A press release is the most controlled form of communication a company puts out. Every word is approved, every claim weighed against what competitors and customers can do with it. That makes a press release about AI adoption valuable and misleading at the same time. Valuable, because a company rarely announces something without an actual change occurring in the process behind it. Misleading, because the announcement always sounds more favorable than the practice on the floor.
The question is not whether the press release is accurate. The question is which part of the work behind it has shifted, and what that means for the comparison between you and that competitor.
Press releases about AI usually cluster around one of three categories, and each category calls for a different reading.
"AI now does X." This is the strongest claim: a task has been taken over, with no human left between input and output. Pay attention to whether the press release refers to a sub-process or to the entire customer journey. "AI processes incoming requests" is different from "AI answers all requests within a minute." The first sentence describes a step, the second a result. Only the second is a claim you can test.
"AI supports our people in Y." This is the middle category: the work remains human work, with AI as a tool that proposes and the employee who approves or rejects. This phrasing is often used to sound forward-looking without making a hard promise about speed or cost. It says something about direction, little about the effect at this moment.
"We continue to invest in personal contact." This sounds like a counter-move, but is often a signal that the company deliberately keeps one dimension outside the AI shift. That can be a strong positioning, or a lack of other options. Without further evidence, the two cannot be told apart.
The reason one company already works with an automated quoting flow and another does not rarely lies in ambition. It lies in the nature of the work. Tasks with a fixed decision pattern and a lot of repetition lend themselves more readily to takeover by AI than tasks that vary strongly per customer. A company with a narrow, standardized range can shift a dimension such as delivery time or response speed faster than a company with custom work in every order.
That means a press release from competitor A cannot simply be applied to your situation, and the absence of a press release at competitor B says nothing about being behind. Some companies actively communicate AI adoption because it strengthens their sales story; other companies apply it without mentioning it, because it is simply an internal process improvement.
From a press release you infer: which dimension the company itself considers important enough to claim, and into which of the three categories that claim falls. That is an indication of what the competitor wants its positioning to win on.
You do not infer from it: whether the claim also applies to the entire customer base, how the customer experiences the change, or whether the workforce has actually been adjusted. That last topic touches on employment-law decisions subject to their own statutory requirements; a press release is neither a source nor a basis for that.
Not every press release deserves equal attention. An announcement around a product launch, a merger, or a quarterly result rarely contains in-depth information about AI adoption; it mentions it at most as a side remark. The press releases that do matter are the targeted announcements: a new customer service tool, a partnership with a software party, a change in delivery times or lead times. These appear irregularly, and for most competitors a few times a year at most.
The value is not in the frequency with which you read them, but in recognizing the pattern across multiple releases: which dimension keeps recurring, and whether the tone shifts from "supports" to "takes over."
A press release rarely stands on its own. It only gains meaning alongside other signals: what customer reviews reveal about a competitor's AI adoption, what lost quotes reveal about where a competitor became faster or cheaper, and what your own salespeople already know but have not yet said out loud. Only in combination does a picture emerge of which dimensions have actually shifted at a competitor and which have only been announced.
To weigh those signals properly, you need to know who you are actually comparing against; see what a peer group is and how you assemble one. And if the pattern points to a dimension on which no one in your market yet scores, that is an indication for what a white space analysis exposes about unseen room in the market.
The underlying question remains the same, whether the evidence comes from a press release or from your own organization: which work in this company can really be taken over by AI. For your own company, the work scan from FTE TO AI answers that question per task.
Name the dimensions on which you think you are winning, and test them. The free dimension check shows which of those claims can be defended with evidence and which rest on assumption. The full benchmark, with the evidence matrix per dimension, is under construction.