Briefing note: Mandate, intelligence, orchestration
Insights from a closed-door dinner with senior B2B marketing leaders, hosted by Bright at Noble Rot, Soho.
In May 2026 we brought together a small group of senior marketing leaders from across B2B, events, regulated industries and information services for the latest in our series of dinners on AI-enabled marketing. Six months ago, the room was still arguing about whether AI would eat their lunch. This time the room was past that. They were debating about how to win commercial outcomes from it – and what their teams need to look like to make it happen.
Three themes ran through the evening. The progression in the room was clear: leaders are moving past point-tool use cases and starting to think in orchestrated agentic workflows – while being honest about the climb between that ambition and what their teams can deploy today. Reading the conversation back, the themes map well onto the maturity stages we already see clients move through – curious and exploring, testing and learning, enabling and scaling, adaptable and embedded.
The mandate gap and the commercial credibility problem underneath
Every leader at the table had an AI mandate from their business. None of them felt it was a well-formed one. The official version is “be more efficient with what we have.” The real version is “do more with less, faster, but it still has to be brand-perfect, and by the way here’s a Copilot licence and limited training.”
The result is patchy: licenses without enablement, governance arriving after adoption rather than before, and team members quietly using whatever tool they prefer with no-one tracking inputs or outputs. Several leaders described running diagnostics across their teams and being surprised — both by how much shadow AI use was happening, by how few people were actually using the tools available to them, less than 50% with only a handful of super users, and by how little of it was joining up into anything strategic.
Sitting alongside the mandate is a quieter question of commercial credibility. Consumption-based pricing is harder to defend internally than per-seat SaaS used to be, and marketing functions are increasingly being asked to attach commercial outcomes to AI spend earlier than the use cases have matured. For at least one leader in the room, increases in revenue per employee to baseline AI impact against felt like the lever most likely to reframe how marketing’s contribution gets measured, which is a more interesting framing than pure cost or efficiency.
The leaders furthest ahead had reframed the mandate. Rather than efficiency, they were asking “how do we use AI to grow?” — including for roles that don’t traditionally sit on the commercial line. That reframing changed which use cases got prioritised, which budgets got approved, and which team members got trusted to experiment.
From tools to the marketing intelligence layer and the GTM engineering threat
The conversation moved past tools quickly. What teams are actually wrestling with is the operating system underneath: workflows, foundational data, value propositions, brand voice, governance, and the role of marketing in the wider business. The shift from running single AI use cases to orchestrating agents end-to-end was the direction of travel — but most leaders were honest about where they currently sit on that curve.
Several leaders described building what we’d call a marketing intelligence layer or brand brain. Essentially a curated foundation that sits underneath the AI stack and pulls together the things that make marketing work in the first place: ICPs and TAM analysis, value propositions and messaging architecture, customer evidence and case studies, sales playbooks, workflow definitions, governance and guardrails, and (yes) brand and tone-of-voice training. The point is not the corpus itself. It’s that any agent, any team member, and any new hire can be pointed at it and produce something that actually sounds and behaves like the business. Building one takes weeks of patient work, much of it interview-led. But once it exists, output quality jumps everywhere it touches.
Foundations beat features here. Without that layer underneath, agentic tooling amplifies noise rather than signal, and the gap between ambition and real-world implementation is wide enough that most leaders were openly working out where to get help to close it.
Sitting next to this is a sharper threat: the rise of GTM engineering pods. Several leaders described work that used to belong squarely to marketing: ICP construction, TAM modelling, list-building, pipeline automation – moving into engineering-adjacent teams running tools like Clay. In some cases those pods sit offshore. In others they sit in revenue operations or sales. Either way, marketing teams that don’t develop the orchestration capability fast risk being narrowed back to brand and comms, while the growth work moves elsewhere.
The transitional pattern that this is generating was one of the most pragmatic findings of the evening. Several leaders described outsourcing part or all the agentic and workflow build for specific marketing capabilities to an outsourced partner further ahead – while their internal team builds capability in parallel. Outsource to learn, in-source when you’re ready. It’s a rational response to the maturity gap, and it’s a model we’re running with a number of clients today with the core enablement focused on the internal team governance and ways of working, so that the outsourced elements work seamlessly as part of the operating model.
The doer-to-orchestrator shift
The most striking observation of the evening was about the next-generation marketing hire. The work itself: the copy, the design, the list, the campaign asset — is moving to agents. The role that remains is the one that orchestrates them: defines the outcome, sets the brief, validates the output, applies critical thinking, and lifts the team’s quality bar.
The hard part is human, not technical. Most mid-career marketers were trained to do, not to manage. Asking them to step back, set the brief and validate the output of something they don’t fully understand isn’t just a skills shift – it’s a change-management problem. People want the new; they also, very reasonably, fear it. Several leaders said openly that retraining their best operators had proved harder than expected, and that the profile they’re now hiring for is different: people drawn to orchestrating something unfamiliar, comfortable sitting with the discomfort of not fully understanding what’s under the hood. One described the future generation of marketers as content engineers. Another, more bluntly, said the next great hire will be a puppeteer – and that puppeteer may not have a marketing CV at all.
The displacement is already visible inside marketing. Junior copywriting briefs, list-build roles, asset-production work, i.e. exactly the early-career rungs the function has historically grown its talent on, are being absorbed into platforms and agentic workflows. The hiring profile that’s growing is the marketer who can sit between the commercial outcome and the agent, fluent in both. That profile is in short supply, expensive, and being aggressively poached.
What we’re watching
- Consumption pricing is starting to reshape vendor conversations. Leaders are increasingly wary of locking in multi-year per-seat SaaS commitments while the shape of AI workloads, and the way they’re charged for, is still moving.
- Foundations beat features. The clients getting commercial value from AI did the unglamorous work first: clean ICPs, sharp value propositions, agreed brand voice, a defined process. Without those, agentic tooling amplifies noise.
- Outsource-to-learn is going mainstream. Marketing leaders are buying capability and adoption from outside while their own teams build alongside. The partners they’re choosing look less like agencies and more like operating-model designers.
Culture and executive air-cover are doing more work than tools. The leaders running the most experimental programmes all said the same thing. Their executive teams are visibly enthusiastic, and the team feels free to fail. Where that’s absent, no license rollout fixes it.
The clearest takeaway of the night: the leaders in the room are no longer asking whether AI will reshape marketing. They are asking what their teams will look like on the other side of it and the honest, repeated answer was we are still working it out. The conversation has moved from anxiety to architecture. Ways of working and governance are very much the backbone combined with solid marketing fundamentals.
If you’d like to explore how we can support your team, we’d love to talk.