hyble is an on-demand point-of-sale marketing platform for the global beverages industry. Used by the world’s largest beer, wine and spirits suppliers and distributors, it enables teams to create brand-compliant menus and materials in minutes, print locally, and track performance across more than 50 markets.
hyble operates with a lean internal product and engineering organisation and had a clear ambition to become AI-first, with a strong desire to bring new AI features to their clients. They engaged Vaul Labs to provide the senior AI and engineering expertise they needed on an interim basis, and the structured delivery approach to help them get there.
The Challenge
hyble had the AI ambition, the customer demand, and the data. What they needed was the alignment and pace to turn that into delivery.
- There was a strategic imperative to align commercial, product, engineering, and data functions around a shared set of priorities, and to build a way of working that would allow those teams to collaborate with customers and ship iteratively.
- Decision-making across the teams needed to be faster and more iterative, with the confidence to commit to a direction, learn from it, and move forward.
- Customer feedback was not yet built into the development cycle, meaning new features were being released without a structured way to measure their impact or improve them.
- hyble had 15 years of valuable menu data and expertise that had not yet been fully shaped for use in new AI features, representing both an opportunity and a foundation that needed to be built on.
Underpinning all of this, hyble had identified the need for experienced product and engineering leadership to set direction, back the team to make confident decisions, and drive the pace of delivery. With AI development accelerating, the CEO and Board recognised that closing that gap was a matter of urgency, not just ambition.
Our Approach
We worked with hyble across two quarters, embedding fractional Engineering and Product leaders directly within their organisation.
In the first quarter, the priority was alignment and pace. We partnered with the product and engineering teams to reorient the roadmap around the AI features that would deliver the most immediate customer value, introducing a structured delivery cadence built around iteration, staged rollout, and measurable progress.
The clear milestone for the quarter was structured customer testing. We established a user testing group with hyble’s largest enterprise customer, involving around 20 users, and worked with the product team to design feedback sessions, define what was being validated, and build a repeatable process for acting on the findings. That testing also surfaced a strategic insight: building AI into the existing platform would not deliver the product hyble’s customers needed. Acting on that finding quickly, leadership made the decision to build a greenfield product instead. What might have taken another year to discover was resolved in a quarter.
In the second quarter, we worked with hyble to make a structural shift in how the organisation was set up to serve its market. Product and engineering were restructured around three distinct customer segments/goals: end users creating menus in the platform, enterprise buyers focused on compliance, cost control, and operational management, and finally a data and infrastructure team to scale and support the roadmap. We also reoriented the data function toward work that would directly support new AI features and give leadership clear visibility over infrastructure costs as AI spend began to scale.
Alongside this, we supported the hiring of a permanent engineering leader, defining the brief and shaping the role around what the organisation needed next. Throughout this process, the Vaul team ensured that any structural changes made would align with the overall approach of the appointee and have time to embed.
The second quarter also produced hyble’s first production AI agent. Built to encode 15 years of menu knowledge, compliance rules, and domain expertise, the agent gives customers a capability that a generic AI tool cannot match: outputs shaped by deep sector knowledge rather than general training data. It is now running in production and represents a durable competitive advantage for the platform.
Across both quarters, the goal was to build durable capability inside the organisation, not to deliver a set of outputs and step away. Every decision, process, and hire was designed to leave hyble stronger and more self-sufficient than before.
Outcomes
“In 2025 we had the ambition to build AI into our product. In 2026, working with Vaul Labs, we have made that a reality. When we showed our biggest customer what we had built, their response was immediate: ‘This is a huge step forward. We’re really excited about this new product and we think our users are going to love it.’”
Craig Letton, CEO, hyble
From twelve months of ambition to live customer testing in six weeks
- Within six weeks of engagement, hyble had a working AI feature in structured testing with 20 users from its largest enterprise customer.
- Presenting the prototype to the hyble board marked a turning point. With a working product in front of senior leadership, the conversation shifted from what AI might eventually deliver to what it could deliver next.
A production AI agent that gives hyble a sustainable competitive edge
- hyble now has a production AI agent that encodes 15 years of domain expertise, menu psychology, and compliance knowledge into every output. It is not a generic AI tool with a company logo on it: it is a capability that could only have been built by an organisation with hyble’s depth of sector knowledge.
- In direct comparison testing, the hyble agent consistently outperforms a general-purpose large language model given the same prompt. For customers, that is a compelling reason to stay in the platform rather than reaching for a generic alternative.
- The agent was designed and built in three weeks, a timeline made possible by the organisational groundwork laid in the first quarter and hyble’s decision to move quickly when the right model capability became available.
An organisation structured and led to scale
- Product and engineering are now structured around the two distinct customer segments that drive hyble’s commercial growth: end users and enterprise buyers. Each team has a clear mandate, its own roadmap, and direct accountability for delivery.
- A permanent engineering leadership hire was made during the engagement, with Vaul Labs shaping the brief and supporting the selection. The organisation now has the leadership layer it identified as critical at the outset.
- Key engineers were mentored and coached directly by embedded Vaul Labs practitioners throughout the engagement. The organisation carries that capability forward independently.
- The data function has been more closely aligned with the rest of product engineering and is now oriented toward growing the foundation of menu intelligence that will directly support the next generation of features.