Built for advisory work where context and confidentiality matter.
Fergus is FGS Global’s AI platform, tailored to the needs of consultants advising businesses on complex, high-stakes situations. It helps them work with quality data while keeping sensitive client information confidential—a critical requirement when supporting major companies and startups through acquisitions, IPOs and other consequential events.
I worked within an international, multidisciplinary team, collaborating closely with data scientists and engineering. I owned the product design for Daily Brief, from the initial concept and user research to configuration, information hierarchy and the briefing experience.
To respect confidentiality, this case study uses reconstructed wireframes and fictional content rather than production screenshots. They explain the design decisions without exposing client information or proprietary interface details.
An early prototype with a live data point opened a wider product effort.
What does a consultant need to understand quickly about a client?
Many variables and dependencies, guided by an agent and visible to the user.
An early prototype opened a bigger question.
Daily Brief began at a team hackathon. I developed the concept with a data scientist and a backend engineer, and within six hours we had an early prototype connected to its first live data point.
That first test opened a broader product effort. After leadership adopted the idea, we interviewed consultants to understand what a useful brief should help them see, and how it should fit into their client work.
What matters at a glance?
The media monitoring experience needed to offer a quick, current overview of coverage and other developments relevant to a consultant’s client. In interviews, we explored the kinds of information that could matter, from competitors and market activity to prediction markets and publications. The answer depended on the client and the purpose of the work.
That made setup a product problem. A useful overview needed a way to capture each workspace’s context, while allowing people to adjust it as their work changed.
A long configuration process needed more than a form.
The setup involved many variables and dependencies. It had to flex across different clients and mandates without making users work through a rigid sequence of fields. At the same time, a conversation alone could make it hard to see what the system had understood.
I designed a configuration panel paired with an AI agent. The panel keeps the developing scope visible and editable. Alongside it, Fergus guides the user, points out missing information and suggests ways to refine the configuration.
A scope the user can inspect
A conversation that moves the setup forward
The agent helps interpret intent, notice gaps and propose refinements.
Review the scope, edit a setting directly, or continue the conversation.
Start with intent
The consultant describes the client and what the team needs to follow.
See the interpretation
The panel makes the evolving configuration visible, including gaps.
Adjust the scope
The user can edit settings directly or work through a suggestion with the agent.
The agent reduces the effort of getting started. The visible panel makes its interpretation understandable and correctable.
Turn a collection of information into a brief worth reading.
Once the scope was defined, the next question was how to help a consultant read the result. A summary, a lead story and distinct formats for different sources give the information a hierarchy and make its meaning easier to assess.
At a glance
A shift in the sector
Sources →Choosing a lead story changed what the system had to select.
I gave one story prominence at the top of the coverage. That decision went beyond layout: it led to an architectural change in how the data was selected, so the brief could identify a lead story.
The overview helps the reader orient themselves; the lead story establishes priority. Source access stays close to the content so they can inspect the material behind it.
Different evidence needs different reading models.
I used distinct card formats to differentiate source types. A press article, a prediction market and a social post carry different kinds of information. Their structure should help the reader recognise that difference before interpreting the content.
A new sector development
Will the sector reach the projected milestone?
Public discussion
Platform · Time · Original postIllustrative card structures with fictional content and values. The social card is a conceptual reconstruction of the source-type distinction, rather than a reproduction of a production component.
Preparing your brief
Gathering and organising the latest information.Make the wait understandable.
The briefing experience also needs to communicate what is happening before the content arrives. A visible status and progress indicator explain that the brief is being prepared, while placeholders preview the structure of the result.
From an experiment to a released feature.
Daily Brief was released to relevant teams within FGS Global. The work grew from an early prototype through user interviews, product definition and the design of a setup that could adapt to different client work.
The work connected product definition, interaction design and the system behind the interface: how the brief was configured, what information was selected and how different signals were presented. My focus was to make those choices understandable and keep consultants able to inspect and refine the result.