Role
Product Designer
ColinnerData
2025
Scope
AI conversation UX
Desktop app
Marketing website
AI conversations, data management, and website
Activity, trends, and data oversight
An English-language application
ColinnerData is building WiseCore, a product that uses conversations with AI to help founders and executives capture their experience and the reasoning behind their decisions.
The founder, Jiang Keyue, brought me in to design an English-language experience for an international audience. My scope covered the Windows desktop app’s conversation and data-management interfaces, alongside the company homepage and product page.
The central challenge was to make an abstract concept tangible. Users needed to understand how to begin a conversation, what to contribute, and how to return to their saved sessions. The interface had to connect the immediate act of talking with AI to the longer-term purpose of building a personal knowledge collection.
Giving AI conversations a clear purpose.
The product’s premise went beyond an open-ended chat. The interface needed to guide users toward reflecting on their decisions and experiences, with clear session types, prompts, and ways to save their contributions.
Connecting conversations with stored knowledge.
Users needed to understand the relationship between a session and the record available afterward. Writing, saving, browsing, and reviewing had to feel like connected parts of the same experience.
Organizing a dense dashboard.
The requirements included trends, an activity calendar, distribution charts, completion tracking, authorization status, and tagged records. These needed a clear hierarchy so users could move between an overview and individual entries.
Supporting personal reflection and data oversight.
Conversations could involve sensitive professional and personal experiences. The design needed a thoughtful tone, visible information about data handling, and accessible record-level status.
I organized the experience around two connected activities: contributing knowledge through AI conversations and reviewing the resulting collection of saved sessions.
For the conversation interface, I focused on session entry points, prompts, response space, and draft and completion actions. I developed the conversation and review screens together so the relationship between a session and its saved record remained visible.
The visual direction combined editorial typography, restrained color, and generous space. This gave the conversation area a calm tone, while the dashboard used a denser layout for scanning and comparing information.
I also designed sample access within the conversation area to help first-time users understand the intended interaction before adding their own content.
The website extended this direction, introducing ColinnerData’s broader concept and showing how WiseCore translated it into a desktop application.
DECISION 1: Use typography to establish a reflective tone
I chose serif typography for key headings and conversation prompts to give the experience an editorial character.
Interface labels, controls, and dense information used simpler typography to support scanning. This combination gave personal reflection and data management distinct emphasis within a shared visual language.
DECISION 2: Organize the dashboard around levels of detail
The dashboard needed to accommodate six modules alongside access to individual records. I used grouping, spacing, and hierarchy to give each type of information a clear place.
Trends and activity provided an overview, supporting modules summarized the collection, and the table exposed record-level detail. Users could move from understanding their overall activity to finding a particular session.
DECISION 3: Make the first AI conversation easier to begin
The conversation area paired a specific prompt with space to respond and access to a sample. This gave users both a starting point and a way to understand the intended format.
Draft and completion actions made the session boundaries explicit. Users could leave an unfinished response or finish the session, while progress feedback distinguished an empty state from work in progress.
DECISION 4: Adapt the workspace to conversation and review
During a conversation, I prioritized the active prompt and response, with previous exchanges available above. Generous space and collapsible navigation helped keep attention on the session.
The review areas used denser layouts, filters, charts, and tables. Shared navigation, typography, and color connected the two modes while allowing each to support a different task.
DECISION 5: Surface data-handling information where users contribute
I placed a data-security message with access to further explanation within the conversation view, close to where users entered personal content.
The data-management views also included authorization labels and status summaries, including information about authorization for AI use. These elements made data handling visible during both contribution and review.
A. The AI conversation
A focused conversation interface covering the initial prompt, sample access, a response in progress, and previous exchanges. Draft and completion actions give users clear ways to pause or finish a session.
B. The data bank
A workspace for browsing and managing saved session records, with search, filters, date navigation, tags, and record-level status information.
C. Making it measurable
A dashboard bringing together activity trends, a calendar, summary metrics, authorization status, data distribution, and goal completion, with a table for reviewing individual records.
D. Modes and states
Light and dark interface treatments, empty and in-progress conversation states, and collapsible navigation, adapting the workspace while retaining familiar controls and structure.
E. The public story
A company homepage introducing ColinnerData’s vision and a product page explaining WiseCore’s AI conversation and knowledge-management experience. Both use a related visual direction to connect the application with its public presentation.
The project established an interface connecting AI conversations with the organization and review of saved knowledge. It covered the path from responding to a prompt to finding a previous session and inspecting its associated information.
The designs also created a consistent visual direction across the Windows application and the website, with layouts adapted to conversation, data review, and product communication.
What I delivered:
AI conversation interfaces with prompts, sample access, and session history
Empty, in-progress, draft, and completion states
A data bank with search, filters, tags, and record details
A dashboard with six modules and a record table
Data-status and authorization displays
Light and dark interface treatments
Collapsible navigation for the conversation workspace
Company homepage and WiseCore product-page designs
A visual direction combining editorial typography with structured data layouts
Jiang Keyue
CEO at ColinnerData
