna Rao, put the demand side in one line: "Enterprise demand for Claude is significantly outpacing any single delivery model."
A week later, on May 11, OpenAI launched a company of its own, called the OpenAI Deployment Company, with more than $4 billion of initial investment. Its people go inside customer organizations; OpenAI's name for them is "Forward Deployed Engineers." The announcement described the work inside each organization as "connecting OpenAI models to the customer's data, tools, controls, and business processes." A typical engagement, it said, "will begin with a focused diagnostic of where AI can create the most value."
Look at what both companies built. Not more seats. Not a bigger interface. Engineers who sit with the people doing the work and build around the workflows already there. OpenAI says it in so many words: connecting its models to the customer's data, tools, controls and business processes.
As I see it, these two announcements show that the model alone isn't the deliverable. Anthropic's own page says it: "Putting Claude to work in an organization's core operations takes hands-on engineering and deep familiarity with how each business runs." The deployment into your context is the work.
The analysts saw it the same way. Writing about the two moves in CIO.com on May 6, Prasanth Aby Thomas quoted Deepika Giri, head of research for AI, analytics and data for Asia Pacific at IDC: "AI model providers are moving beyond being platform vendors to actively shaping the entire AI value chain." By moving into implementation and services, she went on, they were positioning themselves "closer to enterprise outcomes rather than just supplying underlying technology."
That is the argument of this chapter, and here it is from two companies that sell the model.
Context compounds
Organizations with unified context get a loop that feeds itself. Unified context makes the AI useful. Useful AI helps you create value for customers. Value deepens relationships. Deeper relationships produce more context. And around again.
Each turn builds on the one before. The organization that starts building unified context today won't see the full benefit for months. The one that started two years ago is already working with a depth of understanding a newcomer can't simply buy.
The fragmentation disadvantage
Before AI, fragmentation was expensive but survivable. Everyone was carrying the same handicap. After AI, it's a serious competitive weakness. Organizations with unified context can put AI to work on the whole picture. Organizations with scattered data can't.
An AI working from all of your context will do better than one working from a slice of it, and the gap is bigger than the difference in the amount of data. The whole picture shows patterns that no fragment can.
The architecture of your business context sets the ceiling on what AI can do for your organization.
What it means for strategy
Your context architecture is your AI strategy.
A better model on scattered data gives you better-phrased generic answers. A modest model on unified context gives you specific, relevant intelligence you can act on. Context wins.
The time to build context is before you badly need it. For some organizations, February was the moment they found out they needed it. For the rest, it's now. Not next quarter. Not after the planning cycle. Now. Because building context has a slow start: it takes the most effort at the beginning and gives back the most later.
Part III turns to what that looks like when you build it. And if you want to see how an organization makes the change, person by person, The AI-Native Shift is the book for that.