August is shaping up to be another very busy month for CMO moves. In just the first two weeks, 22 new CMOs have been announced globally: with a perfectly even split of 11 women and 11 men.

External hires continue to dominate: 16 came from outside the organization, versus just 6 internal promotions. Half (11 executives) are stepping into the CMO role for the first time. Meanwhile, the “industry traveler” remains a rarity, with only 2 of the 22 making a cross-industry move.

The U.S. remains the busiest market, accounting for 16 appointments across 10 states. California leads with 6 new CMOs, followed by New York with 2. Outside the U.S., England and India each announced 2, while the Netherlands and the Philippines added one apiece.

Tech continues to drive CMO demand, with 8 appointments so far this month, followed by restaurants with 4 and CPG with 3. Government, NGOs & higher education and media, sports & entertainment each added 2, while retail, financial services and manufacturing each added 1.

We find every CMO interesting. This week we really struggled to shortlist the 3 CMOs to feature. We failed.

So here are five of the best: a CMO hired to make Twitch’s weirdness valuable, another trying to save Otter from becoming a feature, a man tasked with explaining what ServiceNow now is, a Square marketer selling warehouses full of Nvidia chips while China questions the whole premise, and an internal promotion charged with making compliance “a little zingier.”

^{As usual, our pen is icy. So our disclaimer is here.}

For years, Twitch embodied everything the modern media business found awkward.

In 12 years working at the world’s largest media buyer, I heard a planner mention it maybe once in a meeting. I saw it namechecked on a few résumés.

It was messy, expensive to deliver, difficult to control, with weak discovery and built around gamers watching other gamers. Elsewhere, social media became shorter, faster, more algorithmic, and much easier for agency planners to buy.

AI now social feeds with synthetic slop at industrial scale. Suddenly, Twitch’s weirdness looks like a plus. That is the setting for Christine Cassis taking over as CMO. After leadership roles at Shopify, Pinterest, Reddit and Google, she has been handed a difficult brief: protect the gaming community that built Twitch while expanding its relevance in music, fashion, sports and entertainment.

Cassis calls Twitch’s problem a “perception gap.” Millions know the platform, but many still understand it as the place where young men watch other young men play video games.

CEO Dan Clancy makes the same argument but more bluntly. He has called social media “antisocial” and said: “AI is trying to take humans out of the equation, and live is one of the formats that keeps humans at the center.”

Twitch has stumbled into an interesting inversion. The chaos is still there, the world changed around it, and chaos became easier to sell.

Her interviews make the fit clearer. At Shopify, she said her team stopped behaving like “a traditional broadcast desk” and started thinking “more like journalists on the ground.” Her better line was this: “Corporate megaphones aren't nearly as effective as passing the mic.”

It’s a fitting philosophy for Twitch. The value comes from communities which are often deeply suspicious of corporate marketing. The problem is that Twitch still has to turn this human advantage into a better business.

Cassis has made discovery one of her three priorities. So, Twitch wants better algorithms to help people find communities while arguing that algorithmic feeds have made the rest of the internet worse. Uh-huh.

Her Shopify background is notable, but I don’t think she’s been hired to create QVC for Gen Z. I would expect creators, products and transactions to be welded together more closely. More commerce could improve creator dollars and reduce dependence on annoying advertising. Push it too hard, though, and “deeply human” starts looking suspiciously like another shop window.

I once had a glimpse of corporate Twitch. The company approached me about a Head of Talent Acquisition role. I expected a conversation about what it was trying to solve and why it had called. Instead, we went immediately into Amazon behavioral-interview mode: “Tell me about a time when…”

I remember thinking: hang on, you called me.

I did not proceed to the next round. One interview proves nothing about a whole culture, but I thought about it when Cassis described Twitch as “live, unscripted, participatory….” One of the internet’s liveliest products sits inside the giant warehouse that is Amazon. There is tension in that.

For years, Twitch’s long, messy and unpredictable humanity made it harder to package than the cleaner social platforms around it. Now the internet is filling with polished synthetic content.

Cassis has been handed the weird stuff. She just has to make sure nobody tidies it to death.

Otter.ai is doomed if it stays a notetaker.

For a while, you met on Zoom, Meet and Teams and Otter followed you everywhere. It transcribed, summarized and gave you something searchable. That was novel and wonderful when native meeting intelligence was crap. Now Zoom, Microsoft and Google are furniture in that house. Zoom AI Companion summarizes meetings. Teams brings Intelligent Recap. Google Meet has Gemini notes in Workspace. In most companies, it’s hygiene with no extra password friction.

Otter solved meeting notes so well, that it stopped feeling like a company. Otter now calls itself a "Conversational Knowledge Engine." (eeew) It connects conversations across a company. People can query what was said across hundreds of meetings. It integrates with Salesforce and HubSpot. It turns calls into coaching and pipeline intelligence. That might be the escape route, but it’s also a much harder business.

Fireflies.ai is already deep in CRM. Fathom has used free recording aggressively. Granola and Krisp made bot-free capture normal. Salespeople already tolerate Gong. Zoom is expanding notes beyond its own meetings. MS and Google own the calendar, identity and documents that surround every conversation.

Otter has to convince companies that their most important conversations should sit above all of that. And neutrality is only valuable if the product is meaningfully better.  It also needs a clear answer to the procurement question: why add another SaaS product to store sensitive conversations when the tools you already pay for do the same thing?

If the new ambition is to own the memory of the company, that’s going to require the new CMO to accomplish a Feat of Strength: awareness, acceptance of the pivot and trust.

Lots of people already know what Otter does. The job is to convince the market that what Otter does next deserves its own category.  And the giants are not waiting politely while Otter figures it out.

Kurt Apen, Otter’s first CMO, says he helped take the company from low-single-digit millions to more than $100 million in revenue. So the new CMO Alex Gay is inheriting a successful machine whose original category is being swallowed by the platforms around it.

Gay comes from Superhuman/Grammarly - another narrow use case made into a broader enterprise play - and he is already describing the bigger prize: “Voice is one of the richest, most underutilized sources of knowledge inside every organization.”

Otter’s old promise: it remembered the meeting, so you did not have to. Its new promise: it remembers the company.

That’s a claim on the enterprise stack. Zoom, Microsoft and Google will not surrender it. Gong, Salesforce and the rest are already fighting over the commercially valuable parts.

Otter has escaped the notetaker category only to enter a far worse neighborhood.

ServiceNow paid $7.75 billion for Armis. The deal came with a CMO.

Simon Mouyal was just named Chief Marketing Officer 3 months after the acquisition closed. He was CMO at Armis. He moved across to lead security GTM, now he gets the whole company.

ServiceNow’s previous CMO, Colin Fleming, left in May to become OpenAI’s CMO for Business – we wrote about that here. ServiceNow lost its marketer to one of the companies threatening to rearrange enterprise software, then replaced him with the marketer from the cybersecurity company it had just bought.

Follow the marketer —— and you learn where ServiceNow thinks the next act lives.

The company says Mouyal will accelerate growth “from workflows to cybersecurity.” Armis expands ServiceNow’s security ambitions. Mouyal has been CMO at CyberArk and athenahealth. He worked at Microsoft, HP and Rackspace. This is a cyber marketer being handed a company that would like the market to stop thinking of it as the place where IT tickets go to die.

ServiceNow now calls itself the “AI control tower for business reinvention.” It sells workflows, AI agents, security, CRM, HR technology, enterprise search and anything else that can be orchestrated through its platform. It has spent billions buying into adjacent categories. Which creates a marketing problem.

What the fuck is ServiceNow?

Mouyal has been thinking about this for years.

In an interview 8 years ago, he said: “The first thing you have to do to define your brand is to really clarify what category you’re playing in.” Good luck, Simon.

He told another story from his time at Medidata. His team removed the logos from competitors’ websites and compared the language. Everyone shouted the same things: faster, safer, cheaper. You could not tell which company was speaking.

Enterprise AI has recreated that problem at scale. Remove the logos today and everyone promises agents, orchestration, intelligence, transformation and measurable outcomes. ServiceNow has the extra complication of operating across so many categories it’s difficult to describe.

Mouyal’s answer at Medidata was to get brutally clear about the business the company was really in, then elevate the conversation above individual products.

There is another Mouyal story that matters more.

He joined Microsoft shortly before the launch of Windows Vista. By his telling, everyone knew Vista was “not a great product,” but marketing still had to produce the biggest launch possible. They did. Sales exploded for two weeks and then fell away when customers discovered that, indeed, Vista was not a great product.

Marketing cannot keep the fries hot.

The Vista episode tells you something about Mouyal. He learned early that category stories only survive if the product underneath earns the language. ServiceNow is making giant claims. CEO Bill McDermott calls the company the “AI control tower for business reinvention.” Mouyal says enterprises face “AI chaos” and argues ServiceNow can discover, secure and govern it through one platform. His better, older line from Armis is simpler: “You cannot secure what you cannot see.”

Security gives Mouyal something tangible to work with. ServiceNow connects assets, identities, workflows and AI agents, and argues that enterprises need one place to understand what all those systems are doing. Armis gives that claim a more concrete foundation.

Comprehension is the issue.

Mouyal has to explain why a workflow company is also an AI company, why an AI company is also a cybersecurity company, and why customers should trust ServiceNow to sit above software from Microsoft, Salesforce, Google and a small army of AI startups.

Eight years ago, Mouyal said the first job in defining a brand was to clarify the category in which it plays. ServiceNow has handed him one of the hardest versions of that assignment.

It spent $7.75 billion buying Armis and found a top marketer inside the box. Now he has to tell us what ServiceNow bought itself into.

Nebius is basically an AI landlord with a very expensive electricity habit. It builds massive data centers, fills them with Nvidia chips, and rents the resulting compute to companies that need absurd amounts of it. Think AWS, except specifically for AI.

Competitor CoreWeave is the American version that got there first and much louder. Same idea: borrow mountains of money, buy mountains of GPUs, sign contracts with Microsoft, OpenAI and friends, then use those contracts to justify buying another mountain of GPUs. Wall Street calls this “AI infrastructure.” Reddit has less polite terminology. At this point the joke that AI capex is holding up the US economy like a jack under a fat cow: the Fed’s polite version says AI-related software, equipment and data-center investment has made a ‘meaningful contribution’ to recent GDP growth.

Nebius wants to become the global alternative. It has hired Lindsey Irvine from Square to help make that happen. She led global marketing across brand, demand, product and hardware there, having previously served as CMO of Benchling and MuleSoft. At MuleSoft, ARR grew from roughly $250 million to more than $1 billion. Nebius CRO Marc Boroditsky says he wants her to build demand “at every layer of the market, from enterprise buyers to the developer community.”

That is the assignment hiding inside this appointment. Nebius has plenty of infrastructure, giant contracts and money flowing toward it. It needs a much larger population of customers who can explain why they chose Nebius rather than AWS, Azure, Google or CoreWeave.

Across the AI ouroboros, the same names keep appearing as chip suppliers, investors, customers and partners. Nvidia sells the GPUs and invests in the companies buying them. Cloud companies borrow against long-term contracts to build more capacity. Giant technology companies then make promises to consume that capacity years out. Everybody gets another reason to spend and build.

Irvine has to turn “we own warehouses full of Nvidia chips” into a differentiated company vs a well-financed utility bill.

If AI suddenly requires far less compute, Irvine cannot brand her way around the economics. Silicon Valley is spending mountains of cash on the assumption that AI use will continue swallowing compute. Meanwhile, China has spent years operating with tighter access to the best American chips, trying to spin up domestic alternatives and giving its engineers a country-wide imperative (and bonus national pride) to do more with less.

If that keeps working, GPU rental prices are toast. Expensive chips depreciate while they are still warm. Data centers financed on presumed scarcity start looking like enormous sheds full of electronics that you never wanted in your zip code anyway.

Nebius is betting heavily that this doesn’t happen.

Irvine says “the AI cloud category is still being created” and describes Nebius as having “a business that is compounding.” She’s not wrong. Her job is partly to make sure that compounding comes from a broad market progression, not an elaborate exchange of billion-dollar promises and premises among the same small group of techbro billionaires.

Then there is the geopolitics.

Nebius emerged from the non-Russian remains of Yandex and is headquartered in Amsterdam. That gives it an interesting story in Europe, where dependence on American technology companies carries new political risk. Trump’s policymaking makes “technology sovereignty” sound like sensible risk management.

So Irvine may eventually have to sell two versions of the same company. In America, Nebius is infrastructure required to keep the country competitive in AI. In Europe, it can make the case that handing the entire AI stack to American hyperscalers is a horrible idea.

Both versions run on Nvidia.

And somewhere to the east, China keeps raising an irritating possibility: perhaps intelligence does not need quite as many warehouses as everyone has ordered.

Irvine has to build demand, establish trust, explain why Nebius deserves to exist alongside much larger clouds and help investors believe the billions going into concrete, electricity and Nvidia chips will earn colossal returns.

She has joined a company betting that intelligence will remain hungry. While China may have other ideas.

If you don't know what SOC 2 is, be glad. It's up there with BASEL 3 and ISO standards in the great arc of things humanity created because apparently doing basic business was not stultifying enough.

SOC 2 is basically a way for a company to prove that it handles customer data responsibly. The process involves controls, evidence, audits, screenshots, policies and large quantities of documentation that somebody absolutely promised would be finished last Tuesday.

The company automates much of the security-compliance work startups and, increasingly, large enterprises need to prove they are not storing your medical records in Martin from Engineering’s Dropbox or DeepSeek’s Google-indexed shared-chat archive. It has more than 16,000 customers, passed $300 million in ARR and was valued at around $4 billion last year. I think they have a bright future, as vibe coding explodes, more SaaS equals more TAM.

Now it has promoted Sarah Scharf to CMO.

Her job gets harder from here because “we make SOC 2 suck less” is a fantastic proposition but a finite category. Vanta now sells third-party risk management, AI governance and a broader trust-management platform. It wants to become what Scharf calls a “trust company.” Oh dear.

Trust is one of those words every corporate brand ‘discovers’ in a strategy workshop. My thesis is that there is less trust in the world than ever, so everyone is making out they have it.

Scharf knows the danger. She joined Vanta in 2020 as its first product marketer and helped build much of the marketing machine that got it to where it is today. Her answer is to keep the brand playful in a category where everyone else appears to have hired the same designer of ominous blue shields.

Vanta once plastered San Francisco with billboards saying: “Compliance that doesn't SOC 2 much.” You must admit, it is good.

Scharf says she wants Vanta marketing to retain “a wink.” Her description of the challenge is wonderfully modest: “Some of that is going to be bland and boring, but what can we do to make it just a little zingier?”

Imagine becoming CMO of a $4 billion software company and your mandate includes making compliance a little zingier. It’s an excellent moment to try.

AI is making trust substantially harder to establish. Somebody eventually has to prove where the information went and whether the robot behaved itself. Vanta would very much like to be that somebody.

Scharf says the opportunity is to show how Vanta helps companies “instill and maintain that trust in a world where AI makes it harder to come by and maintain.” The old Vanta helped prove your systems pass an audit. The new, bigger Vanta wants to prove your company, suppliers and increasingly your AI can be trusted. Bigger market but a much squishier promise.

Vanta is spending on influencers and podcasts and has launched The Tabletop, where CISOs role-play security disasters in a fake situation room. This produces content security professionals might even watch.

We have therefore reached the point where companies film executives pretending to suffer cyberattacks so AI will tell other executives to buy their software.

Scharf is also using AI inside marketing for asset resizing, editing and campaign prototyping. She now screens candidates for AI proficiency, although she assures Business Insider this is “not from a tokenmaxxing perspective.”

There is one final wrinkle.

Vanta hired Scott Holden as CMO last year. About a year later, he is leaving for an opportunity he has not yet disclosed, while staying on as an adviser to Scharf. That is a quick CMO transition, although the handover appears unusually cuddly.

Holden says Scharf wrote the “Compliance that doesn’t SOC 2 much” line he had admired from afar, was his onboarding mentor and became his right hand. CEO Christina Cacioppo goes further. She says Scharf was the person most responsible for architecting Vanta’s shift from automated compliance to trust management - and, more recently, agentic trust.

So this is not an outside CMO arriving to impose a bigger story on a narrow product. Scharf helped write the story, built much of the team that tells it and now has to prove it can travel.

Given how often companies respond to growth by importing an outsider CMO with a bigger vocabulary, moving a proven insider up the ladder. It is a sensible move and we applaud it. The person now being asked to stretch Vanta beyond compliance is also the person most likely to remember what made the proposition work.

Scharf has to keep the wink while turning Vanta into something bigger than automated compliance, without turning it into another SaaS claiming to deliver trust, resilience, transformation and peace of mind through a dashboard.