A website is not an annual report and not a job posting. No one explicitly writes that customer service is partly handled by a chatbot with human oversight, or that quotes go out within an hour because a system does the calculations. But the traces are there, in what is promised and how fast that is promised. The question is not whether a competitor says something about AI. The question is what the site shows about the speed and consistency of the work behind it, and whether that means a shift in the comparison.
Some sites are explicit: a chatbot that gives an immediate answer, a configurator that produces a quote within minutes, a tracking environment that shows real-time status without an employee's involvement. That is the layer that can be read without interpretation. Note the difference between "response within 1 business day" and "immediate answer": the first is a promise about human work with a queue, the second points to a task the system handles itself. Both are legitimate, but they don't compete on the same dimension.
The more interesting layer is not in the text but in the site's behavior. Response times that have dropped, quote processes that have been shortened, a support page that has gone from "send an email" to "chat now": these are traces of a task that has shifted from fully human work to a system that does most of it, with an employee approving or intervening. Compare today's site with an earlier version, via an archive or your own notes, and the shift becomes visible in what has become faster, more widely available, or more detailed without the organization visibly having grown.
The most common mistake is assuming that every gain in speed is AI, and that every mention of AI actually means the work is being taken over. A competitor can put "AI-driven" on the homepage while the core of the work is still entirely human work; marketing language and operational reality don't always align. Conversely, a site that nowhere uses the word AI may already have processes behind the scenes in which a system does most of the work and an employee only handles the exceptions. The site itself rarely shows which of the three categories applies: full takeover, takeover with human oversight, or unchanged human work. You have to add that distinction yourself by looking at what the site's behavior makes logical, not at what is literally stated.
A second pitfall is confusing speed with quality. A faster quote says something about the turnaround time of a task, not about the accuracy of its content. For tasks where oversight is needed for good reason, such as custom pricing or complex deliveries, a faster competitor may just as well be a competitor that has dropped the control step. That is a different choice, not automatically a better one.
The speed at which this changes depends on what is at stake if it goes wrong, on the systems already in place, and on how much customer contact a company routes through repeatable channels. A supplier with many standardized orders can largely let a system take over an order process; a supplier with a lot of custom work keeps the assessment with people, even if it already uses AI internally to calculate variants. This is not a matter of being ahead or behind. It is a matter of what the nature of the work allows, and that differs per company and per dimension, not per sector as a whole.
A website changes more often than an annual report and more slowly than a price list. For the dimensions on which delivery time, response speed, or accessibility of information decide the deal, checking competitors' sites a few times a year is defensible, especially around the moments when competitors revise their offering. For dimensions the website says little about, such as internal production quality, another source is needed. How often a full repeat of the comparison pays off depends on how fast your market moves; how often should you repeat a competitive analysis goes into this further.
The website mainly shows the customer-facing part of the work. For a more complete picture, it works together with other traces: what job postings reveal about the role of AI in a competitor's work, what price lists reveal about the speed and margin behind a competitor's offering, and what annual reports say about investments in automation. Each source shows a different part of the same question: where is the competitor gaining time, and does that change the dimension on which you are compared.
None of these traces tell you with certainty where you yourself win or lose; that is a separate question, worked out in how you know where your company truly wins. And none of these signals say anything about personnel decisions at the competitor: what an employer does with a team when a task shifts is up to that employer, and its own legal requirements apply there. What the website does offer is a first, free-to-read indication of where the comparison is shifting.
The underlying question, which work in your own company can truly be taken over by AI, is answered by the work scan from FTE TO AI on a per-task basis, independent of what competitors show on their site.
Name where you believe you win: speed, availability, precision, price. Set those claims next to what the sites of two or three competitors show, and mark which of your own claims you can support with concrete evidence and which rest on assumption. That is exactly what the free dimension check does: you name where you believe you win and see which of those claims can be defended with evidence. The full benchmark, with the comparison against your peer group on all market-deciding dimensions, is under construction.