Consistency Is the Signal That Decides Whether AI Trusts Your Brand's Story

Ask a language model about a company today and it will answer instantly, distilling every signal ever emitted about the business into a single verdict for someone who has never visited the website or read an earnings call. That verdict now forms the first impression most customers receive. The task facing communications teams is no longer producing more content, but controlling the quality of the signals feeding the models. The ones still measuring success by publishing volume are optimizing the wrong variable entirely.

Michela Ratti is a veteran chief communications and brand officer with two decades of experience building communications, brand, and corporate affairs functions for listed multinationals across consumer technology, FMCG, luxury, and regulated industries. She's an engineer by training, beginning her career at Procter & Gamble as a product scientist before moving into communications and rising to VP of marketing communications. From there, she went on to lead communications at eBay across EMEA and at Swarovski through a multibillion-dollar rebrand, and most recently served as Chief Communications & Brand Officer at Ferragamo. Ratti's central assertion stems from a hard truth about where that first impression of a brand now gets formed.

"The AI answer on your company is the new front door, but it's a front door that the company itself doesn't own," she says. In her view, that loss of ownership is what forces a rethink of the whole discipline, starting with the standards for any signal a company emits.

A new set of Cs, borrowed from the diamond trade

Years in luxury left Ratti with a useful frame. Where a diamond is graded on the four Cs of clarity, color, cut, and carat, she's built her own set for the age of AI, a checklist for any information a company releases into the systems that will later answer questions about it.

"I've got my five Cs for AI, which are about being correct, current, crawlable, corroborated, and continuously monitored," Ratti says. These are the properties that determine whether a model repeats a company's information accurately or garbles it. She treats testing them as routine, regularly querying the major models to see what they say about companies she has worked for, and occasionally about herself. What comes back, she notes, is often surprising, which is precisely the point: the answer is being assembled whether or not anyone in the building is paying attention.

Reputation when there's no one in the room

The deeper change beneath the checklist is a shift in what reputation even means. The old definition assumed an audience of people talking among themselves. The new one assumes an audience of models talking to users, with no human necessarily present at the moment of judgment.

"Reputation used to be what people said about you when you left the room," Ratti points out. "Now it's what a model's saying when there's not necessarily anybody in the room at all." That reframing raises the stakes on everything a company publishes, because a model doesn't hold opinions so much as weigh evidence. As she puts it, AI "does have a weighted average of everything that's ever been written or published about you," much of it produced by third parties rather than the company itself. This is precisely why independent corroboration and earned media carry weight a corporate press release can't buy on its own.

Managing an information supply chain

The scale of reputation management pushed Ratti toward a different metaphor for the job, one that reframes communications as an operational discipline rather than a messaging one.

"What we're talking about is not just managing messages or channels, it's managing an information supply chain," she says. Like any supply chain, it has inputs, quality control, and bottlenecks, and it can't be run by one function alone. Communications should own the meaning and reputation dimensions, in her view, but marketing, investor relations, legal, and digital all feed the same chain, which makes the work a team sport with a hard requirement: clear objectives and defined swim lanes so the contributing functions don't pull the company's story in different directions. When sources are contradictory, she notes, the cost is concrete. "If the models find conflicting information, then they tend to drop you to the bottom of the priority list."

Consistency is the C that holds the rest together

Partway through describing her five Cs, Ratti added a sixth on the fly, and it turns out to be the one underwriting all the others. Consistency, she asserts, is what keeps both people and models from receiving a muddled picture, and it's the discipline most companies underrate.

"Consistency is key," Ratti says, tying it directly to trust: inconsistent information is not merely untidy but actively misleading, feeding conflicting signals to humans and machines alike. The failure mode she warns against is treating the problem as one of volume. Flooding the zone with more blog posts, Q&As, and listicles may help a company show up in answer-engine results, but it does little for corporate reputation and can actively hurt it if the additional material introduces contradictions. Rather than more content, the goal should be more coherent content, aligned to a clear sense of what the company wants to be known for.

Keep a human at the end of the chain

For all her focus on machine-readability, Ratti resists the idea that communicators are now simply writing for machines. Her P&G training left her with a conviction she returns to repeatedly: there is always a person at the end of the chain, and the machine is only a conduit to them.

"We need to make sure that at the end of the day, the machines are being guided by a person," Ratti says, adding that whatever a company publishes still has to be understood and verified by a human rather than optimized purely for machine consumption.

That principle shaped how she introduced AI at Ferragamo, where she brought in and sponsored a platform for the marketing and communications teams. "The thing that really made the difference for the team was linking back to what Ferragamo was in the business of and how AI could really help everyone do their jobs in a more efficient and impactful way." This included tackling work her team wanted to do but never had time for, such as drawing on a hundred years of brand history to tell stories that would otherwise go untold. Handled that way, Ratti believes, AI stops being a tidiness chore and becomes an asset, provided its output is orchestrated, curated, and edited by people who know what the company stands for.