How to Succeed with an AI Wrapper (Part 2): Own the Ecosystem

Copilot is not impressive right now. It is clunky, inconsistent, and often underwhelming compared to standalone models. Many early users try it once, decide it is worse than ChatGPT, and move on. That reaction misses the point. Copilot is not trying to be the best AI. It is trying to be the most unavoidable one. 

Copilot matters not because it is the inevitable future for everyone, but because it illustrates one viable way an AI wrapper can survive. 

Teams Beat Slack Because Ecosystems Beat Products 

The right historical comparison is Teams versus Slack. Slack was the better product. Cleaner interface, faster iteration, more pleasant to use. Anyone who used both early knew this. And yet Slack lost. It lost because Teams shipped bundled with Office 365, inherited identity, security, compliance, and governance by default, and appeared automatically inside organisations that already ran on Microsoft. Slack was software. Teams was infrastructure. 

Lawyers did not choose Teams because they loved it. They tolerated it because removing it would have required effort, approvals, retraining, and an awkward conversation with IT. That is what network effects look like in law. 

Copilot Inherits the Microsoft Moat 

Copilot sits in exactly the same position. It lives inside Word, Outlook, Excel, Teams, and SharePoint. It inherits permissions, audit logs, retention policies, and identity controls without asking for them. It does not ask firms to change how they work. It changes how their existing tools behave. That alone makes it structurally powerful, regardless of how it benchmarks against the latest standalone model. 

Configured properly, with a disciplined SharePoint structure and sensible document hygiene, Copilot becomes the best search I have ever used over my own documents and legal material. Not because it is clever in isolation, but because it understands relationships and context inside the ecosystem where the work already lives. Sorry, Ailira. That one hurt to write. 

This Is a Blueprint, Not a Commandment 

This is not the only way for an AI wrapper to succeed. It is one viable path. Wrappers can survive if they own an ecosystem lawyers already depend on, or become so deeply embedded in one that removal is painful. 

That ecosystem might be a practice management system, a document management system, a billing platform, or a compliance engine. The specific category matters less than the principle. If your AI is optional, it will be treated as such. If it is infrastructural, it gains durability. 

Yes, this favours incumbents. Yes, it feels unfair. But the only person who says life is fair is selling you something. 

The Uncomfortable Takeaway 

If proprietary data was the first moat, ecosystem ownership is the second. It is not mandatory, but it is powerful. Wrappers that understand this can build something durable. Wrappers that do not are not building products. They are buying time. 

Copilot is simply the most visible illustration of this path. The lesson applies far beyond Microsoft. If you want to succeed as an AI wrapper in law, being brilliant is optional. Being embedded is not. 

This article is Part 2 of the How to Succeed with an AI Wrapper series by Adrian Cartland. Part 1 examined why proprietary data is the first and most fundamental moat for any AI wrapper in law. Read it here.