AI Product Closures Show the Risks of Stand-Alone Tools
Relay, an AI-powered workflow automation product positioned as an alternative to Zapier, shut down entirely in August. The five-year-old service used AI agents to automate email and task workflows, but faced a market in which OpenAI, Google and other large platforms were building comparable automation into their own products.
Its closure joins a widening set of AI products, startups and internal company initiatives that have been shut down, delayed, pivoted or absorbed into broader platforms. S&P Global Market Intelligence estimates that about 42% of corporate AI initiatives are ultimately abandoned by the companies behind them.
Why AI projects are being abandoned
The reasons vary across products. Projects can run out of funding, encounter technical limitations, fail to scale, face stronger competition or struggle to establish enough user demand. In several cases, the difficulty is not creating an AI feature but maintaining a compelling stand-alone destination once larger services offer similar functions to existing audiences.
Huxe, an AI audio app founded by former Google NotebookLM developers, closed in May 2026 as larger platforms began offering comparable experiences. Yupp, which enabled side-by-side comparisons of more than 800 AI models at its peak, closed in March 2026 after its founders concluded that it had not achieved sufficient product-market fit.
Costs have also shaped outcomes. Figgs AI attracted more than 1 million users for customizable AI characters, but its developers said sustaining a free service was too expensive. Humane raised $230 million for its AI Pin, yet performance problems and a battery-fire warning involving its charging case preceded the February 2025 shutdown of the AI Pin business. HP acquired most of Humane’s assets for $116 million.
Large platforms face setbacks too
Scale has not prevented prominent companies from retrenching. OpenAI rolled back a July 9 redesign that combined Chat, Codex and Work modes after users criticized the interface as confusing and cluttered. The company also discontinued the stand-alone ChatGPT Atlas browser on August 9, moving its most useful capabilities into ChatGPT, and folded Operator’s web-task capabilities into the same product.
Apple repeatedly postponed its more context-aware Siri, first highlighted with Apple Intelligence in 2024, citing engineering issues and bugs. The delay contributed to a $250 million settlement over claims about the marketing of iPhone 16 AI capabilities. Siri AI appeared in the iOS 27 beta in July, with an English-language rollout beginning this month.
Microsoft delayed Recall for nearly a year after its planned Windows feature, which periodically captured screenshots to make a user’s activity searchable, prompted privacy and security concerns. A cybersecurity researcher later demonstrated a tool that could extract and display Recall-captured data, sustaining questions about the safeguards.
What businesses should take from the pattern
For businesses evaluating AI tools, the record argues for testing demand, reliability, operating cost, security and vendor durability alongside the quality of a demonstration. A useful capability may be discontinued, delayed or incorporated into a broader platform, so deployment plans should account for integration changes and alternatives before the tool becomes central to a workflow.