AI is only as good as your data. Marketing now owns its first-party data, but only in partnership with governance. And the CMO wins by thinking as a business leader, not a function.

Persson was drawn to marketing by the cinema commercials of 1980s Sweden, where television carried no ads so everyone went to the movies to watch them. She grew up on the Levi's and Juicy Fruit lifestyle advertising, read the advertising magazine Resume cover to cover every week, and set herself the goal of becoming a product manager at Procter & Gamble. Then, by accident, she joined the tech company Commodore in 1993, got her first computer and watched a whole world open up, a journey she dates back 33 years.

The through-line has been putting the customer at the centre of everything and at the forefront of the marketing, because customers trust their peers, not vendors. That is hard at a tech company, she says, because tech companies are obsessed with what they can create with the technology and do not think enough about what customers can achieve with it. The discipline is to think from the outside in every single day.

By her account Snowflake had around 40 customers, had been in market for 10 months and had not done a million dollars in total revenue when she joined in 2016, the smallest company she had ever worked for. She was even sceptical of the name. What pulled her in was a problem she had lived: marketing had become highly digitalised while the data infrastructure had not followed, leaving data fragmented across channels with no way to measure the whole mix, no attribution and no real-time optimisation. Snowflake's mission of letting everyone work with all their data in real time convinced her it would change everything.

Persson credits staying constant on the customer's problem and the company's why, in Simon Sinek's terms: helping organisations unlock their full potential with data and AI. The mission expanded over 10 years, from unlocking data for real-time intelligence to becoming the data foundation for agentic AI, but it stayed true to one core, helping customers across their entire data life cycle.

In the early days the pain sat with data teams and IT, serving a business that could not get the data it needed. Now, Persson argues, every organisation runs on data and, in a world of agentic AI, data and enterprise context is the most important asset. Marketing wants agentic journeys, personalisation and real-time optimisation of the channel mix, sales wants customer context in real time to be more relevant, and finance wants to close the books faster, so Snowflake has moved from solving infrastructure for IT to working alongside the lines of business.

The data is only the raw material. The thing everyone is really after is the intelligence, and you need all that data and enterprise context available to you before you can bring it out.

The journey is changing rapidly, and agentic search is becoming incredibly important. Where something like SEO could once be owned by a single team, agentic search pulls in the entire marketing department, from communications to content, because all of it affects the result. It has become a cross-functional discipline, Persson says: someone still owns it, but now it involves everyone, a big change from the past.

Invest in both, Persson argues. Demand generation harvests the buyers looking today and gets you in front of them before competitors, while brand invests in being top of mind for the buyers of the future. Tech companies lean too far into demand because the pressure is on pipeline and performance is easy to measure, but durable growth is about making your numbers in the future, not just this year.

Brand is also where long-term trust is built, and Persson believes trust will be the most important currency going forward. Personalisation depends on customers being willing to share more data, and they only share it with companies they trust. Without that trust, she says, the promise of AI can never be delivered.

There is a race among marketing departments to own their customers' first-party data, which used to sit in ad tech platforms and with media buying agencies. Pulling it in-house now demands a partnership with a centralised data and AI governance team, because marketing can no longer just buy and implement technology on its own. Someone has to own the quality, Persson says, because AI will only ever be as good as the data feeding it, and without the data in order the full potential of AI stays out of reach.

The companies becoming agentic enterprises share two traits: their data and enterprise context is governed, safe and compliant, and AI is a top priority owned at CEO level. Persson is blunt that AI innovation cannot be delegated to someone else in the company. It only works as a mandate from the highest level, with the CEO in the hackathons building agents alongside everyone else.

In B2B, Persson notes, sales teams record customer conversations through tools like Gong with consent, and capturing all that intelligence on an account makes it far easier to become relevant. An account manager with it captured and accessible can quickly see the next best step, what to focus on and who else to involve. Preparing for a meeting can take a rep days, she says, but with the intelligence available the decks can be prepared in minutes.

Persson says CMOs now need AI and data fluency themselves, hard as that is, because you cannot expect your team to be fluent if you are not willing to be. Leaders should lead from the frontier, set AI as a priority and experiment first-hand, and that extends to CEOs too. Marketing's core job, she adds, is to be the steward for the customer and apply these technologies to the customer's benefit.

Given the extraordinary expectations on CMOs, Persson argues you cannot be an expert in everything. Lean into your own superpowers, hire a wingman strong on technology, and treat assembling a team of experts as the number one job. She sees herself as a team coach, a conductor of an orchestra. And to earn a board seat, marketers must think as business leaders: it is all about the business as a whole, not marketing as a function, understanding finance and, in B2B, partnering closely with sales.

What matters most now, Persson says, is judgment, creativity, common business sense and integrity, and that takes a long time to build. The problem is not here today, but within five years it will bite, and companies will have to take responsibility for the next generation. She admires the US focus on internships, absent in many European markets, and suggests they may need to get longer, even two years, so people have time to prove themselves. The younger generation brings AI fluency, but it counts for little without the judgment to understand what truly matters to the customer.

What is different this time, Persson observes, is that people learn from their peers rather than external academies that do not yet exist for this. At Snowflake the learning is built internally by marketers, through weekly lunch and learns, enablement sessions and hackathons, and that is why being in the office matters: someone can say they just found a way to save five hours and it spreads immediately. Her advice to the young is in the same spirit: the best careers are discovered, not planned, so get in somewhere, find your strengths and stay a constant learner.