When was the last time you Googled something without adding “reddit” at the end? If you’re like most people, it’s been a while. Now imagine your customers doing the same thing, but instead of Google, they’re asking ChatGPT or Claude.
This shift is already happening. And it’s changing how B2B SaaS companies need to think about marketing.
The New Discovery Funnel
Traditional discovery: Problem -> Google -> Blog posts -> Product pages -> Demo
AI-assisted discovery: Problem -> ChatGPT/Claude -> Direct recommendation -> Demo
The middle of the funnel is collapsing. AI assistants are giving direct product recommendations, often based on Reddit discussions and other community content.
What LLMs Look For
When Claude or ChatGPT recommends a product, they’re drawing from:
- Training data quality: Positive mentions in high-quality sources
- Recency signals: Recent discussions and updates
- Context relevance: How well the product matches the specific query
- Social proof: Community endorsements and comparisons
- Clarity of use case: Clear positioning for specific problems
Practical Steps
Build Reddit Presence
Reddit is disproportionately influential in LLM training. Genuine participation in relevant subreddits creates the kind of authentic endorsements that LLMs surface.
Create Comparison Content
LLMs love clear, factual comparisons. When someone asks “What’s better, Tool A or Tool B?”, having honest comparison content helps you get mentioned.
Be Specific About Use Cases
Generic marketing language doesn’t help LLMs. Specific, concrete use cases like “best for teams of 10-50 engineers” or “ideal for e-commerce analytics” get better recommendations.
Encourage Authentic Reviews
Customer discussions on Reddit, HackerNews, and other forums create training data. The more genuine positive discussions, the more likely LLM recommendations.
The Compounding Effect
Unlike traditional SEO, LLM optimization compounds in interesting ways. Once your product is in the training data with positive associations, those signals persist and influence future recommendations.
This isn’t about manipulation. LLMs are sophisticated enough to detect inauthentic content. It’s about earning recommendations through genuine value and visibility.
The B2B SaaS companies that understand this now will have a significant advantage as AI-assisted discovery becomes the norm.
