Microsoft and Google: Rethinking Purpose in the Age of AI

The Integration of AI: A Poorly Posed Question

During a product planning session, five managers gathered around a table, with a roadmap displayed on the screen and a budget conversation already decided. The question on the agenda was: how do we integrate AI into our product? This question, while common, is actually poorly posed. It assumes that the initial goal of the product remains valid and that AI is merely an enhancement. However, no one questions the relevance of that goal in the age of AI.

This is the question that all product teams are currently grappling with. And it’s the wrong question. Not because integration is bad, but because this question assumes that the product's goal is still valid — that the reason for the product's existence is correct, and that AI is simply an enhancement that makes it better. Add intelligence here. Automate this workflow. Insert a co-pilot. Announce the launch.

No one in the room paused to ask the more difficult question: what is the purpose of this product now that intelligence can inhabit it?

These two questions are not the same. The first optimizes what exists. The second questions whether what exists should continue in its current form. The first is a product decision. The second is an intellectual decision.

The Gap Between Questions

This gap — between the question everyone asks and the one almost no one takes the time to ask first — is the subject of this article.

The Wrong Question

Every roadmap session I have observed over the past eighteen months follows a version of the same scenario: identify where AI fits in, prototype a feature, test adoption, launch. The framework is solid. The underlying assumption is not.

The assumption is that the product's goal — the reason it exists, the work it was designed to do — is still the right one. AI is just a new capability added to that existing goal.

But what if the goal itself has changed?

Jakob Nielsen identified this change in 2023 by calling AI the "first new user interface paradigm in 60 years." His argument was precise: we have shifted from command-based interaction — where the user tells the computer how to do something — to an intention-based outcome specification, where the user tells the computer what they want. This reversal does not just change the interface. It changes what a product must be.

Ethan Mollick pushed the argument further in Co-Intelligence: AI is not a tool that is added to a workflow. It is a collaborator that changes the very nature of work. When work changes, the product designed to organize that work must also change — otherwise, it becomes an obsolete artifact that people use out of habit rather than necessity.

Clayton Christensen's Jobs to Be Done framework makes the consequence concrete: customers do not buy products. They hire them to make progress in specific circumstances. When those circumstances change — when intelligence can directly solve the intention — the work for which the product was hired may no longer exist.

The Hidden Evidence

In August 2024, Satya Nadella sent an internal memo to Microsoft. He directly examined the founding vision of the company he leads and wrote:

"When Bill founded Microsoft, he did not just envision a software company, but a software factory, unbound by a single product or category. This idea has guided us for decades. Today, it is no longer enough."

The CEO of Microsoft — the company that makes Word, Excel, and PowerPoint — looked at the model that built one of the most valuable companies in history and declared that it was no longer sufficient. Not "needs an update." Not "requires new features." Just insufficient. Structurally insufficient for what comes next.

And here is the tension that this memo does not resolve: Copilot, as it exists today, is built on that same legacy. Intelligence has been added. The goal has not been reexamined. Word still simulates a typewriter. Excel still simulates a ledger. PowerPoint still simulates a slide projector. Extraordinary simulations — refined over decades, loved by hundreds of millions. But simulations of tools designed for a world before AI. And by Nadella's own public acknowledgment, he is not satisfied with the current state of Copilot.

This is not a failure of ambition. It is an honest illustration of the brutal difficulty of moving from recognizing a gap in purpose to effectively closing it. Earlier this year, I wrote about architectural change — how the environment we have designed over the past forty years is being structurally replaced. But what I have realized since is that the architectural question is actually the second question. The first concerns purpose. And that is the question almost no one asks.

The Google Example

Now let’s look at Google.

Google's flagship product was never the answer — it was access to the answer. In 2025, AI Overviews appeared in 60% of queries in the United States, clickless searches soared to 69%, and organic click-through rates dropped by 61%. The product that once gave you a map to the answer is now compelled to become the answer itself. This is not a feature update. It is an existential redefinition.

Meanwhile, Anthropic made a completely different bet. Rather than building a product that competes with Word, Gmail, or Notion, they built the layer of intelligence that these products connect to. Intelligence itself is the interface. Everything else revolves around it. It’s the same structural logic as Airbnb not owning houses or Uber not owning cars — the company is not the asset; it’s the connection between the user and what they need.

The Market Has Priced the Gap

In February 2026, the financial press coined a term for what happened to software stocks: the "SaaSpocalypse."

In about 48 hours, approximately $285 billion vanished from the valuations of SaaS companies. Atlassian dropped 35% after the number of enterprise seats declined for the first time in the company's history. Salesforce fell by 28%.

Bloomberg's summary is worth pondering: "Wall Street looked at the speed of progress in agentic AI and concluded that hundreds of SaaS companies built on a per-seat pricing model were structurally overvalued."

Structurally overvalued. Not facing headwinds. Structurally — meaning the foundation itself.

Forrester stated that "SaaS as we know it is dead." TechCrunch named what the market was actually valuing: "Being 'in the cloud' is no longer enough. You must be the intelligence." Bain & Company framed the choice clearly: "Disruption is mandatory. Obsolescence is optional."

The Cost of an Unexamined Purpose

Per-seat pricing did not collapse because AI became capable. It collapsed because the purpose around which those seats were built — humans performing workflows via a user interface — was no longer the only available model. The gap in purpose was still there. AI made it financially visible.

$285 billion in 48 hours. That’s the cost of an unexamined purpose when the market finally takes a look at it.

The Difficulty of Implementation

I must be honest about the difficulty of this situation.

The argument I am making seems clear. In practice, it is brutal.

Most products have millions of users who depend on them as they are. Most teams operate within organizations with years of technical debt, committed roadmaps, and real constraints on how much they can question things from the ground up. "Reexamine your purpose" is easy to write. It has a real cost to execute.

There is also a legitimate counterargument that deserves to be taken seriously: not all products need a purpose revolution. Some software exists to accomplish a specific, limited, and well-defined job — and AI allows for better execution of that work. That is a valid integration. Forrester's analysis confirms this: vertical and domain-specific SaaS products — those that know exactly what they are for — are expected to grow from $133.5 billion to $194 billion by 2029. Specificity of purpose is a defense. It is the unexamined purpose that is the vulnerability.

The Important Distinction

But here’s the distinction that matters: there is a difference between knowing your purpose and inheriting it. Most product teams today are in the latter condition. They did not choose the purpose — they inherited it from those who shipped the first version years ago, under different constraints, with different capabilities. The question "what is it for?" has never been explicitly asked. It has been assumed.

And the gap works both ways. Humane's AI Pin had the right vision and no inherited baggage — yet it failed for the same reason. The founders never asked what the purpose of their product was in relation to the ecosystem in which users were already operating. They built a new building instead of reorganizing an existing one. The gap in purpose is not just about inherited products. It is about any product built without an honest answer to the first question.

This assumption — or this omission — is the vulnerability.

Three Questions Before the Next Roadmap

I want to be clear about what I am not saying. I am not saying to tear everything down. I am not saying that every company must become an AI-native startup.

What I am saying is simpler: before your team opens the roadmap and starts planning AI features, someone in the room must honestly answer three questions. Not theoretically. Honestly.

The first question concerns the work. What work is this product engaged to accomplish?