Personalization was a signal. AI turned it into noise.

In 2005, the number of internet users worldwide crossed one billion. With one-sixth of humanity online, businesses began moving their selling there too.

Before the advent of the internet, sales outreach primarily consisted of telemarketing, spray-and-pray mailing, and door-to-door soliciting. But now, reaching out to thousands of people had suddenly become easier through emails and web-based communication. However, it was also easier for customers to ignore that outreach, as the messages were often irrelevant and generic. Businesses scrambled to find ways to make prospective customers respond and eventually found the key ingredient: personalization.

Sitting in front of a computer, armed with digital tools like CRM software and online database directories, salespeople were able to find out details about a prospective customer, like their role, what industry they're in, and whether they're using a competitor's product. This gave them an idea about the pain points each potential buyer faced at work, helping them to tailor their sales pitch accordingly. And it worked—customers felt like salespeople were helping them with solutions to their problems instead of just trying to close a deal.

It worked so well, in fact, that personalization has become the standard in sales. According to McKinsey, 71% of consumers expect companies to deliver personalized interactions, and 76% get frustrated when faced with a lack of personalization.

The disappearance of visible effort

As effective as personalization is, it's also the most difficult to scale, given the amount of research it entails. That's why personalized communication carries more weight among customers swamped by generic calls and emails; it shows the amount of effort put in by the salesperson to know them and helps build trust.

Enter AI. With just one prompt, it can browse the internet, get the relevant facts about the customer, and prepare a polished output in the desired format. Suddenly, personalization no longer carries the weight it used to. Why is that?

In 1973, the economist Michael Spence introduced the concept of signaling: conveying otherwise unobservable information through an observable action. Importantly, the receiver regards a signal as credible only when it's hard for others to replicate.

Spence's example concerned job market signaling. For an employer interviewing candidates, who can't directly observe how capable a candidate is, a degree works as a signal not only because of what it teaches but because completing it is harder for a less capable person than a more capable one.

Applying this to sales outreach, a personalized message sent to a prospective customer used to signal effort, because it showed that the salesperson was willing to spend the time to do the research. A buyer receiving this message could then reason that someone who invested this much time probably has something worth hearing.

But AI collapses this arrangement entirely. All it takes is a few minutes with a chatbot to make all the sales touchpoints appear personalized. Sales teams are overjoyed that what used to take hours now merely takes minutes. But what about customers?

AI's trust deficit

A 2025 study on how the use of generative AI in advertisements can impact consumer trust reveals two possible reactions:

• If companies disclose that they used generative AI in a given advertisement, the public might appreciate the transparency, which can result in a small increase in trust. A section of consumers might also deem AI use in advertisements as inappropriate, which could erode trust.

• However, if the companies don't disclose the use of AI in the advertisement and end up getting exposed, either through a third party or when consumers recognize it themselves, the negative impact on trust becomes stronger than through voluntary disclosure.

While the study doesn't talk directly about sales, it's a good indication of how people interact with sales-focused messaging from an organization. Given the rapid pace at which AI technology is changing, the above study is by no means conclusive or even comprehensive (the sample size was 321 people and it was an online experiment). But even if a business were to take the above conclusions at face value, it would imply that prospective customers won't appreciate AI-generated sales messaging if it's not disclosed. And the risk of damaging trust persists even if companies disclose their AI use.

Elsewhere, the study defines trust as the "confidence that advertising is a reliable source of product/service information and the willingness to act on the basis of information conveyed by advertising.” But public trust in AI isn't all that robust: a global survey covering 47 countries reveals that less than half of respondents are willing to trust AI. With increasing AI adoption, people have become less trusting and more worried about the technology.

Showing AI its place

With AI making personalization—the primary signal of effort—appear to be table stakes, businesses run the risk of letting AI-driven communication foil efforts to build trust among potential buyers. So how can businesses make their outreach signal genuine effort again?

The answer lies in how sales and marketing teams use AI. If they see AI as a way to scale output with minimal effort, businesses will merely end up with another version of spray-and-pray, which defeats the idea of standing out.

The real line is between effort the buyer sees and effort the buyer doesn't. Businesses must keep AI away from things that touch buyers: emails, calls, marketing content, and sales demos. These are points in the buying journey that help in building trust with the customer. For instance, a founder talking directly to a prospective customer can build trust more strongly than dozens of emails or calls made by a salesperson.

So where does AI go? Everywhere that the buyer can't see: customer research, brainstorming, list building, call summarization, CRM logging, and data analysis. There's much to gain from automating these areas through AI, as it can buy teams the time and energy needed to go deeper on select accounts and build trust.

We can look at the evolution of sales as bookended by door-to-door selling and AI-powered outreach. But the real difference between the two isn't the amount of effort; it's whether the buyer can see it. In door-to-door, the seller physically showed up at hundreds of doors, and the buyer could see it. AI is the reverse—it merely creates the appearance of effort and leaves buyers suspicious. The solution, therefore, is to put effort back where the buyer can see it: let AI handle everything the buyer never sees, and keep a human on everything the buyer does. Because, for all its capabilities, AI can never build trust with a human.

To wrap up, here are the words of David Ogilvy—who was a door-to-door salesman before venturing into the world of advertising—from his 1935 sales manual, which remains uncannily relevant even in this AI-powered sales age:

"The more prospects you talk to, the more sales you expose yourself to, the more orders you will get. But never mistake the quantity of calls for quality of salesmanship. Quality of salesmanship involves energy, time, and knowledge of the product." 

Also read: AI can't have tacit knowledge—that's why it can't replace you