Enterprise AI strategy: why buy often beats build

February 3, 2026

The question of whether enterprises should buy or build AI solutions was a central theme at the 2025 Central Park AI Forum, hosted by Norm AI.

Across the day, discussions focused on the realities of deploying AI in high-stakes environments, particularly legal and regulatory functions, where reliability, explainability and long-term maintenance matter as much as innovation. A fireside chat between Henry Moniz, chief compliance officer at Meta, and Norm AI founder and CEO John Nay captured many of these tensions.

One of the clearest messages was that building software is not the same as owning it over time. While internal teams may be enthusiastic about developing new tools, sustaining production-grade systems is far less appealing. This challenge becomes acute in compliance and legal use cases, where systems must be continuously updated to reflect regulatory change. “If you’re an engineer, you might be onto the next sexy thing you want to build. You might not necessarily be pushing […] updates and improving. It’s not as fun to maintain as it is to build.” — Henry Moniz, CPAIF 2025.

Talent allocation also featured heavily in the debate. Engineering teams capable of delivering robust, compliant AI systems are scarce, and enterprises must decide where that expertise delivers the greatest value. For most organisations, legal and compliance software is essential but not differentiating. “There’s a cost benefit, right? Like if you have engineers, we have a lot of engineers. All these companies have a lot of engineers. But do you want them working on products that are accretive and commercial and they can really build the bottom line?” — Henry Moniz, CPAIF 2025.

Forum speakers cautioned against two common pitfalls. The first is over-reliance on narrow point solutions from multiple vendors, which can fragment data and create operational complexity. The second is partnering with vendors that simply wrap foundation models without investing in deep, domain-specific workflows. As models from providers like OpenAI and Anthropic rapidly improve, shallow integrations offer diminishing returns.

The consensus view was that enterprises should favour partners with a clear category focus and the ability to adapt alongside their clients. In regulated, mission-critical environments, buying from specialists with industry-wide visibility often leads to faster deployment, lower risk and more sustainable outcomes than building in-house.

For more insights, read the full story here.