Martech 2025: Scott Brinker on new data paradigms & AI’s next act

See, that’s what I do in my downtime. I spend all day running HubSpot’s platform ecosystem, making sure our integrations with hundreds of partner tools work smoothly, so to unwind I dive back into mapping 14,000 products. It lets me relax while sharpening my understanding of where the industry is headed. By plotting every new entrant, I spot emerging clusters such as AI powered analytics platforms, specialized personalization engines or novel customer data infrastructure solutions often before they hit the mainstream. Those patterns tell me how marketing technology evolves and which technical and strategic skills practitioners will need next: data engineering, AI prompt crafting or orchestrating across dozens of APIs. If you find something you love that also fuels your career growth, that is a rare gift. I am very grateful that mapping the MarTech landscape is what I do for fun. ​

It combines my engineer’s curiosity with my marketer’s fascination for innovation. Each time I refresh the landscape, I am effectively doing industry R&D: surveying new startups, tracking consolidations and measuring waves of AI native tools. It is like being a cartographer in a territory that never stops shifting. That continuous discovery fuels my creative thinking at HubSpot. When I see a rise in tools for real time video personalization, I can explore potential partnerships or suggest those capabilities to our product teams. On a personal level, it is immensely satisfying to translate that chaos into visual form, turning 14,000 disparate data points into an intelligible map that helps thousands of marketers worldwide gain clarity.

My first startup was building multiplayer games on bulletin‑board systems as a teenager. In that era, I learned that a single innovation could wipe out entire platforms overnight. During the dot‑com boom, I led technology at a web‑development agency building complex sites for Siemens, Citrix and Seth Godin’s ventures. I mastered client management and large‑scale delivery but felt stifled by one‑off projects where I would build what the client asked for then move on. I wanted to own a product vision end to end. So I founded a company focused on landing pages—a novel idea at the time—that evolved into interactive content such as calculators, assessments and configurators. It was in that work I saw the gulf between marketing and IT. They spoke different languages, had misaligned objectives and often clashed over data and integrations. To bridge that divide, I launched the ChiefMartec blog in 2008 to share insights on how marketers could embrace technology and technologists could understand marketing. After my startup was acquired, I kept complaining about poor APIs and integrations until HubSpot founders Brian Halligan and Dharmesh Shah said, ‘If you want to make it better, come do it here.’ Joining HubSpot let me scale my mapping passion into building a robust ecosystem that now supports thousands of specialist tools. ​

The inaugural landscape was literally Exhibit A in my plea to CMOs. I showed them 150 disparate tools already in use and asked, ‘Who in your organization owns these? Who decides which to add and how to make them work together?’ It created an instant ‘aha’ moment. Each year readers demanded an update, so I refined categories, added new vendors and watched the count double then triple. By around 2016 we hit roughly 11,000 tools, what many thought was peak fragmentation. Then came ChatGPT and the AI boom. Over the next 18 months, an additional two to three thousand tools emerged, many built atop generative‑AI frameworks. The pace accelerated as low‑code and no‑code platforms democratized creation. Any team could spin up a specialized AI assistant for content briefs, support triage or predictive analytics. Today the landscape charts over 14,000 products, each reflecting a slice of marketing’s evolving toolkit.

Categories are inherently imperfect because many innovations straddle multiple segments. I would ditch them if I could, but without some framework the map becomes indecipherable. We maintain broad pillars such as data, content and experience, acquisition and engagement as well as management and operations. Each year we update subcategories to reflect emerging specializations like AI‑driven content assistants or real‑time customer signals. Marketers should start with their objectives—customer journey, campaign goals, data requirements—and use categories only to narrow options. Avoid the trap of thinking you must fill every box. Define your strategy, list required capabilities and then choose the handful of tools that deliver maximum impact.

In the early years, the landscape was about sheer scale and impressing marketers with the proliferation of options. Now, it is more about patterns: where ecosystems are consolidating, which adjacent categories intersect and how AI is redrawing boundaries. We see content‑creation platforms embedding analytics, cloud warehouses offering built‑in CDPs and collaborative suites integrating real‑time personalization. The real value comes from spotting these convergence zones, guiding CMOs to invest not just in standalone platforms but in composable architectures that adapt as new waves of innovation roll in. ​

The report surfaced three core themes. First, the rise of a universal data layer. As more companies centralize data in cloud warehouses and lakehouses, marketing teams gain secure, governed access to first-party signals across sales, product, support and web analytics. That breaks down historical silos and fuels personalization at scale. Second, AI engines have become interchangeable commodities. Every vendor can tap GPT-style models or open-source alternatives. The true differentiator is the quality, governance and freshness of the data you feed into those engines. Our survey found leading organizations investing heavily in data cleansing, compliance controls and real-time streaming pipelines so they can trust AI-driven recommendations and automate decisions with confidence. Third, generative AI is reshaping workflows but has not replaced human strategy. Teams layer AI-powered assistants onto existing stacks to automate content drafts, segmentation tasks and simple orchestrations—but they still rely on skilled marketers to define goals, guardrails and creative direction. Overall, investing in data maturity and governance delivers durable ROI no matter which AI model dominates next year. ​

Capturing a moving target like AI required an agile editorial approach. After drafting each section we monitored product blogs, research announcements and model releases daily. Whenever OpenAI, Anthropic or a major vendor published a new capability, we reassessed our analysis. I rewrote the universal data layer section twice in a single week to reflect fresh examples of warehouse-native CDPs and new governance frameworks. We also recalibrated our predictions on AI agent maturity when vendor previews of autonomous orchestration tools appeared. That iterative cycle ensured the final report reflects the state of the art as of publication. ​

Most current AI agents in MarTech “look a lot like iPaaS” tools: marketers prompt them in plain language—“if this happens, trigger that”—and the agent wires up connectors behind the scenes ​. Common real-world uses include drafting content (“write a social post summarizing this blog”), audience segmentation (“flag leads likely to churn”), and simple orchestration (“when a form’s submitted, sync the contact to CRM and start an email drip”). These remain assistant-led workflows—humans still define triggers, thresholds, and success metrics. Truly autonomous agents that “decide an entire campaign strategy and then run it end-to-end” are still limited to pilot projects at a handful of advanced organizations.​

At its core, a thriving ecosystem balances vision and execution. You need a clear joint value proposition that explains what unique outcomes our combined products deliver for customers. It is not enough to say ‘we’re partners’; you must articulate why a mutual client should care. You also need operational alignment so product roadmaps, sales motions, support processes and marketing campaigns move in concert. Partnerships fail when teams stay in silos. The complexity arises because no single partnership team controls all levers: product managers, solutions engineers, sales reps and support staff all play critical roles. Successful partnerships require what I call influence leadership—carefully weaving these stakeholders into a cohesive program, maintaining communication rhythms, aligning on success metrics and iterating based on real customer feedback.

It usually boils down to two failure modes. First, no compelling joint value proposition. If the alliance does not solve a real customer pain or unlock new use cases, it never gains traction. Signing an agreement is not enough—partners must co‑develop messaging, case studies and joint campaigns to demonstrate impact. Second, inability to operationalize. You can agree conceptually to integrate, co‑sell and co‑support, but unless product, sales, marketing and support teams across both companies are committed, the partnership remains paper‑only. Too often there is no assigned owner for onboarding new integrations, no shared roadmap for feature parity and no coordinated go‑to‑market plan, so the joint solution never reaches customers.

Today I would give us a B plus. Seven years ago, integrations were a terrible afterthought. Now buyers rank integration capabilities among their top three evaluation criteria. Vendors have responded by investing heavily in connectors, app marketplaces and developer tooling. Nirvana would be standardized interoperability where any tool or AI agent can discover, authenticate and call another tool’s services through a common protocol. Anthropic’s Model Context Protocol shows promise. If major platform vendors adopt such standards, we would unlock seamless composability—a fabric of interoperable services rather than dozens of point‑to‑point connections.

 Is it too late to consider a career as a winemaker? … One thing I would say is marketers who are not technologists or marketing ops people should not be responsible for building infrastructure. Their real value is strategy, creative and understanding the customer. Stay plugged into the community of your peers: see what tools they’re posting about, and if something sounds promising, allocate five or ten percent of your time to test it. 

A college math professor told us life is hard so we might as well get used to it. His brutal exams forced us to either freeze or act. He observed that students who started trying even imperfect solutions made progress, discovered insights and eventually succeeded. That lesson—to just do something and iterate—has guided me through unpredictable shifts in marketing technology. In today’s fast moving environment, those who embrace experimentation and learning by doing will always outpace those who cling to perfection or freeze in uncertainty.