What Elite CMOs Are Really Doing with AI: Insider Insights from Canva, Stripe, and HubSpot

Many of us in tech live deep in the echo chamber. We feel the urgency and anxiety to keep up, knowing that Generative AI has been adopted at rates 2-3x the internet or the personal computer. Every week, product releases from across tech announce leaps forward, hurling us toward a science-fiction future. Doomsday warnings of job loss, deepfakes, and evil AGI bounce around the echo chamber. And yet hope and possibility emerges from top thinkers as well; I really enjoyed the recent treatise?Machines of Loving Grace from Anthropic’s CEO Dario Amodei.

But what are CMOs tasked with marketing today doing with AI? How are they thinking about it for their products and their teams? How are they staying ahead of legal and privacy concerns?

I was honored to join the CMOs of Canva, Stripe, and Hubspot to discuss this topic. Watch the replay by registering for the series HERE .

In preparation for the discussion, I talked to a dozen elite CMOs (Including leading a SaaStr session with the CMOs of Snowflake, Carta, and LinkedIn (AI@3:58), and an AI Session with Sprout Social, Seismic, and Deepgram .) This blog will cover a few highlights from the Canva webinar and what I’m hearing from many different CMOs:

1. Top CMOs Are Prioritizing AI for Their Customers
2. Internal Projects are In Experimentation - Fewer Are At Production Scale
3. Governance, Privacy, and Security Are a Top Priority
4. Agility, Resilience, and Learning Mindsets Reign
5. Some Cynicism Exists Behind Closed Doors

#1 Top CMOs Are Prioritizing AI for Their Customers

Top CMOs are excited about productivity and innovation gains for their teams, but even more critical is having exceptional AI IN their own tech products. Canva’s webinar was spurred in part by big AI announcements they’ve made over the last few months with their AI tools that feel like and are called “magic .” Hubspot has had many AI releases across its portfolio and features AI stories on almost all of CMO, Kip Bodnar’s “Marketing Against the Grain, ” podcasts. Stripe just announced a deepened partnership with NVIDIA and AI products for fraud detection , strengthening its leadership.

While everyone is announcing something, incumbents are rightly cautious about disruption from AI-native scaleups (who have unusual access to capital). On the other side, deep-pocketed giants with new AI capabilities could also be coming for our lunch. Everyone in tech feels the urgency to launch a steady stream of effective and impactful AI features in their products and establish leadership in a volatile market. Much of tech marketing includes the vision and delivery of AI (and now AI agent) features. Everyone’s customer advisory boards are discussing AI. Naming, branding, events, content, thought leadership and more revolve around trying to establish credible expertise in an AI world.

Vertical Applications Defend Their Turf, Horizontals Verticalize.

A few years ago, I advised an AI company that had a horizontal product for any and all audiences. As any marketer or salesperson knows, this is an incredibly difficult charter on a limited budget (and even Fortune 500 are limited vs competitors!). Because the product could do “anything,” it was hard for any one person in any industry or job role to visualize exactly what it could do for them. People want templates, best practices, and terminology in their contexts - for instance, the worlds of pharma and financial services are very different. Many technologies with vertical expertise are betting that if they add AI to their products effectively, their vertical expertise will continue to be incredibly valuable. Horizontal products are rapidly establishing more specialized examples, customer stories, processes, nomenclature, and product context by industry or role.

Data Ownership May Underpin Future Leadership

Some companies that already have a lot of unique customer data see their strength in being the source of training data for the AI they deploy. Salesforce announced this strategy at Dreamforce, claiming that their wide swath of process and customer data would make them the most reliable, insightful, and effective AI Agent in the world. Snowflake and Databricks similarly have AI strength built on top of data ownership and a strong data strategy (“Without a good data strategy, you can’t have a good AI strategy”). Many top CMOs are reinforcing aspects of their historical differentiation that underpin their future AI strength.

#2 Internal Projects are In Experimentation - Fewer Are At Production Scale

While CMOs are rapidly sharing THEIR customer stories of AI success in THEIR marketing, many of their own internal deployments are still mostly in the experimentation phase. I’m hearing about projects in the following categories:

• Creation:?generating text, images, and videos for content, blogs, and social. People are finding the tools helpful for brainstorming, refining, outlines, drafts, edits, visual edits... but find that the human touch is very much needed to get a great final product. (I personally find Claude Pro to have a more human tone than ChatGPT Plus, though I used them both for help with my blogs.)
• Massive personalization at scale: AI is a game-changer for helping companies serve the long tail of different audiences, industries, languages, countries, etc. I’m seeing a number of companies being able to massively replicate landing pages, emails, and web pages. This includes some high-ROI projects around translating content for different languages or massive web projects targeting hundreds or thousands of SEO terms with dedicated web pages. Some are beyond the scale of what’s been possible to date; all are at a lower cost than before.
• Sales and Marketing Sequences: Many companies are analyzing and creating better prospecting emails, optimizing their marketing and sales sequences, making SDRs much more effective, and giving a better experience to prospects. AI promises faster responses, more personalized responses, trained on what works best. In our SaaStr session, Snowflake’s CMO shared they are seeing double the meeting rates for SDRs supported by AI-generated emails.
• Strategy: Lastly, many companies are using AI for deeper strategy projects around product strategy, customer listening, positioning, competitive research, and a number of deep thinking, research, and synthesis projects.

However, most projects are still in experimentation versus being at mass-scale deployment. I’m hearing:

• Most CMOs are still testing and experimenting to see which tools they want to standardize, which processes can be meaningfully improved, and how they can accelerate up-skilling their team.
• Some companies have stronger IT teams leading cross-company experimentation; some marketing teams are leading their AI experimentation themselves.
• Lots of companies use multiple frontier models at once determining if their collaboration suite vendor AI (Google Gemni / Microsoft Co-Pilot), is good enough vs. a more specialized option: ChatGPT / Claude / Mistral.
• Many people testing Writer and Jasper head-to-head for marketing.
• A number of companies’ internal IT departments (including Stripe’s) have set up custom, private LLMs for internal use, trained on internal content, data, voice, and tone.
• Marketers are watching search behaviors of buyers very closely, wondering if Gartner’s prediction that search volume will drop 25% by 2026 thanks to AI chatbots and virtual agents will come to fruition. SEO and paid search advertising have been major digital contributors to marketing growth over the last decade. The speed of change here will test marketers' agility and creativity. One CMO told me she was already seeing decreases of 10-15% on search terms “topped” by Google Gemini answers vs. non-AI-topped terms.

#3 Governance, Privacy, and Security are a Top Priority

Many top CMOs (and companies) are rightly cautious and careful about the privacy and security concerns of AI for their customers and their internal security. As providers of AI-infused technology, they are expanding and sharing their own data policies and security standards, if they are large and established, reinforcing how they are more trustworthy than “fly by night,” “do it yourself,” and “wild west” startups. Internally, they are forming governance boards made up of legal, IT, marketing, product, engineering, sales, support, and other departments that are using or considering using AI. Policies on employee use, tool choice, disclosures of use, what customer, financial, or market data can be fed into which LLMs, and more are being addressed. Questions being asked:

• Will those companies use our data to train their models?
• Can competitors in any way extract processes and data?
• Does this in some way give foreign governments or bad actors potentially dangerous access a vector to us or our customers’ data? Remember the SolarWinds attack by the Russian Government? What is China up to these days?
• How do we retain copyright and legal ownership of content created in concert with AI?
• Could the code itself have imbedded or “back-door” code that’s a risk?

Further, there are issues with the AI results themselves:

• Will they hallucinate, and if so, will your company be held responsible? (Read how?Air Canada lost a lawsuit over a hallucinating support chatbot )
• Will they be persuasive or misleading?
• Will they be biased?

Most enterprise vendors, at this point, provide an enterprise edition LLM, where they promise not to use your data to train their model and publish rigorous trust, safety, and privacy commitments. But as any seasoned technology exec will tell you, we often don’t yet know what we don’t know in the security space…

Carilu’s Sidebar on Ethical AI

I’m a huge fan of Anthropic and follow them very closely for their insights on the dangers and opportunities of AI. While ChatGPT by OpenAI is currently leading the consumer Generative AI space, Anthropic is very close in its API revenue and growing at a faster clip. See the revenue data here . Anthropic is the values-based frontier model made up of former OpenAI executives who didn’t agree with the safety and ethics approach of OpenAI. Anthropic is dedicated to researching how to make AI safer, more reliable, more transparent, and steerable. Their research aims to inform consumers, companies, and developers how to make AI safer, and their own Claude LLMs are a demonstration of how effective AI can be, even when and because they have trained it on human ethics and values. Their work on training their LLM based on their Constitutional AI (values from the Univeral Declaration of Human Rights, Non-Western Perspectives, and more is a must-read) (And not to continue my rant, but the constant departure of key leaders at OpenAI, the use of Scarlet Johanssen’s voice without her permission and the potential legal proceedings which have been swept under the rug arouse my concern about the ethical leadership at OpenAI). Having a massively powerful, somewhat black-box solution under a cloud of ethical ambiguity seems risky. I hope Anthropic can gain even more momentum so the market leader can be values and safety-oriented. *(Note - I am in no way affiliated with Anthropic; all opinions are my own.)

#4 Agility, Resilience, and Learning Mindsets Reign

Generative AI has emerged at an interesting time in tech. Many companies outside of AI and security have been living through an economic slowdown over the last several years with slower growth, higher churn, tighter budgets, and related layoffs. It wasn’t the perfect time for stressed marketers to make time to learn entirely new technologies, tools, and skills. But thanks to the promises of increased quality, productivity, and “doing more with less,” everyone is digging in. (Research by?BCG showed ?adopting AI can lead to 40% higher quality and 25% faster output.)

Many top CMOs organized training for their teams on AI skills, and go-getting employees are driving their own learning. Many first-round trainings were horizontal in nature: “how to use generative AI for marketing processes and projects,” “how to use AI safely,” “our governance rules,” etc. Many CMOs who have done continued training for the teams have focused on more role-specific AI projects and training and proof of concepts for specific processes. A number of different AI consultants have emerged to coach and consult on AI for marketing strategy, product marketing, content creation, SEO optimization, and more specific AI use cases. A few that I follow closely include Liza Adams , Nicole Leffer, and Marcel Santilli .

#5 Some Cynicism Exists Behind Closed Doors

I, myself, live DEEP in the echo chamber, as I’m reminded every month when I lead “Pozorski Family Investing Club” with my sister, 86-year-old dad, and cousins from the midwest who are in retail, real estate, construction, and private equity. While I feel like the whole world is changing at warp speed around me, it’s a blip and a source of cynicism for many outside of tech.

While most of my CMO conversations revolve around the excitement and opportunity of AI, I’ve also been privy to private conversations with CMOs who are less caught up in the hype. These CMOs, often at larger scale companies, reflect on the hype cycles and timelines of the Internet boom, the shift to mobile, the shift to cloud, digital transformation, COVID predictions, and more. These aren’t Luddite skeptics but bright, forward-thinking CMOs who are just more measured in expectations about how substantially AI will affect their business or their teams in the near term. They may be launching AI products, have assigned AI leads in marketing, and have set aside budget for AI, but they’re not as obsessed as the rest of us -it’s a component of their work, not the focus. These CMOs think AI is another interesting development and tool that, while having the potential to make us radically more effective, will experience a slower realization. Many have found AI to be a helpful supplement but not ready for full autonomous trust. Some expect the transition to be driven more improvements in their existing stack than a massive changing of the guard to AI-native tools or massive re-skilling of every marketer.

Conclusion & Next Steps

We covered so much more in the?Canva / Stripe / Hubspot webinar ?than I covered here, give it a listen!

My own journey down the rabbit hole of AI innovation is punctuated by more questions than answers. So what’s next? Let’s keep learning from as many different sources as possible. Coming up this week…

OpenAI on How Their Marketing Team Uses ChatGPT

OpenAI has a session on October 31st on how OpenAI uses ChatGPT in Marketing. They’ll cover:

??Best practices for adopting AI into marketing teams

??9 key use cases that marketing teams will get value from today

??Security practices that keep your data safe

??? Live product demos for getting started

?? Date: October 31 ? Time: 11:00am - 11:45am PT

?? Click here to register

(*I also have no affiliation with OpenAI)

Thoughts? What are you hearing?

Carilu Dietrich is a former CMO, most notably the head of marketing that took Atlassian public. She currently advises CEOs and CMOs of high-growth tech companies. Carilu helps leaders operationalize the chaos of scale, see around corners, and improve marketing and company performance.