The AI question the market asks of a company has hardened over the past year, shifting from what the technology might do for the business to whether it makes the business unnecessary. Data providers, software firms, publishers, education companies and professional services businesses have all faced some version of that doubt, and it usually shows up in the share price first. What makes this a communications story is that the doubt is often wrong. Many of the companies being written off hold precisely the assets that gain value as AI spreads: proprietary data, embedded workflows, regulatory trust, distribution, and customer relationships. The gap between how relevant a company is and how relevant it is believed to be has rarely been wider, or more expensive to leave open.

Companies that stay quiet and wait for the perception to correct itself will find that others define them first. Activist investors now target companies on AI from both sides: some are criticized for moving too slowly, others for failing to explain why their business can withstand the technology. Analysts ask about AI on earnings calls whether or not management raises it. And AI assistants have become an audience in their own right: when someone asks one whether a company still matters, the answer is built from whatever is on the public record.

Closing the gap is practical work. Across our engagements with companies facing this exact doubt, five imperatives come up again and again, and they apply whether the business sells software, education or financial data.

The first is to restate what the company is rather than defend what it was. Insisting that the old product still matters rarely works; the stronger move is to explain what the company uniquely holds that becomes more valuable as AI spreads. An education company will struggle to win an argument about content once content is abundant, but its advantage sits in what stays scarce: trusted assessment, credentials and outcomes that can be verified. Every sector has its equivalent, and identifying it is a strategy decision that has to happen before anyone drafts a message.

The second imperative is to lead from the front, and that means the CEO. When the chief executive talks about AI and the company's future in their own words, on earnings calls, in interviews and on stage, the market reads it as conviction. When the subject is handed to the CFO or the product team, the market reads that too, usually as a lack of conviction.

Third comes proof. The companies that shift perception are the ones that show evidence: what is deployed, what has changed, what value it has produced, and which claims describe pilots as opposed to systems running at scale. Investors are one audience for that evidence, and currently the most impatient one. Research from our Edelman Smithfield colleagues found that 65% of CEOs now reference AI in prepared remarks while only 27% of companies provide any AI-related financial evidence. The London Stock Exchange Group showed what closing that gap is worth: when it answered activist pressure this spring with a far clearer account of its AI strategy, Elliott Management publicly credited the improved communication, UBS removed the company from its list of businesses vulnerable to AI disruption, and the shares subsequently recovered their February losses, even though the products themselves barely changed over that period.

The fourth is validation from third parties. A company's own claims will only carry it so far. Perception shifts when customers put numbers on the value, analysts update their view and serious journalists test the claims and find they hold up. The communications job is to give those third parties something to work with: reference customers, real data and access to the people doing the work.

The fifth is the newest: make sure the story shows up where AI answers questions about you. Buyers, journalists and investors increasingly start their research with an AI assistant rather than a search engine, and what the assistant says about a company is heavily influenced by information available in the public record. If that record still describes the company as it was three years ago, that outdated version can continue to shape what AI answers say about the company. Keeping the narrative current, consistent and machine-readable is becoming as fundamental as search optimization was two decades ago.

None of this can rescue a business that AI genuinely makes obsolete, and the approach only works when the evidence behind it is real. What communications can do is make sure the market's view of a company catches up with that reality, rather than leaving the story for others to write. Earlier this year I argued in these pages that communicators need more precise language for describing AI systems. For companies whose relevance is in doubt, that precision now has to cover the whole corporate story, and the communicators who can deliver it will find themselves at the center of the most consequential conversations their organizations are having.

Margot Edelman is Edelman's global tech practice lead.

Note: This article reflects the views of the author and is published as part of PRovoke Media’s opinion section. It does not necessarily represent the views of PRovoke Media or its editorial team. We welcome a range of perspectives and invite readers to submit thoughtful responses or counterpoints for consideration to [email protected].