As buyers use AI to research vendors and check claims, salespeople may lose influence over the decision. Their value may shift toward trials, coordination, and closing deals.
It’s no secret I greatly dislike salespeople. Specifically, I dislike the fact that most salespeople take their wealth of experience and capabilities and spend their time spamming people on the internet, including me. If the best use of their time is cold calling strangers, it really makes you wonder how little value they have to offer.
I want to be clear, I don’t object to sales itself. Explaining the product, showing how it meets the customer’s needs, walking through objections, setting up demos and trials, negotiating contracts, and staying on top of the entire process is all meaningful work. Spamming strangers isn’t. And yet far too many sales jobs involve spending significant time doing just that. AI is going to disrupt sales, and not just in the way people expect.
Sales and marketing lead generation have been early adopters of technology generally, as well as AI. This is not a recent trend.
Physical bulk mailing (sometimes called junk mail) dates back nearly 200 years in the US. One of the first uses was in 1835 by the American Anti-Slavery Society according to the Smithsonian Museum. In the latter part of the twentieth century, telemarketers' calls and fax spam grew so bad the US needed to pass the Telephone Consumer Protection Act (TCPA) in 1991.
Email marketing (spam) remains common to this day. It has moved from general spam to “personalized” (e.g., “Dear {name}” where “{name}” is not resolved or resolved incorrectly a surprising amount of time), to targeted campaigns in which recipients would be segmented based on prior actions. As many of us have experienced, AI now “personalizes” each message, albeit with irrelevant information. Sadly, spamming will likely only increase.
The best description of sales I ever heard came from my VP of Sales at Averon. He was a former army ranger and planned sales meetings the same way he did military actions. His planning and coordination was a thing of beauty. But he acknowledged that his job, regardless of the process he designed to take them through the buyer’s journey, was to hold their hand through what was ultimately an emotional decision.
Buying decisions are based on trust. I trust a high-end steakhouse to have a good quality meal with high-end service; I also trust a fast food restaurant to taste the same and consistently meet my expectations at any location. I trust a car I buy to be safe or gas efficient. If I buy product A over product B, I’m trusting it’s more reliable, cost-efficient, better meets my needs, and/or will help me do well in my job. The product could not work as expected and the steakhouse meal could be bad, but I’m playing the odds over the alternatives based on trust. Brand marketing builds trust in certain experiences, quality, or emotions (think of well known brands like Coke, Mercedes, and Disney) and sales helps close the gap of what marketing can’t do. That may be anything from educating on features, showing how it compares to the alternatives, or overcoming concerns.
There’s an open question about how AI changes trust. In the long term, I believe we must shift to a zero trust model (see “The Shift to a Zero Trust Society”), but it will be decades before that comes to pass. In the meantime, and somewhat orthogonal, people trust AI.
Let’s be explicit about this. There is widespread distrust of AI companies and of AI as a concept. However, people trust the advice AI gives them. A study published by MIT’s Media Lab found that people trusted AI-generated medical advice even when it contained errors (paper | article). An F-Secure survey found that while “90% of AI users say it's their responsibility to verify the answers AI gives them, . . . 70% still only check sometimes, rarely, or never do.” (Disclaimer: F-Secure is a commercial company selling security products.) In short, actions show that people trust what AI tells them, more than they should.
Eighteen months ago I wrote about Generative Engine Optimization (GEO) or Answer Engine Optimization (AEO), although I called it AIMO, which clearly never caught on (see “Welcome to the World of AIMO: AI Model Optimization”). This is going to transform the sales process.
In a traditional buyer's journey, customers begin with a need. From there they will find potential vendors, research them, and create a consideration set, meaning a short list of likely candidates. For example, you may need a car, and you can probably think of dozens of models, but if you’re a straight guy through a midlife crisis and have no kids, minivans and Subarus won’t really be a potential buy but a Chevrolet Corvette might be. (A buyer who’s happily married with young kids would flip that.) After that comes the trial. The last stage (for initial purchase) is the decision. This is where the salesperson overcomes objections, works on pricing, or talks through specific issues.
Now let’s consider this process in an AI world. The buyer begins by asking AI what it recommends. The stages of vendor search, research, and consideration set start to collapse into one or more chats with an LLM. Even if it’s multiple chats, it’s still going to happen over a compressed timeframe. For a large enterprise sale, traditionally each of those steps could be days to weeks as someone finds time to read through websites and marketing materials, or online reviews. Now this can be done in hours or minutes. At the end the potential customer has one or more vendors in the consideration set. Even if the LLM didn’t narrow it down to one, it likely gave directional guidance, e.g. this one is the most flexible, that one makes sense for tight budgets. The salesperson now faces a better-informed (possibly biased) purchaser.
As an analogy, consider car sales. In the mid twentieth century, a consumer walked into the dealership and relied on the salesperson for guidance. This is because, as Dan Pink notes in his book To Sell Is Human, the salespeople benefited from information asymmetry, knowing much more about the cars than their customers did. Fast forward to today and a car-buyer walks into a dealership with the same information. She’ll know about the car’s features, reviews by third parties, online recommendations, recall and other history for the brand, etc.
This is good and bad news. The good news is, salespeople will face buyers further along in the journey. The bad news is, the salesperson will have less room in which to maneuver the prospect.
I noted at the start that purchases are about trust. A good salesperson creates trust in the product or service, necessary to close the sale. The article referenced above summarized the MIT report’s findings: “When the AI model was accurate, it scored significantly higher than human doctors across every measure of trust and satisfaction. . . Even when the AI made clear factual mistakes, participants rated the advice as valid and trustworthy, saying they would follow it at rates comparable to or higher than physician recommendations.”
In other words, patients trusted the AI over doctors. Doctors are highly educated people with years of experience who have both a moral and professional obligation to look out for the interests of their patients. Salespeople can help you decide, but it’s well understood that they get paid commission. Consequently, their interests may not align with yours.
Additionally, salespeople have less training than doctors. That’s not a knock on them, only to say with doctors there’s a clear, large education/training gap between them and the patients. That’s not true for salespeople and customers. If patients think that, assisted by AI, they know better than doctors, the same logic applies even more so. Doctors understand your body better than you do. Salespeople do not understand your job as well as you.
If people trust AI over doctors, how much more will they trust AI over salespeople? Whatever you say, they will likely ask their AI to confirm it. This is especially easy when so many meetings are done by video and transcripts are shared afterwards.
This raises the question: what is it salespeople provide when AI is the trusted authority? It may be as simple as having a human on the other side, so people know there’s someone to call. For example, many people cared about having a US-based customer support team for their purchases, having experienced exasperating IVR systems and overseas call centers who could do no more than follow the manual, but never escalate beyond the support team. However, the current trend is otherwise. As a Gen Xer, I grew up with the rule: don’t call after 10pm. But for millennials the rule is: text don’t call. They want to bypass the smalltalk of a phone call and reduce the “intrusiveness” with direct voice-to-voice communication. Gen Z seems to be continuing the trend. Given that, along with the high association of trust in AI, it would seem they may place less value on that human connection. I do think the pendulum will swing back, but that will be a generational shift, not something in the coming years.
There are still two key elements a salesperson provides, or really two facets of the same value proposition. Salespeople will foster the trial and then they will help close the deal. The trial management role is obvious, since someone has to coordinate not only who gets access to the product or service, but that someone oversees the process to make sure it’s being used correctly and value is being exposed. The seller’s AI likely won’t be quite as helpful in the near future since it may lack much of the customer context and the buyer’s AI likely can’t marshal resources to deliver on the trial as well as a human can.
The second aspect is shepherding the deal. AI itself may even do the final contract negotiations. However, there’s often a lot of legwork in enterprise sales. This includes answering details, addressing stakeholder concerns, etc. Neither the buyer’s AI nor the seller’s AI is likely to have full context. But it means communication, project management, and negotiation all become more important, compared to today’s salesperson who fundamentally must be about building rapport. Note that I included negotiation even though I wrote earlier that it may go to an AI. Negotiation is not just about haggling over price, but often about coalition building, and that matters for business sales where there are multiple stakeholders.
In “The Rise of Sales” I wrote about how the operational improvements of AI make sales the crucial gating factor of new ventures. With other elements de-risked, sales becomes the ultimate early (and continued) signal of success, being expected much sooner in startups and corporate initiatives. However, as AI progresses, particularly as GEO / AEO advance and people will turn more and more to their AI for decision making, the impact any salesperson can make may decline. Outreach may become less important, and a salesperson’s influence during the decision-making process will decrease. That’s not to say it will disappear entirely, but rather that the nature of sales will evolve and parts of the process will shift to AI, as it has with other functional areas. I just hope it cuts down on sales spam.
It’s critical to learn about corporate culture before you accept a job offer but it can be awkward to raise such questions. Learn what to ask and how to ask it to avoid landing yourself in a bad situation.
Investing just a few hours per year will help you focus and advance in your career.
Groups with a high barrier to entry and high trust are often the most valuable groups to join.
