We're on M365. We use Teams. We use Zoom. We have ChatGPT and Claude running as separate subscriptions. That's five tools, five admin consoles, and five different AI contexts with no idea what the others know. On March 10, Google shipped a major Gemini update across Docs, Sheets, Slides, and Drive. A day before that, Microsoft announced Wave 3 of M365 Copilot. Both events landed in the same week, and I've been doing the consolidation math ever since.
Is it worth consolidating your SaaS stack under one AI platform?
For me, the evaluation has never been primarily about cost. When you run the numbers, M365 and Google Workspace end up in the same ballpark. Google raised Workspace pricing roughly 17% in 2025 to bundle Gemini into standard plans. Microsoft is working through its own pricing adjustments, with full Copilot bundling into M365 plans not landing until July 2026. After you account for the standalone ChatGPT and Claude subscriptions we're also paying for, the delta between staying fragmented and consolidating isn't what drives the decision.
What actually matters is benefit gain. The fragmented stack we're running today has a real overhead cost — not in dollars, but in context. Every tool knows only what lives inside it. Teams doesn't know what's in my inbox. ChatGPT doesn't know what's in my Drive. When I'm researching an issue and need to pull together information from three different places, I'm the integration layer. That's the friction I'm trying to eliminate.
Why is Microsoft falling behind on AI integration?
On March 9, Microsoft announced Wave 3 of M365 Copilot — new agentic features, multi-model intelligence, the works. The announcement is fine. The product reality around it is not.
Microsoft's Copilot deployment problem in 2026 isn't feature gaps. It's that the rollout process consistently stalls. Governance gaps. Permission sprawl. An administration interface that's at a genuine complexity breaking point. Frequent UI changes that users hate. Enterprise customers have been vocal about prioritizing AI features over security and stability updates. One analysis of Microsoft's cascading 2025-2026 price changes found a mandatory 25% cost increase on a typical $10M enterprise agreement — EA tier elimination, M365 Copilot bundling, Unified Support multiplier all hitting at once.
The core problem is what I've watched happen with our own M365 instance. Microsoft's product surface is enormous. Realigning Teams, Exchange, SharePoint, OneDrive, and the rest of the M365 stack around a coherent AI experience takes time — time Microsoft doesn't have while Google is shipping every two weeks. The slowness isn't the features themselves. It's the organizational drag of moving a suite that old, that large, that fast.
Microsoft is so slow to move or realign their products that it's actually slowing us down. Google develops so quickly it's almost hard to keep up with.
Can Google Workspace actually replace M365 + Teams + Zoom?
The honest answer: closer than it was 18 months ago, but not complete.
On March 10, Google shipped a meaningful Gemini update to Docs, Sheets, Slides, and Drive. The new "Help me create" feature in Docs lets you describe what you want to build and Gemini synthesizes a first draft using content from your Drive, Gmail, Chat, and the web — not a blank-slate chatbot response, but something informed by your actual work. That's the kind of integration that changes how you think about AI in a productivity suite.
Google Meet is not Zoom. I'll say that plainly. For external calls with customers, Zoom still has an edge in meeting features and user familiarity. But here's the question I keep coming back to: do I want best-in-class video sitting in isolation, or a slightly lesser video product where the AI actually knows what I've been working on? For a small team that can adapt, that tradeoff is worth examining.
What's the real risk of rolling out AI tools too fast?
Definitely too much too fast. But not in a bad way.
We're small enough to be dynamic, but the process changes would happen so rapidly that frontline employees wouldn't be able to keep up with back-office changes. There would be a real delineation in productivity between users who see it happening and users who don't. That gap is the actual risk. Not the AI making mistakes. Not the data concerns. The gap between the people moving at AI speed and the people who haven't gotten there yet.
The answer isn't to slow-roll AI at the company level. It's to roll it out by function, not by flip. Identify who gains the most from it first, get them comfortable, document the workflows that are actually changing, then expand. The risk isn't the tools — it's deploying change faster than your team can absorb it.
Where's the actual productivity gain right now?
This is the concrete part.
We're slow-rolling Claude Cowork right now, and it's already measurable. Yesterday I was on four or five calls about outbound caller ID being flagged as spam — a gnarly issue that touched transcripts, vendor emails, regulatory PDFs, and customer expectations all at once. After the calls, I took everything: the transcripts, the emails, the PDFs — fed them into Cowork, asked it to research and debrief the situation, then draft an update to the customer.
What used to take hours of iteration — write something, paste it into ChatGPT, rewrite it, rewrite it again, finally land something worth sending — took minutes. One email out to the CIO of a large healthcare company, clear and accurate, without the painful back-and-forth. That's not a demo. That's Tuesday.
The ROI from AI tools isn't usually one dramatic win. It's the compounding of dozens of small tasks that move faster. The rapid pace of getting through small tasks is real. Even deep research tasks — pulling together context across multiple sources before a critical call — are measurably faster. That's the case for consolidation: when the AI knows your email, your files, your calendar, and your chat history, the gap between "what happened" and "what do I do next" shrinks.
I haven't made the call yet. But the direction is clear.