Independent, founder-led B2B marketing and sales agency, competing in a market being reshaped by AI. B2B marketing and sales agency, delivering strategy, creative and sales enablement for specialist B2B sectors.
Via helped Velo design and build Olev, an AI operating system, that cut delivery costs by 20% within three months of rollout, while adding new service capability and accelerating delivery timelines rather than trading one for the other. Client feedback has stayed consistently positive throughout, the clearest evidence that speed and cost discipline have not come at the expense of quality.
The Challenge
This was The Shift. Velo, an independent B2B marketing and sales agency, was contending with the same disruption reshaping every specialist agency: AI and martech were changing how marketing is planned, produced and delivered faster than most operating models could keep pace with, and clients were increasingly insourcing the creative and production work agencies had traditionally owned. In the words of Velo’s own leadership, staying current wasn’t a growth opportunity to weigh up at leisure, it was “a threat to our very existence.”.The agency had already tried to respond, through an internal initiative built to explore AI and new ways of working. But exploration on its own wasn’t enough. Without clear ownership, a shared framework for deciding what to adopt and what to drop, and a way of coordinating curiosity across the agency without derailing fee-earning delivery, the initiative risked becoming another well-intentioned effort that quietly stalled.
Impact across the business
20% time saving
>50 models in the ecosystem
Agents, GPTs and GenAI in symphony
Used daily
Our Approach
We worked closely with the Velo team to reboot that stalled initiative into something more fundamental: not a team running occasional experiments, but Olev, an operating system built into the core of how the agency delivers its work. The design started with accountability. Every initiative needed a named owner, a defined output and a deadline, because as the team put it themselves, being “too busy to drive” would kill momentum before it started. We then built a single test, the Prism, that any new tool, workflow or service had to pass before being adopted: does it prove commercial impact, is the cost and rollout time proportionate to the benefit, is it secure and compliant, and is it genuinely relevant to the clients and sectors Velo serves. Initiatives that failed the Prism were retired as deliberately as those that passed it were scaled.
From there, Olev was built out across four workstreams: investigation and adoption, to give the agency the judgement to track what mattered in AI and martech without chasing everything; new services, to respond to clients increasingly insourcing creative and production work; ways of working, to redesign delivery so automation, AI agents and orchestration tools sat inside every team’s practice rather than one specialist function; and agency value, to keep renewing what Velo can credibly claim as its edge as the market keeps moving. Governance was built in from the start: a visible roadmap tracking every initiative’s status, adoption and commercial impact, reviewed face to face every month with the authority to advance, pause or stop.
We took 20% out of our delivery costs within three months, and we did it while adding new services and moving faster for clients – not by cutting corners.”
The results
- Olev embedded AI across every part of the agency’s operating model, not a single tool or team: MCP-based orchestration, AI agents, custom GPTs and generative AI now sit inside strategy, creative production, sales and client service alike.
- 20% reduction in delivery costs achieved within three months of Olev’s rollout.
- New capability added alongside that cost reduction rather than traded off against it, with both strategic and production briefs moving from approval to delivery faster than before.
- Greater agility and sharper strategic and creative thinking, letting the agency move further and faster without diluting the calibre of the work it produces.
- Client feedback has stayed consistently positive throughout, the clearest evidence that speed and cost discipline have not come at the expense of quality.
- Agency-wide adoption now underway against fixed 2026 deadlines: full automation of identified repeatable tasks by 31 December 2026. AI to be used to service by every active client, and to be used by every member of staff by 31 August 2026.
