The first version of a product isn't software. It's understanding.

Retreve

Retreve didn't begin with a product specification. It began with a conversation about a recurring frustration: despite significant investment in analytics, business users still relied on technical teams to answer relatively simple questions. That observation became the foundation for a much broader exploration into how enterprise intelligence could become more accessible, intuitive and commercially valuable.

As engineering progressed, it became clear that building the platform was only part of the challenge. Before customers could experience the product, investors could support it or engineering could confidently prioritise it, the idea itself needed to become tangible. The product had to be understood before it could be fully realised.

Rather than waiting for software to reach maturity, the product evolved through research, product strategy, brand development, experience design and successive Product Experience Models. By the time the platform was ready, the vision had already been tested, refined and communicated through many conversations with stakeholders, customers and investors. Retreve became more than an AI platform. It became proof that clarity is often the first product an organisation needs to build.

Retreve conversational analytics platform

Executive Summary

Retreve began with an ambitious idea: make enterprise intelligence as natural as asking a question. While organisations had invested heavily in data platforms and reporting tools, accessing meaningful insights often remained dependent on technical expertise. The opportunity wasn't to build another dashboard. It was to rethink how people interacted with enterprise knowledge.

The engagement started with a founder's vision rather than a fully defined product. Engineering commenced in parallel, but many of the decisions that would ultimately shape the platform had yet to be made. Before software could be built with confidence, the product itself needed direction: its purpose, positioning, interaction model and commercial narrative all required definition.

Extant's role was to turn that early vision into a product that people could understand long before the platform reached technical maturity. Through product discovery, market research, strategic positioning, brand development, experience design and successive Product Experience Models, the idea evolved into something customers could experience, investors could evaluate and engineering teams could confidently build towards.

The result was more than a product design engagement. It showed that commercial validation and technical development don't need to happen one after the other. By allowing both to evolve together, Retreve entered development with greater clarity, stronger alignment and a much clearer understanding of the product it was becoming.

Retreve platform from concept to release

One Conversation Changed Everything

Retreve didn't begin with a product roadmap or a detailed specification. It began with a conversation about a recurring frustration that many organisations had quietly accepted as normal.

Despite years of investment in business intelligence platforms, dashboards and data infrastructure, people still struggled to get timely answers from their own information. Business users depended on analysts, and analysts depended on technical teams. The data existed, but reaching it was needlessly difficult. The question wasn't whether organisations needed more data, but whether they needed a better way to interact with it.

That single observation shifted the direction of the engagement. Rather than starting with technology, the focus moved towards understanding how people actually searched for information, how decisions were made inside organisations and where traditional business intelligence platforms continued to create friction. The objective wasn't to improve reporting. It was to rethink the relationship between people and enterprise knowledge.

As those conversations developed, it became clear that the opportunity extended well beyond analytics. Retreve wasn't simply about answering questions faster. It was about making enterprise intelligence accessible to the people making everyday business decisions, and everything that followed grew from that starting point.

Diagram of the Retreve opportunity

Understanding the Problem

The temptation with every new product is to begin designing solutions. We deliberately did the opposite. Before defining features, interfaces or workflows, we spent time understanding the people who would depend on the platform: what questions they were trying to answer, why existing business intelligence tools continued to frustrate experienced users, and where confidence disappeared between asking a question and receiving an answer.

One insight surfaced repeatedly. The problem wasn't access to data, because most organisations already had more data than they knew what to do with. The challenge was turning that data into something people could confidently use without depending on technical specialists.

That changed the conversation. Instead of designing another analytics platform, the product evolved around a different objective: reducing the distance between a business question and a trusted answer. Every decision that followed, from information architecture and interaction design to AI conversations and workflow prioritisation, was measured against that principle.

By understanding the problem before defining the product, engineering gained clearer direction, stakeholders aligned around a common vision and the product developed around genuine user needs rather than technical assumptions.

Retreve user personas
Understanding who would use the platform and how their needs differed across the organisation.
Retreve user journey map
Mapping how questions moved through existing systems and where delays, uncertainty and dependency emerged.
Retreve research findings
The observations that shaped the product before interface design began.

Defining the Product

With a clearer understanding of the problem, the next challenge was deciding what Retreve should become. There was no shortage of possibilities. Enterprise AI was evolving rapidly, expectations were changing almost weekly and it was easy to be distracted by new technologies and feature requests. The discipline lay less in deciding what to build than in deciding what not to build.

Rather than creating another analytics platform with conversational capabilities, the product was positioned around a much simpler ambition: helping people access business intelligence through natural conversation instead of traditional reporting interfaces. That decision influenced every part of the product. Information architecture was organised around questions rather than reports, and user journeys were designed around decision-making rather than navigation. Features were prioritised by how effectively they reduced friction between curiosity and insight, not by how many capabilities could be demonstrated.

As the ideas evolved, the product became progressively easier to explain. That simplicity wasn't accidental. It came from continuously refining workflows, removing unnecessary complexity and challenging assumptions before they reached engineering. The objective was never to create the most feature-rich platform, but the product that organisations would understand, adopt and trust.

By the end of this phase, Retreve was no longer just an interesting idea. It was a clearly defined product with a shared vision, a coherent structure and a roadmap that engineering could build with confidence.

Retreve product blueprint
The first complete view of the product, aligning business goals, user needs and engineering priorities.

Designing Conversations

Traditional business intelligence platforms were designed around navigation. Users were expected to know where information lived, which report to open and how to interpret the results, and finding an answer often meant working through layers of dashboards, filters and visualisations. For experienced analysts this became routine. For everyone else, it became a barrier.

Retreve challenged that assumption. Instead of asking users to learn the system, the system was designed to understand the user. Conversation became the primary interface because it reflected how people naturally think: business leaders don't wake up wanting to explore dashboards, they ask questions, and those questions became the foundation of the product experience.

Designing those conversations required far more than placing a chat interface on top of enterprise data. Every interaction needed to establish context, understand intent and guide users towards meaningful answers without overwhelming them. The experience had to feel intelligent without becoming unpredictable, and powerful without becoming complicated.

This led to a significant shift in how the product was designed. Navigation became secondary to intent, and workflows were organised around conversations instead of screens, allowing users to move naturally from one business question to the next. Rather than forcing people to adapt to the product, the product adapted to the way they already worked. By the time the interaction model was complete, Retreve no longer felt like another analytics platform. It felt like a conversation with the organisation's knowledge.

Retreve agentic conversation storyboards
Exploring how natural conversations could replace traditional navigation and reporting workflows.
Retreve conversation flow map
Designing interactions that understood intent, maintained context and guided users towards meaningful business insights.
Retreve interface evolution
From early conversational concepts to a refined enterprise experience that balanced intelligence with usability.

Building the Brand

As the product became clearer, another challenge emerged. People could understand what Retreve did, but they still struggled to understand what it was. Was it another business intelligence platform, an AI assistant, an enterprise search engine or a knowledge platform? The technology could be described in many ways, but none of those descriptions communicated the product's real purpose.

The brand became the bridge between the vision and the market. Rather than beginning with visual identity, the process started by defining the product's position within an increasingly crowded AI landscape. Every discussion returned to the same question: what should people remember after their first conversation with Retreve? The answer wasn't a list of features or technologies. It was the promise of making enterprise intelligence easier to access through natural conversation.

From that positioning, the rest of the identity evolved naturally. Naming, messaging, visual language and interface design were developed together so that every customer touchpoint communicated the same idea. The brand didn't exist to decorate the product. It existed to explain it.

This mattered most once the Product Experience Models were introduced. Customers, investors and stakeholders weren't simply evaluating screens. They were judging whether the product felt credible, differentiated and commercially viable, and a clear identity made those conversations easier because people understood the product before they understood the technology behind it. By the end of the engagement, Retreve had evolved from an interesting concept into a product with a distinct identity, a clear point of view and a consistent narrative across every interaction. The brand wasn't the finishing touch. It became part of the product itself.

Retreve brand exploration concepts
Exploring different directions before arriving at a positioning that reflected the product's purpose rather than its technology.
Retreve logo evolution
The visual identity evolved alongside the product, creating a brand that felt confident, contemporary and enterprise-ready.
Retreve brand mark
The resolved mark that carried the product's positioning into every customer and investor conversation.
Retreve design system
Creating a visual language that connected brand identity with product experience across every touchpoint.

Product Experience Model™

One of the biggest challenges in early-stage product development is that engineering and commercial momentum rarely move at the same pace. Software takes time to build, but markets don't wait. Investors need something to evaluate, early customers need something to experience and product teams need meaningful feedback long before the platform reaches production, and waiting for engineering to answer every question often delays learning until it becomes expensive to change direction.

Retreve followed a different path. While development progressed in parallel, the product itself continued to evolve through successive Product Experience Models. These weren't prototypes created to imitate software. They were strategic representations of the product, designed to communicate how it would behave, how people would interact with it and why it mattered.

Each version became a catalyst for conversation. Customer demonstrations revealed where expectations differed from assumptions, investor presentations tested whether the product proposition was commercially compelling, and internal workshops aligned engineering, product and business teams around a shared understanding of what was being built. Every discussion generated insights that shaped the next iteration before those decisions became embedded in code.

The Product Experience Models also changed how the product was presented. Rather than explaining ideas through documents or presentations, stakeholders experienced the product directly. They could navigate workflows, explore conversations and understand the value proposition in a way that static specifications could never communicate, so the product became tangible long before it became technically complete.

This approach changed the relationship between product strategy and engineering. Development no longer carried the responsibility of validating every assumption, because by the time functionality was implemented, many of the most important product decisions had already been challenged, refined and aligned through real-world conversations. Retreve showed that commercial readiness doesn't have to begin after development. It can evolve alongside it.

The Product Experience Model became the bridge between vision and execution, allowing the product to mature through understanding before it matured through software. For Extant, this has since become a core principle of how products are shaped: the first version people believe in is often more important than the first version they can install.

Retreve Product Experience Model, version 1
The first interactive experience focused on validating workflows, product direction and commercial understanding before engineering reached maturity.
Retreve Product Experience Model, version 2
Insights gathered through customer demonstrations, investor conversations and internal workshops informed a more refined product experience and clearer product narrative.
Retreve completed product experience
The completed platform reflected months of strategic refinement, allowing engineering and product thinking to converge around a shared vision.

Outcome

By the end of the engagement, Retreve had evolved from an early product idea into a clearly defined enterprise platform with a shared vision, a credible market position and a product experience that people could understand long before engineering was complete.

Instead of relying solely on technical progress to communicate value, the product became a communication tool in its own right. Product Experience Models, strategic positioning and refined user journeys enabled meaningful conversations with customers, investors and stakeholders while development continued in parallel. Feedback that would traditionally arrive after launch was captured much earlier, allowing important product decisions to be validated before they became expensive to change.

Equally important was the alignment it created internally. Engineering teams worked towards a clearer product vision, business stakeholders gained confidence in the direction of the platform and every discussion was anchored around a shared understanding of what Retreve was becoming. The outcome wasn't simply a better interface or a more polished demonstration. It was a product with greater clarity, stronger commercial readiness and a development process informed by continuous learning rather than assumption.

Retreve reinforced a belief that continues to shape how Extant approaches product development: when people understand the product, better decisions follow, and when better decisions happen early, better products emerge.

Retreve completed platform
The completed enterprise experience, reflecting the evolution from an early concept to a commercially credible AI product.

What We Learned

Retreve reinforced a lesson that extends far beyond enterprise AI. Many organisations believe products begin when development begins. In reality, development simply gives shape to decisions that have already been made, and the quality of a product is often determined long before the first line of code is written.

The most valuable work happened before software existed. Assumptions were challenged instead of accepted, users were understood before features were prioritised, and the product was given a clear identity before it was introduced to the market. Customers, investors and engineers could all see the same vision, even though the platform itself was still evolving.

Perhaps the biggest lesson was that commercial understanding and technical development don't need to happen one after the other. They can, and should, evolve together. Every conversation with a customer, every investor presentation and every Product Experience Model created an opportunity to refine the product before engineering committed to implementation.

For us, this changed how we think about building products. We no longer see strategy, branding, experience design and engineering as separate phases. They are different expressions of the same product journey, each reducing uncertainty and increasing confidence before the next decision is made.

Retreve didn't prove that AI can answer business questions. It showed that when people understand a product, they make better decisions about it, and better decisions build better products.

The first version of a product isn't software. It's understanding.

Facing a similar challenge?