Sales teams have not found generative AI to be a major differentiator when it comes to making work easier, according to a study from agentic sales messaging platform company Breakthrough. The primary reason: AI fails to capture the human element.
In an interview with TechRepublic, Breakthrough CEO and co-founder Adit Abhyankar stated, “On the one hand, you would think that it’s the best application for AI because sales is about communication, language models are also about communication, but the difference is that when I sell, when I’m trying to communicate something to you, I have an opinion.”
Breakthrough surveyed 500 US and UK-based sales directors, managers, and executives, all of whom were employed by companies with more than 50 employees, between February and March 2025.
Sales personnel struggle with applying AI to specific tasks
Sales as an industry is still in the process of choosing which tasks really require a human touch. Those tasks benefit from trust between individuals, Abhyankar said. Meanwhile, many people are using AI — the ubiquity of its use in the survey surprised Abhyankar — but the struggle comes in finding the best tasks to use AI for.
“What AI does is it gives you things that seem plausibly correct super easily,” said Abhyankar. “And so what ends up happening is that it’s really easy to get a wow effect because the moment you ask it for something, you get something back, and when you read it, at first glance, it looks like it’s really good. But when you look at it in a little bit more detail, you realize, actually it’s not really good because it doesn’t really reflect what I was trying to say.”
SEE: How to Use AI in Business from TechRepublic Premium
Working with AI becomes “cumbersome” when sales personnel have to spend a long time telling the generative AI what it did wrong, Abhyankar said.
Nearly half (47%) of sales professionals surveyed said they spend between 31 and 60 minutes daily on generic AI tools — the same amount of time they spend on CRM tasks.
Primary ways sales staff use generative AI
“The primary use case for AI right now for salespeople is around content generation,” said Adit Abhyankar, chief executive officer and co-founder of Breakthrough, in an interview with TechRepublic.
Most teams put AI to work on content creation: 55% of teams used it for sales material generation and 42% for general content generation. Of individuals using AI, 47% used it for prospecting research, 40% for meeting preparation, and 39% for presentation generation.
Which tasks should be designated for humans, and which tasks should be designated for AI?
“I think what’s more likely to happen is that we will get better at saying, this is an AI task, and this is a human task.” For example, he said, a salesperson can use AI to create a slide, freeing up more time for them to meet people at an event.
AI may be able to scan through a conversation and interpret what part of the conversation holds the most actionable insight, but it can’t find leads for the business.
Pitching something to a potential client with no period relationship to your business is a job for a human, he said. Genuine interactions and genuine warmth help build trust.
Other pain points reported by sales personnel in the study included inaccurate results and hallucinations, AI tools not understanding their specific needs, and too many iterations being needed to produce useful results.
Privacy is not as much a concern as ‘losing control’
According to Breakthrough, 23% of sales professionals expressed concerns about privacy and data confidentiality when using AI. Privacy is a relatively minor concern because sales professionals are likely to use an enterprise version of an AI product. A bigger concern, Abhyankar said, is “losing control” – the probabilistic system spitting out unexpected, off-topic content.
“I would say the privacy risk is a well known problem that has been solved before, whereas this risk that AI generates things that you don’t control and that might actually hurt your brand, that is not something that we know how to control yet, and that’s not a a well known problem,” he said.
Proving ROI of AI is still difficult
It is difficult to prove the ROI of AI for several reasons. In sales, AI may not be able to provide strategic guidance that can be proven to make money. Another reason: It can be difficult to judge productivity savings by comparing how much time has been saved versus how much extra work has been done.
“You’re doing more with less time, so you as a person are more productive, but the company as a whole is not because it has not really figured out what to do with those extra three hours that you just saved,” Abhyankar said.
This uncertainty is a natural progression; it’s “how technology normally works,” he said. “You end up with capability and only once you understand the capability are you really able to translate into business value.”
