What is wrong with AI services businesses
The article discusses concerns about the long-term durability of AI services businesses, questioning what fundamental flaws or weaknesses make them feel unsustainable despite current market demand.
Background
- The Hacker News thread "What is wrong with AI services businesses" discusses a growing concern in the tech industry: companies that build custom AI solutions (chatbots, automation, integrations) for clients often struggle with low margins, high competition, and weak defensibility.
- "AI services businesses" refers to agencies or consultancies that sell bespoke AI work rather than a product. Unlike SaaS (software as a service), which scales with low marginal cost, services require ongoing human labor for each client.
- Key issues raised include: clients can easily take the solution in-house after it's built; the underlying AI models (e.g., GPT) are commoditized and available to everyone; and there are few barriers to entry, leading to a race to the bottom on price.
- This mirrors earlier cycles in tech — e.g., the "web development agency" boom and bust, or the "app development" gold rush — where early movers found it hard to build lasting businesses once the technology became accessible to non-specialists.