📰 Key Takeaways

Relay, an AI-powered workflow automation tool positioned as a Zapier alternative, shut down completely in August. It let users automate email and task workflows with AI agents, but as big platforms like OpenAI and Google built similar automation straight into their own tools, five-year-old Relay gradually lost its reason to exist as a standalone product. Chatbot platform Character.ai also filed for bankruptcy recently, just five years after launch. According to S&P Global Market Intelligence, about 42% of AI projects end up abandoned by their parent companies, because of lack of funding, technical challenges, competition, difficulty scaling, or weak user demand. The article puts together a growing “AI graveyard” list that tracks AI products and investments that shut down, pivoted, or fell far short of expectations, in the hope that these cases offer lessons for the industry.

Even the big players aren’t immune. OpenAI tried to turn ChatGPT into a “super app”: the July 9 update combined modes like “Chat,” “Codex,” and “Work,” and renamed the original version “ChatGPT Classic.” But users called the new interface confusing and cluttered, and OpenAI rolled the redesign back soon after. On the other hand, OpenAI had better results shutting down several standalone apps and folding their features into ChatGPT. ChatGPT Atlas, the standalone AI browser, stopped service on August 9, less than a year after launch. Operator, which could browse the web and complete tasks for users, followed the same path. Image generation was gradually built into ChatGPT, which shrank DALL-E’s role as a standalone entry point. The video-sharing platform Sora was shut down in March 2026 because of high operating costs and poor user retention. As for Apple, when it announced Apple Intelligence in 2024, it pitched a new Siri that could understand context, track what’s happening across apps, and get tasks done, but the release schedule has slipped several times. The original summary cuts off here; see the source link for the full details.

💬 JudyAI Lab Take

Relay, an AI workflow tool positioned as a Zapier alternative, shut down completely in August. What makes this worth paying attention to is that it didn’t close because the tech wasn’t good enough. It closed because OpenAI and Google built similar features straight into their own platforms.

The original article shows this isn’t a one-off. Five-year-old Relay lost its reason to exist as a standalone product, Character.ai filed for bankruptcy just five years after launch, and even OpenAI itself shut down standalone apps like ChatGPT Atlas and Operator and merged their features back into ChatGPT. According to S&P Global Market Intelligence, about 42% of AI projects end up abandoned by their parent companies, because of lack of funding, competition, difficulty scaling, or weak user demand. For AI builders, this means a thin wrapper that does only one thing can easily get absorbed by the platform’s next update. The ChatGPT super app redesign is another reminder: bundling things together doesn’t automatically make them better, and users will send back a cluttered interface just the same. For a product to survive, it needs a reason to exist that the platforms won’t build in anytime soon.

Here’s what I’d suggest: make a list for your own product of which features the big platforms are most likely to build in next. Then ask yourself, if you took those features away, what reason would users still have to stay?

📅 Source Info

🔗 Further Reading