In the era of AI, SaaS and COTS are not dead, if vendors adapt their model

Organizations with internal software development teams are finding they can build quality production-grade software - architecturally sound, secure, easy to use, built on modern technology, and delivered at a pace that meets business expectations - faster and cheaper than ever before — not just for AI-specific use cases, but for any use case.

AI now accelerates the entire development lifecycle, from the first line of code to production. That shift gives organizations more leverage: if a vendor isn’t bringing something genuinely hard to replicate — deep domain knowledge, critical data, or superior User Experience (UX) — the case for buying over building gets harder to justify.

What vendors needs to do to stay relevant

Sofware Vendors benefit from economies of scale that most organizations can’t match internally. The cost of building, hosting, securing, and maintaining a platform is distributed across thousands of customers. A single organization building the same thing absorbs the full cost — and that math often doesn’t work, even with AI lowering development costs. However, that calculus is shifting. In recent years, many vendors have aggressively raised prices — and what was a clear financial win a decade ago is no longer obvious.

Beyond cost, the vendors that will stay relevant are those that compete on dimensions AI-assisted internal development cannot easily replicate:

Domain Expertise: The strongest vendors go beyond the software itself — they bring deep knowledge of the domain they serve. That means pre-built content, defined business processes, proven customer implementation experience, and hands-on support for change management and adoption, not just deployment.

Too many vendors deliver a sofware product but leave customers to figure out the rest. They don’t help align business processes, they ship no reference content, and they delegate system integration and change management to partners who often lack the domain knowledge — adding unnecessary delays and costs.

Data as a Product: Vendors that can serve as a high-quality data source for their customers’ AI use cases will remain relevant — and central — to how those customers implement AI in their organizations. This positions the data feed itself as a new monetization layer, offered via APIs as part of the vendor’s business model.

Many organizations are investing heavily in AI initiatives, but hit a wall: either their SaaS/COTS vendor hasn’t kept pace with AI capabilities, or the platform simply doesn’t hold the data needed to build an effective AI model. Vendors that close that gap have a clear competitive advantage.

User Experience: AI has made it easier than ever to build polished interfaces — but great UX still requires understanding the people using the software. An office worker and a field worker have fundamentally different needs, contexts, and expectations. Vendors that deeply understand their users’ personas and embed that knowledge into their product are offering something genuinely hard to replicate.

This is where many established players are falling behind. Some of the largest vendors in the market — with significant market share — have interfaces that haven’t meaningfully evolved in a decade. As AI dramatically lowers the cost of building great UX, the gap between these incumbents and newer or more agile alternatives will only widen.

When to Choose SaaS/COTS vs. Building Internally with AI

AI has significantly lowered the cost of building software — but maintaining and supporting it is a different story. At the same time, strong vendors still bring things that are genuinely hard to replicate internally. The answer isn’t binary; it depends on what you’re getting from the vendor beyond the software itself.

A good approach for organizations is to frame this as a decission matrix, “choose SaaS/COTS when…” vs “build internally when…”.

Every organization has its own strategy with regards to this topic, but here few common examples

SaaS/COTS:

• The vendor has deep domain expertise you’d spend years accumulating (e.g., regulatory content, industry-specific workflows, niche technology).
• The vendor’s data is central to your AI use case. Speed to value matters more than fit.
• The cost of development & support would exceed the cost of buying the SaaS/COTS solution.
• …

Build internally:

• The use case is still evolving and committing to a vendor too early would create vendor lock-in before requirements are clear.
• The process is highly specific to your organization.
• The SaaS doesn’t expose the data you need.
• You’re embedding AI into a workflow that’s a competitive differentiator.
• …