Real people versus AI avatars and how buyers decide who to trust

A few years ago, a researcher at the Stanford Internet Observatory got a sales message on LinkedIn from a friendly account rep. The rep did not exist. Her face was computer-generated, and pulling that thread led researchers to more than 1,000 profiles using AI-generated faces, many of them doing lead generation for real companies selling software.

That was 2022, with clunky tools. The tools have improved. Buyers have noticed.

The scale of the fake problem

LinkedIn takes this seriously, and its own numbers show why. According to its community report, 99.7% of the fake accounts it stopped in the second half of 2025 were caught proactively, before any member reported them. In the first half of 2022 alone, it blocked 16.4 million fake accounts at registration. That is a good catch rate. It also tells you how much there is to catch.

The platform's answer has been verification. In December 2025, LinkedIn said more than 100 million members had added a verification to their profile, and that verified members get up to 60% more profile views. The market is starting to price in proof that you are a person.

When the fake joins the video call

Fake profiles are the cheap end. The expensive end is synthetic people in real meetings.

In early 2024, an employee at the engineering firm Arup joined a video call with what appeared to be senior colleagues and was persuaded to transfer about $25 million. The other faces on the call were deepfakes. Arup's chief information officer later described it as technology-enhanced social engineering, since none of the firm's systems were breached.

The FBI has issued similar warnings. A May 2025 public service announcement described campaigns using AI-generated voice messages to impersonate senior US officials, build rapport and then steal credentials.

Your buyers read the same headlines. Their security teams brief them on it. Their default posture toward an unknown sender has moved from "maybe" to "prove it."

How buyers sort real from synthetic

Senior buyers have always been skeptical. What has changed is the cost of being wrong and the number of plausible fakes in the inbox. Gartner found 73% of B2B buyers actively avoid suppliers who send irrelevant outreach, and AI has made irrelevant outreach nearly free to produce.

In practice, buyers run a quick, informal check before engaging:

• Is there a history? A profile with years of posts, real colleagues, conference talks and a work record that hangs together. A three-month-old account with a flawless headshot fails here.
• Does the message know something? Specific, checkable insight about their business. Generic praise for their "impressive growth" reads as a template, because it is one.
• Can someone vouch? Mutual connections, a referral, a name they have seen on a panel or in a trade publication.
• Will a person get on the phone? This is now the deciding test. A real expert will take a 20-minute call, answer hard questions live and manage without a script.

Think of it like currency. When counterfeits flood the market, shopkeepers don't stop taking cash. They hold every bill up to the light. Your outreach is being held up to the light.

What this means for how you sell

AI is a fine tool for research, preparation and drafting. Used to fake a person, or to fake familiarity at scale, it spends trust you will need later in the deal, when legal and procurement start asking exactly who they are dealing with.

The firms that stand out now do a few unfashionable things well:

• Put a named, verifiable expert in front. A real partner, producer or practice lead with a public track record and a verified profile.
• Show the work. Points of view a buyer can test, published under a real name.
• Move to voice early. Offer a short call with someone who can answer the hard question on the spot.
• Own every word. If a tool helped draft a message, the human who sends it should be able to defend each sentence on a call.

In a market full of synthetic confidence, the scarce thing is a credible person with a phone number.

A check for this week

Look at your own outreach the way a skeptical CFO would. Open the profile of whoever sends your firm's first touches. Is it verified? Does it have a real history? Read the last three messages sent. Could a stranger tell each one was written for one specific person? Then ask the hardest question: if a prospect replied "call me in ten minutes," who on your team would pick up?