OakminerAI
AI, LLM, Conversational, RAG
OakminerAI (now oakie.ai)
Product: AI augmented insights platform
Role: Head of Design
Scope: Web platform, AI chat.
Skills: Design strategy, Conversational design, Competitive research, AI platform, LLM, Wireframing, Prototyping, User research, 0-1 product.
oakie.ai
“An early-stage startup revolutionizing the way organizations leverage the expertise of their most valuable employees. ”
Unlocking Expertise with Oakminer.AI
Expert knowledge is gold in specialized fields like law, portfolio management, and investing. These seasoned professionals have spent years building their skills and knowledge - but they can only share their expertise with so many people at once.
OakminerAI solves this challenge by capturing how these experts research and make decisions. By documenting these processes, the platform helps automate workflows and train employees, effectively sharing specialized knowledge across your entire organization.
Research
As AI/LLMs are an emerging technology, and we built the product from the ground up, not many design resources were available. My preliminary research was done in several vectors:
Competitive research
Ongoing discovery of the different branding and usability strategies going into LLM/AI and other chat interfaces. Including ChatGPT, Claude, Gemini, Copilot, SlackBot, Notion, Adobe ai features, Intercom, and other specialized AI tools.
I’ve collected some of my insights into a short review, published on Bootcamp: Contemporary usability challenges for Chat & LLM-powered applications
Technological literacy
Theoretical research (papers) on AI technology and hands on practice. Using workflows and data sets on ChatGPT etc, as well as iterating with OakAI platform to test the strength and weaknesses of the technology in each case.
Identifying Use Cases
Another aspect of the research and preparation was becoming familiar with our potential users. This was done via user interviews and flow modeling in Figjam. Our early focus was on Investment strategists, Contract and negotiation attorneys, and Policy advocates. We had done deep interviews with professionals who shared documentation and processes with us. We also had them use available LLM products to assess their technical knowledge and LLM skills, which had major impact on the UX of the product.
User workflow and script example
Wireframes & Mockups
“Mockups turn vague ideas into clear, actionable goals. ”
I have used the knowledge gained via interviews and research, user flows and scripts into wireframes. The styling of the wireframes suited the audience and the purpose of the discussion. For example, developers would sometimes get overwhelmed by too much details, but, investors wanted to see a market-ready experience with high level of details.
Wireframes in varius fidelities
Designing for Conversational UX
“ Conversational UX design focuses on creating intuitive and engaging dialogues between humans and machines. It combines UX principles, linguistics, and AI technologies to craft interactions that feel humane and natural.”
Why use a conversational interface?
Currently, LLMs primarily use a chat interface for user interaction. Exploring the possibilities of LLM UI opens up many considerations and expectations, both for users and the product team. Our observations suggested that we could use chat as a primary mode of interaction, but we must preserves some principles in order to mitigate potential usability challenges.
Guidance & Context
We wanted the UX be self-explanatory, and have the LLM push for progress and engagement. This also helps the system to learn more about the user’s preferences.
Clarity & Confidence
LLMs are a new technology, and so, it lacks the credibility of older, more seasoned platforms. To gain the trust of our users, we developed a few strategies to help users test and verify the results. Here is one example:
Persona & Tone of Voice
We had to balance between the AI’s roles as both a student and a teacher, depending on the user. We framed the agent’s tone as Friendly, Helpful and Trustworthy.
Fallback Strategies
We needed to create flows that addressed potential technology challenges. Issues like lag, hallucinations, and inaccuracies needed to be resolved with the help of UX.
The Save, Re-generate, Copy, and Feedback tools serve a dual purpose: they enhance user functionality and, with tracking, allow the product team to assess the AI's performance from the user’s perspective.
Putting it all together
Now that I had the knowledge I needed and strategies to make the UX better, I set out to connect the scenes into an end-to-end workflow.
Basic Flow Prototype
Components library
It may not seem useful to “bake in” UI components in a product that changes daily, but, working with components made my work much easier. Setting up a style library based on Atomic Design concepts and evolving it to create increasingly complex components, helped the engineers feel more grounded when developing the UI. Not all states made it in in earlier iteration, but the growing library help the developers and myself move quickly and test internally and externally the validity of new UX ideas.
Outcomes
Research and user interviews enabled us to plan our product with accurately. Prototyping and a flexible components system enabled us to iterate very quickly. In the course of 6 months, we built the live, front-end experience of Oakie AI. We were able to onboard early clients, and received permission to track their interactions to improve the platform. This was a unique and impressive achievement to a tech-heavy startup company still in the seed round.