The debate over AI marketing team size has moved from conference panels into budget spreadsheets, and Colin Piper’s account of running a 130-person department before cutting to a team of 20 puts the operational logic on the table in unusually candid terms. Piper, CMO of BuildOps, describes a period at his previous employer, Autodesk, where headcount had grown to a point that actively impaired the work it was supposed to enable.
The problem, as Piper tells it, was not ambition but structure. More than 130 people sat within his marketing department globally. Layers of management multiplied to the point where he felt disconnected from day-to-day execution. A target of one quarter (13 weeks) to take a complex campaign from inception to launch regularly stretched to 20-plus weeks. A planning session with 30 marketers in a room became an exercise in circular discussion rather than decision-making.
The symptom that crystallised things was a role his team created: the programme manager. The function existed, Piper says, almost entirely because the team had grown too large to coordinate itself without a dedicated go-between for product marketing, demand generation, operations, and external stakeholders. A leaner organisation, he argues, would likely not have needed it at all.
The structural cost of empire-building
Piper is direct about his own culpability. He describes falling into what he calls the empire-building trap: the assumption, common in large organisations, that headcount is a proxy for influence and budget. Passion projects found shelter within the larger structure, he says, because size created opacity. In a tighter organisation, every piece of work is more visible and more likely to align with what he calls the North Star marketing goal.
Autodesk, for its part, said in a statement that it has modernised its operations and transformed its marketing over the past four years. ‘We’ve embraced AI to move faster and get more done, and we’ve made accountability and experimentation core to how we work,’ an Autodesk spokesperson said. ‘The result is an effective, modern growth machine that’s delivering for our customers and our business.’
The broader context matters here. The last 18 to 24 months have accelerated a structural shift that was already under way. Tools such as Claude Code and others, Piper says, have forced marketers to confront how much of their output can now be produced at scale without equivalent headcount. His framing is not that people are dispensable, but that the ratio of people to output has changed materially.
AI marketing team size at BuildOps: 20 people, six-week campaigns
At BuildOps, Piper runs a team of 20. He describes it as a miniature version of the Autodesk structure, without the layers. The organisation uses what he calls first-party, enterprise-level AI: a system connected to internal documents, a data warehouse, and email, rather than a generic large language model. Different agents are orchestrated to work in parallel across the team.
The practical result: a campaign developed at a leadership offsite with five people went from idea to fully launched in under six weeks. The output included a long-form research asset and a suite of creative work. Piper calls it mind-blowing by the standards of what was previously possible.
External data supports the direction of travel. According to Marketing Mary, content production velocity increases by 35-50% with AI assistance, and a marketing operations team augmented by AI can deliver output comparable to a team 30-40% larger. The implication for headcount planning is direct: the denominator in the output-per-person calculation has shifted.
Execution speed is also improving at the tooling layer. Averi documents a case in which one organisation replaced a collection of disconnected tools with an integrated AI platform, reducing execution time by 40% and improving campaign performance by 25%. The pattern Piper describes at BuildOps (consolidation, integration, fewer handoffs) tracks closely with those findings.
Piper’s advice to other CMOs is to resist the instinct to delegate execution upward. Get closer to the work, he says. Care about the numbers and the operations, not only the vision or the brand. In a structure where AI is working horizontally across the team, a single well-positioned leader can cover ground that previously required several layers of management to traverse. The team that launches a campaign in six weeks is not doing less work; it is doing the same work with a fundamentally different ratio of people to output. That ratio is now the number CMOs need to be able to defend.