The soft technology profile is largely gone. Tech and AI reporters now start from skepticism: they assume the claims are inflated, they want to know how the system actually works rather than what the marketing says, and they treat safety, bias, labor practices, and energy use as core parts of the story rather than footnotes. A spokesperson prepared for the friendly coverage of a few years ago will be caught out.
Have a real answer to “how does it work”
This is now a standard question and it is asked to test whether you understand your own product. “It’s proprietary,” “it uses AI,” and “the model learns from data” are not answers, and a reporter who has done their homework, or has a technical colleague, will note that you did not give one.
Prepare a genuine explanation at two levels. For a general reader: what the system takes in, what it produces, and roughly how, in plain terms. For a technical reader: the approach, the meaningful architecture choices, the data, and the known limitations. You do not have to reveal trade secrets to be specific and honest about the shape of the thing.
Name the risks before the reporter does
Every AI or platform technology has real risks: bias in outputs, failure modes, misuse potential, effects on the people whose work it changes. A spokesperson who claims otherwise is not believed, and the interview turns into a series of “but what about…” questions with you on the back foot.
List the genuine risks and limitations, and for each, the specific thing you do about it: the evaluation you run, the guardrail you built, the human review in the loop, the disclosure you make to users. Rehearse saying these plainly. Acknowledging a real limitation is what gives your positive claims credibility.
Close the hype gap
Put your public claims next to what the product does today. If the website says “automates your entire workflow” and the reality is “automates three of eight steps, with the rest on the roadmap,” a reporter who tests the product will write about the gap. Rewrite every future-tense capability into its present-tense reality for the interview: “today it handles X; Y is something we’re building.”
Prep the founder’s specific tells
Founders are the most credible voice and carry a specific set of risks in tech interviews:
- Dismissing competitors (“they don’t really have a model”), which reads as insecure and invites a fact-check.
- Overpromising timelines (“we’ll have that by summer”), which becomes a broken promise on the record.
- Sweeping impact claims (“this changes everything for…”), which age badly.
If a founder is doing the interview, rehearse against these directly and flag each one the moment it appears in the mock.
Set the business-question lines
Layoffs, funding, valuation, and burn come up in almost every tech interview now. Decide the exact answer for each before the interview and use it consistently:
- Layoffs: a straight account of what happened, why, and how people were treated. No euphemisms.
- Funding and valuation: either a firm “we don’t discuss those” every time, or one agreed number, never a different framing per reporter.
Frequently asked questions
Why has tech coverage gotten harder?
The default posture of tech and AI journalism has shifted from curiosity to skepticism. Reporters now assume claims are inflated, ask how systems actually work rather than what they promise, and cover safety, bias, labor, and environmental impact as central rather than peripheral.
How do I answer "how does your AI actually work"?
Have a real, honest, non-hand-wavy explanation at two levels: one for a general reader and one for a technical one. Reporters increasingly bring or consult technical expertise, and a vague answer ('it's proprietary', 'it learns') now reads as either evasive or as not understanding your own product.
What do I say about the risks of our technology?
Name them honestly and specifically, then explain what you do about each. A spokesperson who claims their AI has no meaningful risks is not believed, and the interview becomes about the denial. Acknowledging a real limitation is what makes your other claims credible.
Should a founder do these interviews or a comms person?
Founders bring credibility and detail, but founder overconfidence is a specific risk in tech: dismissing competitors, overpromising timelines, making sweeping claims about impact. If the founder does it, prep them hard on those tendencies.
How do I handle questions about layoffs, funding, or valuation?
Decide the exact line in advance. For layoffs, a straight account of what happened and how people were treated. For funding and valuation, either a consistent 'we don't discuss it' or a single agreed figure, used every time.
Tech or AI story on the calendar? Press Prep builds the skeptical questions from the reporter and the current coverage, and runs a mock that pushes on the how-it-works and risk answers. See media training for founders or message development.