Coté

Your Boss Doesn’t Know What to Do With AI

Enterprise AI Has a Product-Market Fit Problem. Enterprise AI isn’t stalled because the models are weak. It’s stalled because we haven’t discovered product-market fit inside the enterprise yet.

You don’t find real AI value by theorizing in workshops. You find it by running experiments for months inside your actual systems - against real data - in a governed environment.

That requires a platform.

Without one, AI pilots turn into disconnected experiments, shadow infrastructure, and compliance risk. With one, experimentation compounds into institutional learning.

In this video, I break down:

• Why enterprise AI is still in discovery mode
• Why experimentation must be long-running, not one-off
• How governance enables innovation instead of blocking it
• Why a secure platform foundation is the baseline for AI ROI

If you’re thinking about AI strategy, platform engineering, or how to make AI experimentation safe and scalable, this is where to start.

Featured:

• Sean Goedecke.
• Product Manager Playbook.
• Free book! - The Business Bottleneck.
• What is AI Middleware, and Why You Need It.
• Three Questions That Will Define AI In 2026.
• Building an Enterprise MCP Server Marketplace with Tanzu Platform.

From The Book of Lairs, volume one.

If OpenAI fails, the most likely mode is the Yahoo path: not a dramatic collapse but a slow fade into irrelevance through a thousand mediocre product extensions. ChatGPT becomes a utility everyone uses but nobody pays premium for. Enterprise goes to companies with better compliance stories. The consumer product goes ad-supported. Revenue grows but margins compress. The valuation becomes unjustifiable. They never die – they just stop mattering.

🔗 OK, It’s a Bubble. Now Tell Me How It Pops.

Proof of value lies in the results. To date, more than 90% of the top 10,000 VMware customers have purchased VCF, including nine of the top 10 Fortune companies. Leading companies such as Audi, ING Bank, Lloyds Banking Group and Walmart are adopting VCF and deepening their partnerships with Broadcom. Broadcom’s own internal IT teams have adopted this technology and a cloud operating model to consolidate datacenters and toolchains while improving overall system reliability, improving time to provision applications and infrastructure, and decreasing costs. Most important, the number of workloads managed by Broadcom IT increased during this private cloud transformation."

🔗 One Platform for All Workloads - VMware Cloud Foundation (VCF) Blog

“Greedy work” refers to jobs in which earnings are convex in hours - meaning that working longer, more continuous, and less flexible hours is rewarded disproportionately rather than proportionally. {Summarized by the robot.}

And, from Claudia Goldin:

Greedy work can be defined as a job that pays disproportionately more on a per-hour basis when someone works a greater number of hours or has less control over those hours. It could be a rush job, a demanding client who calls at 11 PM, or a supervisor who asks that the worker give up a vacation day for the project. The firm has deemed the additional payment worth it to have the work done over more hours, at a particular time, or during odd hours. The other critical aspect is that the worker agrees to do it at that wage. Supply and demand, all over again.