The race to embed AI into every corner of the corporate world is on, and it’s creating a fascinating, high-stakes land grab. But this isn’t about who has the flashiest chatbot. The real prize is something far more fundamental, something happening a layer or two beneath the surface. The battle is for AI infrastructure ownership – the digital ground upon which every future business decision will be built. Who gets to own this foundational layer will, quite simply, control the brain of the modern enterprise. And if you’re not paying attention to who’s building the plumbing, you might wake up to find someone else controlling your water supply.
This brings us to one of the most interesting companies in this space: Glean. Its journey tells the story of this entire shift.
The Quiet Evolution of Corporate AI
Let’s be honest, the first wave of enterprise AI was a bit underwhelming. We were sold a vision of sentient assistants and got glorified chatbots that could barely handle a holiday request form. They were front-end novelties, digital receptionists with a limited script. But the game is changing, fast. The new frontier isn’t about simply answering questions; it’s about executing complex, multi-step tasks that require access to the messy, sprawling, and siloed data that makes up a company.
This is precisely where Glean has engineered a clever pivot. As detailed in a recent TechCrunch interview, the company started life as a very smart enterprise search tool. Now, it’s billing itself as an “AI work assistant.” That sounds like more marketing jargon, I know. But the distinction is crucial. Glean isn’t trying to be the interface you chat with; it aims to be the intelligence underneath all the other interfaces—be it a Microsoft Copilot, a custom-built app, or a Slackbot.
The Brains Behind the Operation
So, what exactly is an “AI work assistant” in this context? Think of it less as a chatty colleague and more as the company’s central nervous system. Its job isn’t to talk to you, but to connect every other application to the company’s collective knowledge, securely and intelligently. For any of this to work, two things are absolutely non-negotiable: deep workflow integration and rock-solid permission management.
An AI can’t help a sales director if it can’t access Salesforce data, review past email chains in Outlook, and check product specs in a Confluence document. Seamless integration is everything. More importantly, the system must understand the intricate web of permissions. It has to know, with absolute certainty, that the finance director can see payroll data, but a marketing intern absolutely cannot. This is a monumental challenge in enterprise architecture, and it’s where most AI initiatives fall flat on their face. Getting it wrong isn’t just inefficient; it’s a catastrophic security breach waiting to happen.
Insights from the Man Building the Plumbing
During a discussion at Web Summit Qatar, the Glean CEO, Arvind Jain, laid out this vision with striking clarity. Jain, a veteran of Google’s search team, understands that the value isn’t just in finding information, but in synthesising it within the right context and security constraints. His strategy is to make Glean the indispensable, neutral layer that powers everything else.
Investors are clearly buying it. The company recently raised a cool $150 million, pushing its valuation to a reported $7.2 billion. This isn’t money for a shiny demo. This is a massive bet on the idea that the foundational intelligence layer of a company should be a specialised, independent platform. Jain isn’t trying to out-build Microsoft or Google at the application layer; he’s selling the picks and shovels—or perhaps the digital railway—that everyone else will need to build upon. It’s a classic aggregator play, but for corporate intelligence itself.
A Battle of Giants and Specialists
Of course, Glean isn’t operating in a vacuum. The competitive landscape is brutal. Microsoft is weaving Copilot into the very fabric of its Office and Azure ecosystems. Google is doing the same with Gemini across Workspace and its cloud platform. Their strategic advantage is immense: they already own the digital desktops of millions of workers. Why would a company pay for a separate AI layer from Glean when they can get a seemingly “good enough” version bundled with their existing software?
This is the central question for any business AI strategy. Do you go all-in with a single vendor like Microsoft, creating a deeply integrated but potentially locked-in ecosystem? Or do you opt for a more modular approach, using a specialist like Glean as a neutral knowledge hub that can feed intelligence into various best-of-breed applications? As the TechCrunch article highlights, this is a classic tech power struggle, echoing the platform wars of previous eras.
Cutting Through the Hype
Let’s be sceptical for a moment. The air is thick with AI hype. We’re promised revolution and often get mildly improved spreadsheets. The real test of any business AI strategy is whether it solves a tangible problem or just generates impressive-sounding nonsense. Can your AI answer a question like, “Summarise the top three technical challenges we faced in our key Q3 projects and link me to the engineering tickets that resolved them”?
This is not a simple search query. It requires the AI to understand natural language, access project management systems, read technical documents, and respect who is allowed to see that information. This is practical, high-value work. This is the promise of a truly integrated system, and it’s a world away from asking an AI to write a poem about synergy. The companies that focus on these practical, complex use cases are the ones that will win.
Who Will Hold the Keys?
So, we come back to the main question: who will win the fight for AI infrastructure ownership? The tech giants are betting their scale and existing foothold will make their offerings irresistible. Specialists like Glean are betting that enterprises will value neutrality, specialisation, and control over their own core intelligence.
The future of enterprise architecture hinges on this decision. Choosing a path will have decade-long implications for flexibility, cost, and innovation. The allure of an all-in-one solution from a tech giant is powerful, but it risks creating a dependency that could be difficult to escape later. A thoughtful approach to workflow integration and a clear-eyed view of who owns the central “brain” is no longer just a technical detail; it is the most important strategic decision a company can make today.
What do you think? Is a neutral, third-party AI layer the smarter long-term bet, or is the convenience of an integrated suite from a tech giant simply too compelling to pass up? The choices businesses make in the next 24 months will define their competitive landscape for years to come.
