Mark Abramowitz Wants Marketing to Put Its Name on the Number

Mark Abramowitz has worked in enterprise software long enough to remember when the cloud needed a sales pitch. The technology was new, the buying model was stranger, and the word itself could make an executive glance upward as if someone were discussing weather. He joined Salesforce in 2005 and stayed for 15 years, moving through product management, enterprise strategy, regional marketing, and product marketing. The vocabulary changed. The work did not: make a complicated thing legible, find evidence that it works, and give customers a story in which they can see themselves.

Two decades later, the clouds have filled with agents. They write, classify, summarize, route, recommend, and occasionally surprise the person who switched them on. Abramowitz, now chief marketing officer of Dataiku, is again standing inside a category whose enthusiasm has outrun its operating manual. His response is not to dampen the enthusiasm. It is to make somebody accountable for it.

That sentence reveals as much about his approach as any polished company manifesto. Leads are a marketing noun. Pipeline is a company noun. Revenue is the noun waiting at the end of the corridor. Abramowitz has said he wants marketing to supply more than half of Dataiku's pipeline and, eventually, to commit to a revenue number. He adds a comic sliver of honesty: he feels “fine, maybe not great, but fine” about being held accountable. Accountability is rarely sold with such accurate emotional packaging.

A career spent where the labels are still wet

Abramowitz began as a product manager, including a stint at Grand Central Communications before Salesforce. That origin matters. Product marketers sometimes live downstream from the thing they describe; product managers have to stand near the awkward machinery. At Salesforce, he worked on Financial Services and Sales Cloud, then spent five years in Australia in enterprise strategy and online marketing roles across Asia Pacific. He left for SiteMinder, the Sydney-based hotel technology company, as vice president of global marketing, then returned to Salesforce to lead marketing for Service Cloud.

He scaled the Service Cloud marketing team behind what became a $6 billion business. The more telling detail is the range: product, region, digital acquisition, global marketing, cloud platform. He was learning the whole route a technology story takes, from roadmap to search ad to sales conversation.

ServiceNow was a different stage of the same education. He joined in 2021, when the company was around $5 billion in revenue, and departed four years later after revenue had more than doubled. His remit moved from platform and creator workflows to all product and solutions marketing. He later wrote that he was proudest of shifting the go-to-market model from products toward solutions and personas: connecting brand, platform, industry, and buyer, while widening the conversation from CIOs to security, operations, people, and finance leaders.

Then ChatGPT arrived. Abramowitz's team brought Now Assist and AI Agents to market through launches, keynotes, demos, campaigns, and customer evidence. The timing gave him a front-row seat to the odd compression of the AI era: a category can travel from technical curiosity to board mandate before the organization has agreed on its definitions.

The traveler learns to notice

There is a personal thread beneath the platform history. Abramowitz grew up traveling. In a 2026 interview, he connected those early encounters with different countries and cultures to a habit of curiosity. Travel made the ordinary visible. What one place treats as natural, another does differently. He now calls noticing an underrated leadership skill.

The idea is modest until it reaches an organization. Companies become accustomed to their own misalignment. Sales and marketing define success differently. Product ships features the field cannot explain. Everyone's calendar is full, so the friction acquires the respectable appearance of work. A leader who can look from outside may see that the pieces are individually busy and collectively incoherent.

Former colleagues describe a manager who supplies direction and room. One former Salesforce report recalled being trusted to repair a leads and pipeline problem in Japan despite not speaking Japanese and never having visited Tokyo. The project worked. The same recommendation remembers Abramowitz as available when needed, but willing to let the person solve the problem. It is a useful managerial balance: clarity without choreography.

He has said that people and culture matter more than strategy, and that in another life he might have become an elementary school teacher. Both remarks fit a product marketer whose actual material is comprehension. A strategy that cannot be taught will struggle to travel. A culture that cannot learn will merely admire the slide.

Give every agent an owner

Dataiku presented Abramowitz with a particularly timely problem. The company had crossed $350 million in annual recurring revenue, served more than 750 organizations, and built its argument around governed enterprise AI. Yet the market was changing from models that produce answers to agents that take action. Bad data in a dashboard produces a bad chart. Bad data in an agent can produce a bad decision at machine speed.

Abramowitz's governance language is unusually domestic. If companies insist on calling agents a workforce, he argues, they should follow the metaphor. Someone hires the agent. Someone gives it access. Someone owns its budget and measures its work. Someone retires it when it stops performing. The point is not that software has become a colleague. The point is that autonomy without a manager is a charming demo and a dreadful organization design.

Inside Dataiku marketing, he is putting the principle to work through three levels. The scheme is deliberately simple enough to remember, which is generally where governance schemes win or lose.

This is governance as graduated responsibility, not a velvet rope around experimentation. Low-impact tools remain close to the people with the ideas. The consequential ones acquire review in proportion to their reach. Abramowitz has described Dataiku's emerging campaign operating model as gold-level work because its decisions cross teams and systems.

His outward message is similarly grounded. In an Ai4 interview, he cited research of 900 CEOs across eight countries in which 96 percent believed employees were already using generative AI without approval. Whether every respondent knew the precise extent is almost beside the point. The executive assumption is that adoption has escaped the official inventory. Prohibition cannot be the operating model for something already present.

The answer, in Abramowitz's formulation, is visibility and business ownership. Governance becomes the condition for moving quickly because it lets the organization know what exists, what it touches, and whether it earns its keep. Speed without that knowledge is merely velocity wearing a conference badge.

The CMO moves closer to the product

Abramowitz also sees the CMO role bending toward product. At Dataiku, he participates in roadmap, strategy, and product-market-fit discussions. The arrangement suits his past and the present market. Enterprise AI products are too complicated for marketing to arrive after the decisions and decorate them. Positioning can expose a product problem; customer evidence can change a roadmap; a buying committee can punish a story that speaks fluently to one persona and nonsense to the rest.

He segments Dataiku's story accordingly. Data teams care about pipelines, predictive work, and reliable data. CIO and AI teams care about building agents around business processes. Operational technology teams care about governing and observing a sprawl of agents built across platforms. One umbrella phrase will not do all three jobs. Precision is not fussiness when a buying committee spans the building.

This brings the story back to numbers. Better AI does not automatically produce better marketing data. Abramowitz is quick to note that customer, opportunity, intent, and conversation data still have to be cleaned and reconciled. The ancient rule of garbage in, garbage out has acquired a motor. The consequences now move on their own.

His career has moved through cloud, hospitality technology, customer service, enterprise workflows, generative AI, and agentic systems. The enduring practice is less glamorous and more useful: notice the seams, sharpen the language, invite the customer into the proof, and name the owner of the result. Technology keeps inventing new nouns. Abramowitz keeps returning them to verbs, budgets, and people.