Key Takeaways
- Most B2B organizations are not failing at AI because of tool access, but because ownership, data integration and insight-to-action workflows remain fragmented.
- AI-ready marketing teams connect trusted data, clear accountability and fast activation so insights can influence live campaigns without delay.
As B2B marketers embrace artificial intelligence (AI) tools, they were not handed them the data, the ownership or the workflows to actually deliver on it. That’s the tension pulling at B2B marketing right now. AI tools are everywhere, adoption is racing ahead, and yet most teams are bolting new technology onto foundations that were never built to carry it.
The numbers from Supermetrics tell the story plainly as only 6% of organizations have fully embedded AI into their workflows. Fewer than 11% describe marketing and data as fully integrated functions. Meanwhile, more than 60% flag the gap between insight and action as a major holdup for their campaigns. The pattern is clear: the constraint isn’t ambition or access to AI. It’s the plumbing underneath it.
To unpack what separates the AI-ready from the merely AI-curious, we sat down with Andrea Linehan, CMO of Supermetrics. Linehan brings a practitioner’s view to where marketing keeps getting stuck, and what it takes to break through. In this Q&A, she digs into the AI ownership gap and who should really be accountable for execution, what a best-in-class insight-to-action workflow looks like, and how to tell whether your organization is genuinely AI-ready or just experimenting.
Demand Gen Report (DGR): Andrea, thanks for taking time out of you schedule to answer our questions today. Your report highlights an “ownership gap” around AI. In B2B organizations, who do you believe should ultimately own AI-driven marketing execution and why: marketing ops, data teams, revenue ops, or senior marketing leadership?
Andrea Linehan: Happy to be with you. Senior marketing leadership has to be accountable for what AI is expected to achieve for the business, but ownership needs to be much clearer at each point in the process. Collection should be shared, but governed by whoever sits closest to the customer, and marketing has to be meaningfully involved there. Without that involvement, organizations optimize for what’s easy to capture rather than what’s useful to understand, and the result is data that’s technically clean but strategically hollow. Management of that data, the integrity, security, governance and scalability layer, should sit squarely with data and engineering teams, not marketing. Analysis is different again: it has to be jointly owned by marketing and analytics, because insight without behavioral and commercial context becomes statistical theatre, and analysis without that context becomes guesswork.
Where I feel most strongly about ownership is activation. Marketing has to control how an insight becomes an action in a campaign. Otherwise, you can end up with something I see far too often: data that is visually impressive but still leaves you with the ‘so-what’ question. You have dashboards explaining what happened without giving marketers enough confidence to decide what to do next. That’s where the accountability gap becomes real. You can’t reasonably ask a CMO to own an outcome while giving their team very little influence over the inputs and workflows that determine it.
DGR: You found that only 6% of organizations have fully embedded AI in workflows. What separates those few organizations operationally from the rest, especially in complex B2B MarTech environments?
Linehan: The biggest difference is whether AI is connected to a system that already works. Our research found plenty of AI activity. Teams are using it for content creation, market research and reporting. Those are useful applications, but they tend to sit at the task level. The harder step is connecting AI to reliable data and then allowing an insight to move through the company quickly enough to affect a decision. That requires clear ownership, consistent data, integrations between analytics and activation platforms, and agreed rules around what AI can do autonomously and where a person needs to make the call.
You can see why so few companies have reached that point. Our research spans all business sizes. Fewer than 11% describe marketing and data/analytics as fully integrated functions. Meanwhile, more than 60% of these teams flag the gap between insights and action points as a major holdup for their campaigns. The test I like is very practical: if an insight surfaces on Monday morning, can the marketing team confidently change what is live in the campaign by Monday afternoon without opening a ticket and waiting three days? That tells you much more about AI readiness than the number of AI tools in the stack.
It also comes down to a shift from volume to quality. Most teams built their data habits when more data was automatically better. Now the constraint isn’t volume, it’s whether people trust what they’re looking at enough to act on it. If the effort of using the data outweighs the value someone gets back from it, it gets deprioritized no matter how sophisticated the tooling is.
Where Should MarTech Teams Invest First: AI Tools, Data Integration or Governance?
DGR: For MarTech professionals trying to justify stack investments, how should they prioritize between AI tools, data integration, and governance if budget only allows one major investment area?
Linehan: For many B2B companies, I would put the investment into the data foundation and integration layer first. That recommendation comes directly from what marketers told us. Forty-six percent said better data integration across platforms would do more to close their capability gaps than any other single investment. And only 7% say their data requests are answered in real time. Half are waiting one to three business days for an ad hoc answer.
If the underlying data is fragmented, slow or inaccessible, adding another AI application can give you faster output without giving you greater confidence in that output. That’s a dangerous place to be, particularly in analytics, where an incorrect answer can look perfectly plausible. Governance still matters, but good governance needs something concrete to govern. I would start by making the data reliable, accessible and connected, while establishing clear ownership and controls around that foundation. Then you have something AI can genuinely build on.
DGR: The report suggests many teams are using AI mainly for content and reporting rather than decision-making. What practical steps can B2B marketing teams take to move from AI-assisted output to AI-informed optimization?
Linehan: Start with a decision rather than an AI use case. Pick a recurring decision the team already has to make. Which audience should receive more investment? Which campaign is underperforming relative to pipeline contribution? Where should we reallocate budget this week? Then work backwards and ask what data is required to make that decision confidently, how quickly that data is available, and whether the resulting insight can actually reach the system where somebody can act on it. That exposes the gaps very quickly.
Our research shows that AI use cases today cluster around content creation, research and reporting. Shifting the way teams use AI, however, can accelerate the speed of decision-making and increase confidence in those decisions. The next stage of maturity is connecting AI to specific decisions with trusted inputs, clear success measures and a feedback loop, while treating every recommendation as evidence to weigh rather than an answer to accept. Did we act on the recommendation? What happened? Was the recommendation correct? That feedback is what begins turning AI from an output generator into something useful for optimization.
What Are The Most Important Integrations Martech Teams Need In Place For AI Tools To Succeed?
DGR: Since proving ROI across channels remains such a major challenge, what measurement framework do you think is most realistic for B2B teams managing long sales cycles and multi-touch attribution?
Linehan: I’d be careful about promising B2B marketers a perfect attribution model. Long buying cycles, multiple stakeholders, offline interactions and complex journeys make that unrealistic for many businesses. A more useful framework connects marketing activity to a hierarchy of business outcomes. At the top, you have revenue and profitability. Beneath that, pipeline creation and progression. Then you have the leading indicators that marketing can influence more quickly: qualified demand, engagement from priority accounts, conversion rates and changes in channel or campaign efficiency. The important part is making those layers traceable enough that the business can see how a marketing decision relates to a commercial outcome over time.
Part of why this is hard is that marketing has spent years over-optimizing for explainability rather than believability. We’ve become very good at reporting what happened, impressions, clicks, dashboards, and much less confident arguing why it mattered or what should happen next. A measurement framework only earns trust if it closes that gap, not if it just produces more of the same activity metrics with more precision.
Our research actually shows marketers already behaving this way— 59% link marketing goals to revenue growth, 36% to profitability and 28% use data-driven predictions. At the same time, 40% say proving ROI across channels is their biggest challenge when justifying resources. The goal should be decision-grade measurement with enough evidence to decide where the next dollar or hour should go, and transparency about the assumptions behind that decision. AI can help with that analysis, but it also makes transparency around those assumptions more important.
DGR: The report mentions that fragmented systems are a core blocker. From your perspective, what are the most important integrations MarTech teams need in place before AI can actually improve campaign performance?
Linehan: I’d map integrations around the journey from signal to action. First, your core marketing sources need to feed a consistent data layer through paid media, web and product analytics, CRM, marketing automation and the relevant first-party customer data. For B2B companies, CRM is important because campaign activity has to connect with accounts, opportunities and revenue. Second, the analytics layer needs access to that data quickly enough to support a live decision. Third, and this is the piece teams often miss, information has to flow back into the activation systems. If AI identifies a valuable audience, budget opportunity or underperforming campaign but someone still has to export a CSV, create a ticket or manually update another platform, you’ve preserved the bottleneck. That last mile is showing up very clearly in the research: roughly 36% of teams say they’re missing integrations between analytics and activation platforms.
DGR: The data shows marketers are accountable for outcomes without controlling the data strategy. How can MarTech leaders build stronger influence over data governance when ownership sits outside marketing?
Linehan: Bring the conversation back to the decisions the business needs to make. Data teams have legitimate priorities around architecture, security, quality and governance. Marketing brings another essential perspective: which data needs to be available, at what level of granularity and at what speed for it to be commercially useful. That’s why I think marketing needs a seat at the table from the collection stage onwards. Without that input, it is very easy for a company to optimize for the data that is easiest to capture rather than the data that helps somebody make a better decision.
A MarTech leader can make this concrete by documenting a handful of high-value decisions and showing where the current data process slows them down. “We need better governance” is abstract. “It currently takes us three days to answer this question, so the campaign opportunity has passed before we can act” gives data, marketing and leadership something specific to solve together. Shared service levels can help as well with agreed definitions, owners, refresh frequencies, quality thresholds and escalation paths. That turns marketing’s involvement in governance into a requirement rather than a debate.
What Measurement Framework Works for Long B2B Sales Cycles?
DGR: For B2B teams with intermediate data maturity, what early AI use cases create the most strategic value without increasing the risk of misleading insights from poor-quality data?
Linehan: I’d choose bounded use cases where a marketer can inspect both the source data and the recommendation before anything changes in market. Anomaly detection is a good example. Flag a sudden change in spend, conversion rate or campaign performance and ask a marketer to investigate. Summarizing performance changes across channels is another, provided the AI is working from governed data. AI can also help surface patterns in audience or account performance and identify questions worth investigating. Those applications reduce the amount of manual searching without asking the system to make an unsupported leap.
The filter I actually use is simpler than a maturity checklist: which decisions could this quietly distort, and where would false confidence be most dangerous? Anomaly detection and performance summarising pass that test because a marketer can still see the underlying evidence. Autonomous budget allocation doesn’t, yet.
I’d be much more cautious about autonomous budget allocation or executive-level performance narratives when the underlying data is still inconsistent. Reporting errors are uncomfortable because they can be difficult to spot. A bad creative suggestion is usually obvious to a marketer. A plausible but incorrect attribution number may make it all the way to a leadership meeting. Intermediate skills themselves shouldn’t be seen as a disqualifier. Fifty-one percent of teams in our research rate their data competency as intermediate. What matters is giving those people reliable, accessible data and enough visibility into how the AI reached its answer to challenge it when necessary.
DGR: Your findings point to a gap between dashboard insight and campaign activation. What does a best-in-class “insight-to-action” workflow look like for a modern B2B demand gen or lifecycle marketing team?
Linehan: It should feel almost boring because the mechanics happen in the background. The data from your relevant marketing, CRM and customer systems arrives in a common environment automatically. Definitions are consistent. An anomaly, opportunity or audience change is surfaced quickly. AI can help analyze what changed and suggest a course of action, but the marketer can see the underlying evidence and understand why that recommendation appeared.
From there, the recommendation reaches the right person with enough context to approve, reject or adjust it. Once approved, it flows back into the activation platform without another export or manual handoff. Then the resulting performance data comes back into the same environment so the team can see whether the action worked. That final feedback loop is important. Otherwise, you have automated the movement of data without improving the quality of the decision. The benchmark I keep coming back to is asking if something you learn in the dashboard on Monday morning can affect what a customer or prospect sees on Monday afternoon? If the answer involves a ticket queue, spreadsheet export and three separate teams, there’s still work to do.
DGR: If you were advising a VP of Marketing or Head of Marketing Ops at a B2B company today, what three signs would tell you their organization is genuinely AI-ready versus just experimenting with AI?
Linehan: The first thing I listen for is whether a team is asking which decisions are actually slow, wrong or politically compromised, or whether they’ve jumped straight to which tools they should be using. Tool-first questions are usually anxiety dressed up as strategy. I’d then look for three things. First, somebody can tell me who owns the outcome. There’s a defined AI strategy, specific use cases, clear owners and success measures. That sounds basic, but early findings from our AI Readiness Poll showed only 15% had reached that level of clarity. Second, the team trusts and can access the data AI is using. They know where it came from, how fresh it is, what the definitions mean and what limitations exist. They aren’t spending three days finding an answer that AI is then expected to act on in seconds. Third, insights can change execution. AI is connected to a real workflow and the company can show examples where an insight led to an action and the impact of that action was measured.
I would ask for evidence rather than a list of tools. Show me a decision AI helped you make last week. Show me the data behind it. Show me what changed as a result. If a team can do that comfortably, I have much more confidence in its readiness than I would from hearing that everyone has access to an AI assistant.
