The biggest AI companies have identified one of the most common problems for not just knowledge workers, but for the broader economy, and solved it.

They’re all converging around a similar offering - the all-in-one platform with a chat interface, a vibe coding tool, and agent building. The platform can connect to all of your existing tools so, increasingly, you never have to leave. With so many people feeling bogged down by too many SaaS tools this move makes sense. Which means we could be looking at another layer of platform lock-in on the level of Google or Microsoft.

For business and IT leaders, choosing a platform might feel more like picking your seat on an exploratory submarine than choosing which airline to fly. The platform lock-in problem is not a novel concept to consider when evaluating software. But what’s changed is the speed of development around the new platforms. They’re shipping new features at a breathtaking pace. With record-breaking valuations these companies can seem “too big to fail,” meanwhile the news headlines and frequent outages suggest these businesses are not yet as robust or entrenched as the software giants built in the early 2000’s. All of which makes for difficult decision making.

I joined a 9x workshop where Alexandre Kantjas had a great breakdown of the difference between Claude Chat, Cowork, and Code. Chat for thinking. Cowork for working. Code for building. This feels about right and is how I’ve personally been using the tools. Most work for me now happens in Cowork and I still use chat a lot for strategy, using skills like my Advisory Board Skill. I think Cowork+Chat will continue to be the strongest combination for contractors.

The team behind Claude has been relentlessly dropping new features since the beginning of the year. Here are the ones that matter:

The takeaway from this is that it's important to build malleable systems and consider what the new platform lock-in looks like with AI. Perhaps it’s a non-issue, whichever platform we pick can connect with all our existing tools. But stewarding your AI systems with strong documentation and internal policies will de-risk businesses. Assuring leaders that prompts, frameworks, and agents are transferrable and flexible in the unsteady platform landscape.

When talking to new clients, I often hear similar versions of the same story: We already have so many apps. I know firsthand that project managers and site crews feel burdened by the disjointed apps they’re forced to use. Top-down reporting mandates, time logging, material tracking, material ordering, project management software, notes, punch lists. These things stack up and leave crews feeling burnt out.

My instinct is to help businesses implement tools that don’t interrupt their existing work. While the goal may eventually be to build new systems from the ground up - become AI-native - I think the first step on the road to adoption is quick win implementations. Contractors are under constant stress and pressure to deliver, which means that tools that add friction will sit unused.

Your teams are not clamoring for another app, in fact, they’re more likely to be asking to get rid of apps. Given that reality, leaders need to think about orchestration and how all of these apps can actually work together. And if there are several apps that can be rolled into one, that’s even better. This is why I argue for picking a platform and developing alongside it.

Strip things away to what’s essential, then choose a tool that allows you to connect apps altogether under one roof. Crews will thank you when they can work efficiently in one tool rather than many.

A recent client was able to pilot Claude across a small group for essentially zero new cost by reallocating ChatGPT licenses. Combined with real, hands-on training they have seen 100% weekly active users since we launched.

The starting goal was to pilot their first AI tool and establish a proven deployment method. After 10+ discovery interviews across several departments, the decision felt obvious. We chose to adopt Claude as their AI platform, allowing users to connect it to their existing tools such as MS365.

IT leaders feel pressure to deliver immediate ROI on AI spend. My theory - now proven through client work - is that if we can move internal tinkerers into AI power users, the adoption and use cases will start to flow. Every recently published a great piece by Mike Taylor about this:

https://every.to/also-true-for-humans/seven-things-i-ve-learned-getting-companies-to-use-ai

This connects to the convergence topic covered above. As the tech giants continue racing toward an all-in-one enterprise platform, businesses would be best served to choose a platform and start learning and developing alongside the tech itself.

If you're feeling the platform fatigue, you're not alone. Most of the contractors I talk to are sorting through the same noise. My advice is to pick a lane, start learning, and build your documentation habits now so you're not locked into anything you can't walk away from.