How to Lead AI Transformation Across an Enterprise Marketing Team
Most enterprise marketing teams are not being held back by a lack of AI tools. They are being held back by a lack of strategy.
In a recent episode of the Marketing AI Sparkcast, host Aby Varma sat down with Chole Tambe, director of AI transformation at HubSpot, to explore what it actually takes to drive AI adoption at scale. Tambe is responsible for enabling hundreds of marketers to put AI to work in their day-to-day roles. Her approach is deliberate, people-first and grounded in data gathered directly from her team. The conversation is a practical guide for any marketing leader navigating this same challenge.
Start With a Survey, Not a Strategy
Tambe's first move when stepping into her role was not to roll out tools or launch training programs. It was to listen. She sent an anonymous survey to the marketing team to understand where people were in their AI journey, what was holding them back and what they genuinely needed. The anonymity was intentional. When people feel safe, they tell the truth.
The survey surfaced four clear barriers to AI adoption. Tambe used those barriers as the foundation for her entire strategy. Rather than building a plan around what was available or popular in the market, she built it around what her team actually needed. This approach gave her a defensible rationale for every initiative she launched and helped her stay focused in a space where distraction is constant.
For enterprise marketing leaders, this is a replicable starting point. Before investing in tools, licenses or training programs, ask the team what is stopping them. The answers will shape a more effective strategy than any industry framework.
Build a Strategy Around Human Barriers, Not Technology
Once Tambe had her survey data, she organized her work into four strategic pillars: comprehensive AI enablement, a knowledge-sharing infrastructure, setting the pace for AI innovation and driving operational efficiency. Each pillar maps directly to a barrier her team identified.
Comprehensive AI enablement means showing marketers what they can practically do in their roles, not just handing them a license and expecting results. Knowledge-sharing infrastructure means creating systems for peer-to-peer learning so the team is not entirely dependent on one person. Setting the pace for AI innovation means piloting tools with small groups before rolling them out broadly and communicating expectations clearly. Driving operational efficiency means partnering with the marketing technology team to build workflows and agents that make the work easier.
This framework is useful because it is grounded in outcome rather than activity. Each pillar answers a specific question the team was asking. Enterprise marketing leaders building a similar marketing AI strategy should anchor every initiative to a named barrier or goal. Without that connection, programs become difficult to prioritize and even harder to sustain.
Design Training That Works for Both Beginners and Power Users
One of the more candid moments in the conversation came when Tambe described what she called falling flat on her face. In an early proof of concept with roughly 160 marketers, she ran four weeks of training sessions where everyone attended the same class regardless of skill level. By the second week, half the group said the content was too easy and the other half said it was too hard.
Her response was to redesign the approach entirely. She moved to a 101 and 201 model where marketers could opt in based on a clear description of what each course covers. The 101 course is designed for people who are new to AI or hesitant about it. The 201 course is designed for those who are already comfortable and want to go further. People choose their own track, which reduces the likelihood of disengagement and meets learners where they are.
This kind of differentiation matters enormously in large marketing organizations. When a single training approach is used for a team with widely varying levels of experience, the result is that almost no one is well served. Enterprise leaders who invest in AI training and enablement should design learning programs with at least two tracks from the start and describe them clearly so that self-selection is easy.
Build an AI Center of Excellence That Marketers Actually Use
Tambe is building an AI center of excellence for the HubSpot marketing team. She thinks about it as a repository with three distinct sections: one for marketers who are new to the company, one for those who are on their AI journey and one for power users. Each section offers resources tailored to where that group is in the adoption curve.
Beyond segmentation, she emphasized the importance of featuring use cases from within the team itself. Seeing a colleague's short video explaining how they use AI in email outreach or demand generation is far more compelling than hearing from an outside expert. It creates a direct connection between the tool and the actual job, which is what drives adoption.
The center also includes recordings of live training sessions, curated AI news and a link to the team's dedicated Slack channel, where marketers share what they are learning in real time. Tambe described that channel as one of the most valuable signals in her work because it shows her what is resonating without any filter.
Manage Change With Transparency, Not Just Communication
Tambe's approach to change management is built on visibility. She created an internal wiki that outlines her plan by month so any marketer can see what she is working on and why. This removes the uncertainty that often accompanies transformation efforts and replaces it with a clear narrative.
She is also partnering with HubSpot's CMO to document expectations on both sides: what the marketing team can expect from her function and what is expected of them in return. Framing it as a mutual commitment rather than a directive changes how the message lands.
For skeptics, her approach is direct engagement rather than avoidance. She leans into one-on-one conversations, asks why people are resistant and looks for the specific use case that will create an "aha" moment for that individual. Some skeptics, she noted, become the most vocal advocates once they find one practical application that works for them. This is worth remembering when organizations feel tempted to focus entirely on early adopters.
Treat AI Governance as a Shared Responsibility
Tambe is clear that AI governance in marketing cannot rest on one person or one team. At HubSpot, procurement, legal and security review every AI tool contract in detail before anything is approved. The marketing team operates within those guardrails, and expectations are communicated clearly: do not copy and paste AI output without reading it, and do not put sensitive customer data into tools that are not approved for that purpose.
She also flagged an area that many marketing leaders overlook: governance around AI meeting notetakers. Questions about whether recording is on by default, whether participants are notified and how that data is stored require deliberate decisions. These decisions carry real implications for trust within the team and with external stakeholders.
Marketing AI governance frameworks will continue to evolve as new tools emerge. Tambe's advice is to document what you know, partner cross-functionally and keep common sense at the center of every decision. The standard will keep moving, but teams that have built the habit of asking the right questions will be better positioned to keep up.
Make AI Agents Approachable Before You Deploy Them
Agentic AI is one of the most discussed topics in marketing right now and also one of the most misunderstood. Tambe addressed this directly by running a session where marketers built agents that had nothing to do with work. The only rule was that the use case had to be personal. People built a movie recommendation agent, a gift finder and similar tools.
The result was a shift in how people thought about agents. An agent stopped feeling like a complex technical system and started feeling like something working on your behalf while you focus elsewhere. That reframing is important because it changes the question from "how do I build this?" to "what do I want this to do for me?"
Tambe's broader point about agents is one that enterprise leaders should take seriously. Build with intention. Agents created for novelty or to meet an arbitrary technology goal will not deliver meaningful value. The same discipline that applies to other marketing investments applies here: start with the problem, connect it to a measurable goal and build something that makes work meaningfully easier.
What Marketing Leaders Should Prioritize Now
The final advice Tambe offered applies whether AI transformation is a formal role or an informal responsibility. Think about the team you have, not the team you wish you had. Start from where people are, not where you want them to be.
Find the champions who are already curious and bring them into the effort as collaborators. Do not avoid the skeptics. Seek them out, understand what is driving their resistance and see whether a practical demonstration can shift their perspective. Acknowledge that transformation is overwhelming for everyone, including you, and let that acknowledgment be part of how you lead.
Write a plan for the next month and the next quarter. Do a small amount each day. Transformation does not happen in a single initiative. It happens in accumulated small moves made consistently over time.
Ready to Build Your AI-Enabled Marketing Team?
Building a human-centered AI transformation strategy takes more than the right tools. It takes the right approach. If you are a marketing leader navigating AI enablement, adoption or governance, contact us to explore how Spark Novus can help your team move forward with clarity and confidence.
FAQs for AI Transformation in Enterprise Marketing
-
Start by surveying your team anonymously to identify the specific barriers to AI adoption before building any strategy or investing in tools. Grounding your plan in what the team actually needs produces more sustainable results than starting with technology.
-
An effective AI center of excellence should include segmented resources for different skill levels, peer-generated use case content from within the team, training recordings and a dedicated channel for real-time knowledge sharing. Use cases from colleagues tend to drive adoption more effectively than content from external experts.
-
Engage skeptics directly through one-on-one conversations, identify their specific concerns and find a practical use case relevant to their role. Many skeptics become advocates once they experience a personal result. Avoiding them slows transformation and misses valuable feedback about gaps in your approach.
-
Design at least two training tracks based on skill level, such as a foundational 101 course and an advanced 201 course, and allow team members to self-select based on a clear course description. Running everyone through the same program leads to disengagement at both ends of the skill spectrum.
-
AI governance in marketing should be a cross-functional effort involving procurement, legal and security teams. Clear expectations around data use, tool approval and output review should be documented and communicated to the entire team. Governance frameworks need regular updates as new tools and use cases emerge.
-
Introduce agents through low-stakes, non-work exercises that let team members build and interact with agents on personal use cases. This shifts the perception from a complex technical system to a practical tool that works on your behalf while you focus on something else.
-
A human-centered approach means prioritizing where team members are in their adoption journey, addressing resistance individually and measuring success through organic behavior change rather than tool usage metrics alone. It treats adoption as a people challenge first and a technology challenge second.
