Five ways to use AI in SaaS: a framework for founders

Executive overview

Most founders worry about falling behind on AI — but they're asking the wrong question. The real question is whether they're using AI the right way.

There are five distinct categories of AI use in startups, each with different risk and reward profiles. Knowing which category you're in changes how you evaluate tradeoffs.

The right question isn't "am I using enough AI?" — it's "which category am I in, and what are the risks?"

AI as your core business

• Remove the AI and there's no product left — foundational models, Jasper, Midjourney, Podsqueeze.
• Upside: potential to become the default solution in an emerging category.
• Risk: commoditization — today's moat can become tomorrow's API call.
• Risk: platform dependency — a single model change from OpenAI or Anthropic can end you.
• Risk: market education burden — customers don't yet know what's possible.
• Foundational model companies also face capital intensity.

AI as a feature

• AI enhances the product but isn't the product — remove it and the core value survives (e.g. Notion, Zoom, Loom).
• Can justify 20–50% price increases for premium AI tiers.
• Adds differentiation and can improve retention through habit-forming workflows.
• Risk: not a durable advantage — competitors can copy AI features within weeks.
• Risk: bad AI is often worse than no AI; quality vetting is essential.
• Risk: API costs can destroy unit economics if usage is high.

AI for building your product

• Using AI tools (Cursor, Claude Code, Copilot, Lovable) to ship faster — AI never touches customers.
• Reported velocity improvements of 3–10x; lets small teams compete with larger ones.
• AI is effective at test writing, a task most developers avoid.
• Risk: subtle bugs that are hard to spot without senior review.
• Risk: technical debt accumulates quickly if you move fast without understanding the code.
• Risk: security vulnerabilities — multiple tools have introduced exploitable patterns.
• Note: productivity gains are still debated; evidence is mixed.

AI for growing your business

• AI in the acquisition layer — outreach personalization, SEO content generation, ad copy testing.
• Examples: Clay for outbound, Jasper/Copy.ai for content, AI-powered ad optimization.
• Potential for 10–100x increase in outreach capacity and lower customer acquisition costs.
• Risk: customers may detect and reject AI-generated outreach.
• Risk: AI going off-brand can cause public damage.
• Risk: channel saturation — everyone is running the same tactics.
• Risk: compliance — AI outreach can violate regulations if unmonitored.

AI for operating your business

• Internal use only — your team is the user, not customers.
• Examples: tier-one support triage, resume screening, feedback analysis, internal knowledge bases.
• Potential for 30–80% cost reduction in specific departments; enables 24/7 availability.
• Risk: AI mishandling sensitive customer issues damages experience.
• Risk: employee morale if the team fears replacement.
• Risk: without human oversight, AI can violate laws or internal policies.