Fern Puts a Senior Sales Rep on Every Call Without Actually Hiring One

The Macro: Most Sales Reps Are Not Great and Training Does Not Fix It

Sales organizations have a consistent problem: the top 20% of reps close 80% of the deals. The rest struggle with objection handling, competitive positioning, discovery questions, and technical product knowledge. Companies spend millions on sales training programs that show minimal long-term improvement in close rates.

The existing sales tech stack is mostly post-call. Tools like Gong, Chorus, and Clari record calls, transcribe them, and provide analytics after the fact. Sales managers can review calls and provide coaching later. But “later” does not help a rep who just lost a deal because they could not answer a technical question or handle a competitor objection in the moment.

Real-time assistance changes the equation. If an AI can listen to a live call and provide relevant information, suggested responses, and coaching in the moment, it effectively gives every rep the knowledge of the best rep on the team. The rep still runs the conversation, but they have a knowledgeable assistant whispering in their ear.

Fern (formerly Ishiki Labs), backed by Y Combinator, does exactly this. It runs sales playbooks live during calls, providing real-time coaching, competitive intelligence, and product answers.

The Micro: Meta AI Researchers Building Sales Intelligence

Amit Yadav (CEO) was a research scientist on Meta’s LLaMA team and Reality Labs, with a PhD in AI and 20+ publications. Robert Xu (CTO) worked on multimodal AI and the Orion AR glasses at Meta, plus research infrastructure at Citadel Securities. These are not typical sales tech founders. They are AI researchers who recognized that real-time language understanding could transform sales execution.

The product listens to the live conversation and provides contextual assistance. When a prospect mentions a competitor, Fern surfaces competitive positioning. When they ask a technical question, Fern provides the answer with supporting details. When the conversation stalls, Fern suggests discovery questions that advance the deal.

This is technically challenging. Real-time speech recognition, context understanding, and relevant information retrieval all need to happen in seconds. Latency kills the product: if Fern’s suggestions arrive 30 seconds after the moment passes, they are useless.

The “senior AE on every call” framing is the right way to think about it. New reps with six months of experience can access the collective knowledge and playbook execution of someone with six years. This compresses ramp time and lifts average performance.

Competitors include Gong (post-call analytics), Outreach (sales execution platform), and Dialpad (which has some real-time AI features). Real-time coaching during live calls is still a relatively uncrowded space, partly because the technical requirements are demanding.

The Verdict

Fern is building at the intersection of real-time AI and sales execution. The research backgrounds of the founders suggest they have the technical depth to solve the latency and accuracy challenges.

At 30 days: are reps actually using Fern’s real-time suggestions during calls, or are they ignoring them?

At 60 days: what is the measurable impact on close rates for teams using Fern versus control groups?

At 90 days: is Fern reducing ramp time for new sales hires? If new reps reach productivity faster, that is an easy ROI calculation for sales leaders.

I think real-time sales coaching is an inevitable product category, and Fern’s AI research pedigree gives them a technical edge. The question is whether sales reps will actually adopt real-time assistance or find it distracting. The UX needs to be seamless enough that it helps without interrupting the natural flow of conversation.