.png)
This article is based on a conversation between Vivun CEO, Matthew Darrow and Product Leader, Russell Witham, on The Unexpected Lever podcast.
GenAI advancements are fundamentally changing how go-to-market teams will operate in 2025.
CROs are actively looking to AI to give their sellers new competitive advantages (2024 Gartner Generative AI Planning Survey), whether by scaling customized buyer interactions, automating administrative tasks, augmenting revenue and business intelligence, or other popular applications (Gartner).
Market leaders, like Salesforce, are betting big on agentic AI, primarily focusing on the bookends of the customer journey: SDR agents to help generate leads and Support agents to help resolve customer tickets.
And these are critical applications, don’t get us wrong. They’re just not the only applications.
Most of revenue generation actually happens between the bookends – between when someone is interested, has shown up on your doorstep, but has not yet taken the next step to becoming a customer or expanding further.
Technical validation represents the majority of the sales cycle but the minority of RevTech strategy.
This is the realm of the mid-funnel, where AEs and SEs work together to drive revenue.
This is the work to be done that wasn’t covered at Agentforce. And it all has to do with the work Sales Engineers focus on every single day.
So what would it look like for agentic AI to move the needle on the technical selling work that represents the majority of buyer time and seller efforts in the sales cycle?
In general, a lot of companies are starting to experiment with the foundational LLMs, and we’re seeing small vendors emerge with chat bots to answer questions, surface pre-loaded battle cards, summarize calls, and fill out RFP documents.
While helpful, this doesn’t get at the root of the core strategic work of what experts ellers do.
This approach really misses the mark on the essence of what makes the work of your best performing reps so valuable, what makes companies spend hundreds of thousands of dollars per year on individual AEs, and why buyers don’t trust anyone else on your selling team.
Where do sales reps actually spend their time? According to Salesforce’s 2024 State of Sales report, only 28% of a seller’s week is spent actively selling. The rest is consumed by tasks like managing deals, logging data, coordinating internally, and preparing materials—activities that slow teams down and don’t directly drive revenue.
Focusing AI on low-impact tasks like inbox triage or meeting notes may sound productive, but it doesn’t solve the real problem. If your reps are spending most of their time off the field, you won’t hit your number by automating the sidelines.
The greatest opportunity for AI isn’t task automation—it’s sales execution. Specifically, enabling reps to craft and communicate solutions faster, answer buyer questions confidently, and move deals through the funnel with precision. That’s where AI Sales Agents deliver impact—not by saving minutes, but by winning back the hours that close revenue.
You can’t just ask ChatGPT or Claude to create a value case or stakeholder map and expect a quality output, because foundational models are not useful for replicating the complexity of hybrid human roles.
Why? There are 2 major problems:
These LLMs operate via a “bottom-up” approach, processing vast amounts of data and generating responses based on statistical patterns. This works well for general queries but lacks the structured, top-down knowledge framework that sales engineering demands. Without a holistic understanding of AE objectives, AI’s responses often seem generic and lack the strategic depth that an experienced rep provides. It’s like reading all the words in a book without understanding the plot or the characters’ motivations—AI might know facts but misses the underlying story.
These LLMs lack the context of how to structure and integrate the information they intake. ChatGPT can pull facts without any understanding of how to link them meaningfully. Even the latest chain-of-thought models lack the appropriate context for a field as complex and nuanced as B2B selling. This often requires a real person with actual experience to understand how to relate variables like customer needs, product capabilities, technical criteria, and the competitive landscape into a cohesive solution.
Foundational models will get better and better, but the best way to get leverage across the entire team (and not just rely on one person being enterprising), you need a system that does this en mass.
We are in the middle of an incredibly exciting transformation moment for not only B2B selling and buying.
We believe it’s feasible to create an AI agent that will actually work as a strategic partner to SEs, AEs, CSMs, etc., and not just a knowledge retrieval engine. (In fact, we already created one).
But you’re not going to get leverage from just an LLM with bottom-up knowledge. You’re only going to get a great outcome from a system that incorporates deep domain expertise of the best sellers and product experts in B2B.
The current AI approaches “for” Sales reps miss the mark because they have a reductive understanding of selling work. Until they appreciate the complexity and true value of the function, they can’t be expected to “do the work” in a meaningful way that maximizes value to your customers and the internal teams with whom they work.
For a deeper dive, watch the full conversation below. For more discussions like this about the future of GTM, follow The Unexpected Lever to hear not only from Vivun experts but also leading voices across B2B.
.avif)
.png)
.png)
.png)
