This page collects what used to sit at the top of Digital Impacts as preamble. It is the answer to “where does this come from and how far can I trust it”, moved out of the way of the research itself.
Where the material comes from
Digital Impacts collects the evolution directions described in the workshop Shaped by AI — Phase 2: the impact on digital professions. The two phases of the project are set out in the project plan; the sources behind the corpus are in the literature review.
The rule on sources
No figure enters without a verifiable source. Where a claim rests on a single vendor study, the page says so and says who benefits from the result. Where the estimates diverge, the range is given rather than the most quotable end of it — the drop in organic click-through, for instance, is reported as the twelve independent analyses actually report it, agreeing on direction and diverging by a factor of five on magnitude.
What this section does not claim
Three limits, stated because working around them would be worse than declaring them.
- Where a measure does not exist, the pages say so instead of simulating one. There is no rigorous, independent, brand-specific measure of how often AI gets a brand wrong; the solid studies measure news, and the ones on brands come from vendors selling the solution to the problem they measure.
- On brand identity and on the long horizon the literature offers no metrics. Every decision card inherits that limit.
- The professional profiles are a reading of the corpus, not a survey. A profession appearing rarely on the site means the corpus names it rarely — not that the change touches it lightly.
How it is structured
Three sections, split by what kind of thing a page is rather than by subject.
- What is happening — eighteen phenomena, on two levels. Three chapters follow routes people take deliberately, in order of how far adoption has got: AI-mediated search (which also covers how AI interprets who we are, because it happens in the same response), conversational chat, agentic systems. One chapter does not depend on adoption at all — the cross-cutting themes, which hold either way.
- What has to be decided — eight cards, grouped by what the choice puts at stake: where we are and through whom, how much we let the system do and who answers for it, what we keep counting. Each states the question, the default that applies if nobody answers, and the cost of both branches.
- What it changes for the work — the professions, each with the phenomena that touch it, the decisions that condition it, and the decisions it has to take itself. Who decides belongs here rather than with the decisions: who holds a mandate is a question about people and their remit, not about the choice.
Each phenomenon page runs the same way: the evidence, then the decision it forces, then what it changes for each profession named. The decision block is deliberately part of the argument rather than a callout — the point is that the choice follows from the evidence.
Corrections
Figures date, and several of these will. If something here is wrong or has been superseded, the comment thread at the foot of each page is the fastest route, or get in touch.