On July 6, 2026, Cloud Wars published an excerpt from a keynote by Thales Teixeira, a professor at UC San Diego’s Rady School of Management. His central argument: the enterprise software winners in the AI era will be those that organize around customer outcomes, not products. He distilled it to a simple lesson—customers don’t want a drill, they want a hole. For Microsoft, which is embedding AI into every layer of the Windows and Microsoft 365 stack, this principle isn’t just academic. It’s a blueprint for how IT leaders should evaluate the coming wave of feature releases, license changes, and workflow promises.
The keynote, delivered at a gathering of business leaders and reported by Cloud Wars, drilled into a familiar but often ignored truth. Most enterprise software is sold as a product: a set of features, a licensing matrix, a support agreement. But what the buyer actually needs is a business result. Teixeira used the power tool analogy: a drill is a means; the hole is the end. In software, the “hole” might be a completed sales report, a resolved customer ticket, or a forecast that prevents a shortage. AI, he argued, lets companies sell the outcome directly—not just faster drills, but the hole itself, delivered automatically.
What does that look like in practice? For Microsoft, the answer is Copilot. Not a stand-alone product, but an outcome layer woven into Word, Excel, Teams, and soon the Windows shell. Ask Copilot to “summarize last quarter’s performance” and it drafts a report. Ask it to “find me a time slot for the team next Tuesday” and it schedules a meeting. These aren’t feature improvements; they’re task completions. That’s the hole.
The shift from features to finished work
The news isn’t that Microsoft wants to sell subscriptions instead of perpetual licenses—that transition happened a decade ago. The change Teixeira pinpointed is more fundamental. AI can now package human judgment alongside raw data processing. A traditional enterprise search tool would return a list of documents. Copilot with Graph grounding reads those documents and synthesizes an answer. The buyer no longer needs to train employees on search syntax; they get the answer directly. The drill (search) has been replaced by the hole (the synthesized insight).
This outcome orientation is already visible in Microsoft’s roadmap. Windows 11’s 2026 update includes Recall, which isn’t a file-by-file index but a semantic memory of everything you’ve done on the PC. It’s designed to answer “where was that budget figure I saw?” not “open this folder.” Enterprise agreements for Microsoft 365 are increasingly measured by active usage and productivity scores rather than just seat counts. The underlying shift is away from selling tools and toward guaranteeing results.
Why home users, IT pros, and developers should care
The drill-and-hole lesson lands differently depending on your role.
For home users, Copilot in Windows can feel like an add-on. But its real purpose is reducing friction: adjusting system settings by typing plain English, summarizing web pages in Edge, creating images for a school project. These are outcomes that previously required multiple steps and app switches. The value isn’t the Copilot button; it’s the task that gets finished without you becoming an expert.
For IT administrators, the message is cautionary. When Microsoft pitches Copilot for Microsoft 365 at $30 per user per month, the cost isn’t for a product—it should be for a measurable increase in business output. Yet many organizations cannot quantify the return. If you can’t define the hole you’re buying, you’re paying $360 a year per user for a very expensive drill. Before adopting, IT must work with department heads to set outcome metrics: hours saved on report generation, reduction in help-desk tickets, faster onboarding completion.
Teixeira’s framework suggests a practical audit: list the top five time-consuming tasks your company performs in Office apps. Ask whether Copilot accelerates them and by how much. If the vendor can’t provide real-world data, you may just be buying feature count.
For developers and ISVs, the lesson is even sharper. Microsoft’s platform play—Copilot Stack, Azure AI Studio, plugins—gives you the drill bits. But your customers are also buying holes. A line-of-business app that adds a chat interface but still requires users to navigate five screens to approve a purchase order has missed the point. The winning approach is to embed AI where it finishes a workflow: voice command that triggers an entire approval chain, anomaly detection that auto-generates a corrective ticket in ServiceNow. That’s outcome-first development.
How we got here: from boxes to business results
Enterprise software wasn’t always sold this way. For decades, the product was king. Windows 95 sold on features like the Start menu and plug-and-play. Exchange Server sold on mailbox limits and clustering. Competition was feature-checkbox warfare. The rise of SaaS forced a partial rethink: you weren’t buying a CRM product; you were subscribing to a customer relationship outcome, with uptime and updates handled. Salesforce’s “no software” campaign in the early 2000s was essentially a hole-selling pitch.
Microsoft adapted with Office 365, then Microsoft 365, bundling communication and collaboration into packages aimed at vertical outcomes—Frontline Worker, Information Protection. AI accelerates this because it can directly digest a company’s data and produce something final: a contract draft, a risk score, a translated proposal. That’s why the drill-and-hole distinction matters now more than ever: the technology can finally deliver the hole without the user needing to master the drill.
Teixeira’s research (as cited by Cloud Wars) shows that companies structured around customer outcomes grow faster and retain customers longer. They are harder to dislodge because switching costs are outcome-based, not product-based. If you rely on Copilot to auto-generate your monthly board deck from live data, ripping it out means more than losing a tool—it means rebuilding an entire reporting workflow.
What to do now: an outcome-driven IT checklist
- Map your software portfolio to outcomes. List every major enterprise tool, then write one sentence describing the business result it produces. If you can’t, that tool is a drill—evaluate if a cheaper drill or an outcome-based alternative exists.
- Pilot Copilot with a productivity metric. Choose a department (finance, HR, marketing) and define a baseline: hours per month spent on routine documentation, meetings scheduled manually, or customer queries handled. Run a 90-day Copilot trial and measure the delta. The data will justify licensing or expose a mismatch.
- Demand outcome SLAs from AI vendors. When Microsoft or others release AI features, press them: what percent of tasks are fully automated? How is error rate measured? Can the system integrate with existing LOB apps to deliver hole-completing workflows rather than just chat interfaces?
- Train users on outcomes, not buttons. Adoption of AI tools often fails because people are taught features (how to prompt Copilot) instead of targets (how to finish a quarterly report in half the time). Frame training around “what work gets done today that you hate doing manually”—and show how to achieve it.
- Watch the Windows client. Upcoming Windows updates will blur the line between OS and outcome engine. Recall and Click-To-Do are designed to fulfill user intent rather than expose file systems. For IT, this means rethinking endpoint management: instead of locking down the Start menu, measure how often employees find information faster.
Outlook: The hole economy arrives on your desktop
Microsoft’s next moves are already visible on the horizon. Integration of Copilot with Power Platform will allow citizen developers to build automated outcomes—a budget approval that triggers a Teams post and an Excel update—all from natural language descriptions. Dynamics 365 Copilot already generates customer opportunity summaries for sales teams. As these systems connect, the distinction between “using software” and “getting work done” collapses. The most successful enterprises won’t be those with the most feature-rich tools, but those that can identify and buy the right holes.
Teixeira’s lesson lands at a critical moment. With every major vendor now shipping AI, the temptation is to compare benchmark scores and model parameter counts. That’s drill thinking. The real question for Windows users and IT leaders is simpler: what’s the hole you need, and who can deliver it?