The AI sector is seeing a growing list of projects that have either shut down or fallen short of expectations, according to TechCrunch. For owners and managers of small‑ and mid‑sized businesses, these failures underline the importance of realistic planning, clear value propositions, and careful risk management when adopting new AI tools.
What happened
Relay, an AI‑driven workflow automation service positioned as a Zapier alternative, closed in August after five years, unable to compete with native automation in OpenAI and Google products. Character.ai, a chatbot platform that let users converse with custom bots, filed for bankruptcy in October after legal battles and a costly acquisition by Google left it financially strained.
S&P Global Market Intelligence reports that about 42 % of corporate AI initiatives are eventually abandoned, citing insufficient funding, technical hurdles, competition, scaling issues, or weak user demand. OpenAI’s attempt to turn ChatGPT into a “super app” in July 2026 was quickly rolled back after users complained about a cluttered interface. The company has also retired several stand‑alone tools, including ChatGPT Atlas (a web browser discontinued on August 9), the Operator AI agent, and the image‑generation service DALL‑E as a separate product. Its video‑sharing platform Sora was shut down in March 2026 due to high operating costs and low retention.
Apple’s promised Siri upgrade, branded as Apple Intelligence, suffered repeated delays, culminating in a $250 million settlement over misleading marketing of iPhone 16 AI capabilities. The updated Siri finally appeared in the iOS 27 beta in July, with broader rollout pending.
Microsoft’s Recall feature, billed as an AI‑powered photographic memory for Windows PCs, faced immediate privacy backlash and was delayed for nearly a year while security concerns were addressed. A recent researcher demonstrated that data captured by Recall could still be extracted, keeping the controversy alive.
Other notable closures include Notion Mail, which will cease operations on 22 September 2026 after users preferred external AI agents for email handling; the Humane AI Pin, a wearable AI device that shut down in February 2025 following performance issues and a battery‑fire warning; and the Rabbit R1, an AI companion that, despite early sales, was criticised for unreliable performance.
Huxe, an AI audio app that turned text into podcast‑style audio, closed in May 2026 as larger platforms like Spotify introduced similar features. Yupp, a free playground for comparing responses from hundreds of AI models, shut down in March 2026 after failing to achieve product‑market fit. Figgs AI, a role‑play character platform, ended in 2024 due to unsustainable free‑service costs.
Why it matters for businesses
These examples illustrate that even well‑funded, high‑profile AI projects can stumble when market demand is overestimated or when larger ecosystems absorb their core functionalities. For small‑ and mid‑sized enterprises, the lesson is clear: adopting AI should be driven by concrete operational needs, not by hype.
The 42 % abandonment statistic warns that a sizable share of AI spend may not deliver returns, especially if projects lack clear metrics or if integration with existing tools is weak. Privacy‑related setbacks, such as those faced by Microsoft’s Recall, highlight the regulatory and reputational risks that can arise from data‑intensive features.
Legal challenges, exemplified by Character.ai’s lawsuits over user safety and intellectual‑property infringement, remind businesses that compliance and ethical considerations are essential from day one. Finally, the pattern of larger platforms subsuming niche AI products suggests that differentiation is fleeting; companies must either build strong network effects or focus on specialised use cases that larger players cannot replicate easily.
Practical steps for business owners
- Validate demand before investing – Run a pilot with a small user group and measure concrete outcomes such as time saved or revenue uplift. Abort early if the results are marginal.
- Assess integration costs – Ensure the AI tool can connect to your existing CRM, email, or workflow systems without extensive custom development.
- Prioritise data privacy and compliance – Review how the solution stores and processes data, especially if it involves screenshots or personal information, to avoid regulatory penalties.
- Monitor vendor roadmaps – Choose providers that demonstrate a clear, sustainable product strategy rather than one that may be folded into a larger platform later.
- Plan for exit – Keep data export capabilities and contractual clauses that allow you to switch vendors without disrupting core operations.
The bottom line
The AI graveyard is expanding, with projects ranging from workflow automators to wearable assistants failing for reasons that many small businesses can anticipate. By grounding AI adoption in validated use cases, ensuring seamless integration, and keeping a close eye on privacy and vendor stability, owners can avoid the pitfalls that have sunk larger players. Thoughtful, measured implementation remains the safest path to realising AI’s promised efficiencies.