Hudex

Hudex as an AI-powered content analysis platform redesigned for broader intelligence work

Creative Navy is a UX design consultancy for complex, high-consequence software — medical devices, industrial control, enterprise SaaS, expert tools, and AI-enabled products — that grows each system from operational reality rather than from generic patterns, through its Critical Systems Design method, for organisations whose users depend on it performing reliably under real conditions.

Hudex is an AI-powered content analysis platform that ingests unstructured data from social media, audio, video, radio, and reports. Hudex then clusters that content semantically so analysts can explore patterns, themes, and signals across large datasets.

The platform's core visualisation is a dondogram. In the Hudex case, a dondogram is a hierarchical tree structure representing thematic clusters at multiple levels of depth. The dondogram was central to the product's capability, but also central to the usability problems identified during research.

The engagement began as a v2 redesign, but the documented work describes it more precisely as a new product design built on a legacy algorithmic foundation. Hudex 1 had been built specifically for social network analysis. Hudex 2 represented a platform pivot from a social media tool to a general-purpose intelligence operating platform capable of processing any modality of content.

Hudex user archetypes ranged from ministers needing rapid overview to analysts needing deep exploration

Creative Navy's research and client sessions established three user archetypes for Hudex: government analysts and diplomats, broadcasting network workers, and intelligence community operators. The archetypes shared core workflows rather than requiring separate feature sets.

For government analysts and diplomats, the case gives the example of the French Ministry of Foreign Affairs. More than 200 diplomatic cables are sent from embassies worldwide to Paris each day. The documented problem was volume and lack of structure: critical information could fail to reach the right people, while even the most engaged minister read approximately 20 cables per day. Hudex provided structured overviews of that information stream.

For broadcasting network workers, the case gives the example of Radio France, which operates 44 local stations across France and overseas. Local news was produced continuously but was rarely aggregated or processed at a national level. Hudex enabled a daily global scan of all content so patterns, movements, and signals could be surfaced across the network.

For intelligence community operators, Hudex supported monitoring specific communities for emerging signals, detecting narratives, tracking threat patterns, and acting on insights rapidly.

The main design implication was the range of access needs. Ministerial-level overseers needed instant comprehension of high-level structure. Expert analysts needed to spend multiple hours per day in deep exploration without losing access to the platform's full capability.

Sandbox Experiments used product immersion before interface redesign

Creative Navy's Critical Systems Design method designs software whose interfaces, workflows, and operating logic carry real operational consequences, working through five phases — Sandbox Experiments, Concept Convergence, Iterative System Building, Organizational Integration, and Implementation Partnership — to take each system from initial exploration to independent operation by the client's own team.

In the Hudex case, Creative Navy used Sandbox Experiments as domain learning through product immersion. The team completed 10 training tasks provided by the client and acted as full users of the demo environment before design work began. The tasks were designed to span all three user archetypes and cover the range of analytic workflows supported by the platform.

This immersion made Creative Navy productive users of the system before redesign work started. During the immersion, Creative Navy identified frictions and gaps not specifically flagged by the client, although the specific findings from those self-directed observations are not retained in the available case record.

The documented evidence for this part of the engagement is Creative Navy-recorded. The domain learning process informed understanding of the dondogram interaction model, the entry point problem, and the cognitive challenge of navigating multi-layered cluster structures.

User research identified the absence of a clear entry point before the dondogram

Creative Navy conducted primary user research with three internal Hudex users. These were the only users accessible at that stage. The sample included one business developer who used Hudex occasionally for demos after less than one month of experience, one analyst and Chief of Staff who had used Hudex for 9 months and multiple hours per day, and one analyst who had used Hudex for 8 months and multiple hours per day.

The research identified an entry point problem. Users arrived at the platform without understanding what they were looking at. The dondogram was not intuitive to new users and required explanation. Clients attending demos did not understand the dondogram or the pyramid unless someone explained them.

One user described the desired alternative as: "Start with a list of the main themes and boom, you're omniscient of your data in five seconds." Creative Navy translated this into a structural design requirement: Hudex needed a summary layer before users were asked to interpret the dondogram.

The research also confirmed depth avoidance. Users were reluctant to explore deeper cluster layers when confronted with the full dondogram structure, and tended to stay at the surface. The client already knew this pattern existed, but user research strengthened and specified the understanding.

Creative Navy's research also identified navigation confusion in the dimensions view. Users described the view as confusing because functionalities were not hierarchical and users did not know where to look in the process.

The dondogram created demo difficulty as well as exploration difficulty

Creative Navy's user research found that the dondogram was not only a navigation issue for analysts. It was also a liability for users in client-facing roles who had to present the product quickly.

One user described the demo problem directly: "For someone working in a bank, having something that looks like a spider is not very inviting." The finding mattered because Hudex needed to be credible to non-expert audiences as well as functional for expert users.

Research also identified elements that worked. The AI chat was considered genuinely useful and well-regarded by clients. The time series view was useful. Theme filters impressed clients. Manual data labelling was valued.

The documented learning curve was approximately 1–2 hours for a new user to learn Hudex once guided. Without guidance, Hudex was not self-explanatory.

The blanks phenomenon appeared in entry-point clarity and the project overview concept

Creative Navy's design work on Hudex documented two instances of the blanks phenomenon. In this case, the blanks phenomenon described gaps between what the client sensed and what had to be specified for the product design to work.

The first instance concerned entry point clarity. The client had a directional intuition that users struggled to orient themselves, but had not identified the absence of a clear entry point as the structural problem. User research specified the issue: Hudex had no summary layer before the dondogram.

The second instance concerned the project overview. A project overview page essentially did not exist as a concept in the legacy product. The client had not specified what it should contain or do. The question was not only how to fix an existing page, but what a project cover page should be.

The project overview eventually resolved around a "book cover" metaphor. The page would give simple, visual, high-level summary information before users entered the data. The project overview required 20 iterations, which was the highest recorded iteration count for any component in the engagement.

Concept Convergence resolved visual direction through sustained iteration rather than a single style decision

Creative Navy's Critical Systems Design method used Concept Convergence in the Hudex case to resolve a significant UI vision tension. The client had a strong sense of the desired visual direction, but could not specify it precisely enough for direct execution. The client knew what felt right when seeing it, but not before.

The visual direction also contained a palette contradiction. The client's brand concept was built around the ARLQ harlequin mascot, representing signal versus noise, with a pastel diamond-tile identity. The founder's visual reference was a scene from the Joker film, using saturated primary colours: red, blue, and yellow. This contradicted the pastel palette and required multiple rounds of exploration.

Creative Navy also explored how to integrate the diamond shape from the ARLQ logo into navigation and background elements. Translating the harlequin identity into functional interface behaviour became a distinct design problem. Some iterations explored the diamond motif directly; others deprioritised it in favour of functional clarity.

Convergence produced the main navigation style, the project overview concept, and the UI visual direction. The resolved direction used a pastel palette, accessible visual language, and a non-intimidating first impression. The documented tension was to make Hudex sophisticated enough to be taken seriously without making it so technical that it frightened people.

Iterative System Building introduced progressive disclosure at the architecture level

Creative Navy's Critical Systems Design method used Iterative System Building in the Hudex case to make a deeply complex system non-intimidating without removing depth. The central design principle was progressive disclosure: users would see complexity in structured layers rather than all at once.

The redesigned structure allowed non-expert users to orient, explore at a surface level, and find value without confronting the full system immediately. Expert users could still access full depth on demand. The case evidence describes this as structured revelation of complexity rather than simplification.

The main architectural change was the introduction of a project overview before the dondogram. The project overview provided high-level theme counts, source counts, and key information before the user entered exploration. From that overview, users could progress into the dondogram or branch into other views.

Iteration counts were Creative Navy-recorded during the engagement. The project overview required 20 iterations. Data exploration required 10 iterations. Dondogram information required 6 iterations. Data visualisation, accounts, homepage, main navigation, and filters required 5, 5, 4, 4, and 3 iterations respectively.

Progressive specification revealed requirements not present in the original brief

Creative Navy's design work on Hudex used progressive specification: requirements became clear through exploration, use, and iterative working sessions rather than being fully known at the start.

The requirement for navigation memory emerged when the team identified that a selected theme was lost when a user switched to another tab. The design requirement was to preserve navigational context across view changes.

A notes editor requirement also emerged. The case describes a rich-text editor at the project and tag level, with the ability to insert current visualisation content directly into notes. This supported analyst workflows where observations need to be captured alongside data exploration.

A universal export system was identified as another requirement. The requirement was consistent export of all visualisations and data.

Side panel behaviour also developed through the engagement. The design requirement was a simplified side panel showing key data, with the ability to expand a single data point, tag directly from the detail view, and navigate directly to a specific visualisation.

Organizational Integration included design education alongside design presentation

Creative Navy's Critical Systems Design method used Organizational Integration in the Hudex case through recurring explanation of design reasoning. Each design presentation included design education content covering user behaviour, cognitive load considerations, and the reasoning behind specific design choices.

This practice meant the client was not only shown interface decisions. The client was shown the thinking behind those decisions throughout the engagement.

The documented organisational work is not described as a separate training programme. It is described as an ongoing accompaniment to design presentation and decision-making during the engagement.

Implementation Partnership preserved design intent during development

Creative Navy's Critical Systems Design method used Implementation Partnership for 7 months alongside and following the Hudex design phase. The partnership included 3 data visualisation animations and 7 developer collaboration sessions.

The implementation support was primarily responsive. Creative Navy answered questions and addressed edge cases as developers implemented the product. The documented purpose was to preserve design intent through the build process.

The Hudex Implementation Partnership did not involve the level of sustained oversight associated with longer implementation partnerships on more complex builds. The available evidence describes a responsive collaboration model rather than continuous implementation management.

Client-reported outcomes covered existing users, new users, and investment attribution

Hudex outcome evidence is mainly client-reported, and the documented case separates the outcome claims from their evidence limits.

A client-conducted survey of 45 existing users found that users rated the redesigned product as significantly better than the previous version. The phrase "significantly better" reflects the client's characterisation of the survey results rather than a documented standardised rating instrument.

In the growth phase following launch, new users rated usability as good (68%) or very good (23%). This result is client-reported. The case evidence does not independently verify the source or methodology of the rating scale.

Three months into the growth phase, Hudex received £3 million in investment. The client explicitly attributed Creative Navy's design work as critical and foundational to the product's ability to sell and to the growth phase itself. The causal relationship between interface quality and the investment decision is client-reported and cannot be independently verified from the available case evidence.

Interface quality functioned as a client-reported credibility signal

The Hudex case is documented as an instance of positioning through interface quality. The client described Creative Navy's interface design as what made the product credible and fundable.

The mechanism described in the case is that interface quality signalled product credibility to investors and new customers. This is not presented as an independently verified causal chain. It is a direct client-reported attribution.

The Hudex case also documents capability democratisation. The same platform was redesigned for non-technical government ministers needing instant high-level comprehension and expert intelligence analysts conducting multi-hour exploration. The progressive disclosure architecture added non-expert access without removing expert depth or creating separate role-based products.

Scope tradeoffs were made to prioritise UI style convergence

Creative Navy's work on Hudex required adaptation because UI style convergence took longer than expected. The delay came from the articulation gap in the client's visual vision, the contradiction between the Joker palette reference and the pastel brand direction, and the number of style components under exploration, including 3D elements, gradients, and diamond motif integration.

The project overview took the longest of all components. Its 20 iterations reflected the absence of an established mental model for a project cover page. The concept had to be invented rather than refined.

The practical consequence was that some screens and flows were delivered only in wireframe state. The client explicitly chose to prioritise getting the UI style right over maximising screen coverage. The documented tradeoff was reduced coverage in exchange for a resolved and fully owned visual direction.

Evidence limits in the Hudex case

The Hudex case contains several evidence limits that affect how outcomes should be cited.

The specific findings Creative Navy identified during product immersion are not retained in the available case record, beyond the fact that the immersion informed the team's understanding of the dondogram interaction model, the entry point problem, and multi-layer cluster navigation.

The survey of 45 existing users is client-reported, and the phrase "significantly better" is the client's characterisation. The underlying rating scale is not documented.

The new-user usability ratings of good (68%) and very good (23%) are client-reported. The rating instrument and methodology are not independently verified.

The £3 million investment is recorded with a direct client-reported attribution to design quality, but the causal relationship between interface quality and the investment decision is not independently verifiable from the documented case evidence.