When AI Tools Vanish

I received an email this week with difficult news: Relay.app, a workflow automation tool I've been exploring for the last six months, is shutting down. Free users lose access on August 15, 2026; paying customers get until September 14. The company thanked us for our loyalty and promised a smooth transition, but the message was clear: this product is done.

What made Relay.app noteworthy wasn't just that it worked, it was how it worked. Compared to similar automation platforms, it was refreshingly intuitive, easier to navigate, and significantly cheaper. It did one thing well: making workflow automation accessible to people who aren't engineers. For the last half-year, I've watched it simplify tasks that would otherwise require either expensive tools or technical expertise.

But here's what struck me as I read the shutdown notice: Relay.app didn't fail because it was a bad product. It failed because the market fundamentally shifted beneath it. The capabilities that made Relay.app valuable six months ago can now be replicated, often more powerfully, through OpenAI's code generation tools and Anthropic's Claude Cowork. The niche it occupied is being absorbed into broader platforms. And Relay.app, despite its elegance and efficiency, couldn't compete with companies that have the resources to expand into every adjacent space.

This isn't an isolated incident. It's a preview of a larger consolidation happening across the AI tools landscape.

The Rise of "Too Big to Fail" in AI

We're watching the emergence of mega-platforms in AI (OpenAI, Anthropic, Microsoft, Google) that are expanding relentlessly into adjacent capabilities. Each release expands their scope. What started as language models is becoming code generation, image creation, reasoning engines, and now workflow automation. These platforms have network effects, billions in funding, and the ability to absorb losses while iterating. They're becoming systemically important to how work gets done.

Meanwhile, thousands of niche AI tools have launched in the past 18 months, each solving a specific problem brilliantly: workflow automation, AI-powered writing, code generation, design creation, video synthesis, customer service intelligence, data analytics, research synthesis, SQL query generation, personalization engines. Each one filled a gap. Each one had early adopters who built workflows around it, integrated it into their processes, and became dependent on it.

But niche tools face a structural problem. They can't absorb the losses that mega-platforms can. They can't pivot as quickly. And they can't compete when the platforms they might integrate with decide to build the capability themselves.

The 2008 financial crisis taught us what happens when critical infrastructure depends on institutions deemed "too big to fail." When Lehman Brothers collapsed despite being a pillar of the financial system, it triggered cascading failures across everything built on top of it. Banks were bailed out precisely because the infrastructure of the economy rested on their survival. We're entering a similar dynamic with AI.

OpenAI and Anthropic aren't facing bailout threats, but they're moving toward that same systemic importance. They're consolidating capabilities that smaller players built in isolation. And organizations building workflows around niche tools are, unknowingly, creating dependencies on infrastructure that might not survive the next 18 to 36 months.

The Comfort of Specialisation, the Risk of Consolidation

There's a reason people like Relay.app. It was built by people who understood the specific problem deeply. The interface was clean. The pricing made sense. It didn't try to do everything; it did one thing elegantly. That focus created loyalty.

But focus is also vulnerability in a market where generalist platforms are rapidly becoming capable enough to replace specialists.

Lovable, which generates web applications from natural language descriptions, is beautiful technology. Base44, which helps with code generation, is genuinely useful. Perplexity and You.com, which layer AI over search and research, serve real needs. But each one exists in a space that Claude, ChatGPT, or other mega-platforms are actively moving into. Claude just launched Fable for design. OpenAI and Anthropic are expanding into code generation, reasoning, and task automation. Google and Microsoft are integrating AI across their entire suites.

The math is simple: if a mega-platform can capture 80% of what a niche tool does, and bundle it into something users already pay for, the niche tool's value proposition collapses.

What This Means for Organisations

The lesson here isn't that niche AI tools are bad; Relay.app proved they're often better than their generalist alternatives. The lesson is about risk and dependency.

Organisations that have built critical workflows around niche tools should start thinking about what happens when those tools disappear. Not as a catastrophe scenario, but as a business reality. Your automation layer, your code generation workflow, your design process: if it runs on a single niche platform, you're carrying invisible risk.

This doesn't mean abandoning specialised tools. It means being intentional about which ones you embed into your core operations. It means staying aware of which platforms are consolidating toward dominance. And it means occasionally testing whether the underlying capability can be replicated through broader platforms, just in case.

Relay.app's shutdown isn't tragic; it's clarifying. It shows us how quickly the AI landscape can shift. It reminds us that elegance and efficiency don't guarantee survival when market forces are working against you. And it suggests that the future of AI tools looks less like a diverse ecosystem of specialists and more like concentrated power among platforms that can afford to expand relentlessly into every possible niche.

The smart move for organisations isn't to panic or become cynical about niche tools. It's to hold them lightly, maintain awareness of alternatives, and avoid building irreplaceable dependencies on platforms whose survival depends on factors outside their control.

Relay.app's goodbye is one small data point. But it's part of a much larger story about how AI is being concentrated into the hands of a few dominant platforms and what that means for everyone else.

Some Niche Tools and Their Purposes

Cultivating the Future of Work

Empowering businesses through human-centered AI.

© 2025. All rights reserved.

LetUsBuild | UEN: 53522535X