header-image-disco.jpg

Evolving Watson Discovery into a product for an everyday business user

How I led the design for the most ambitious release our product has had, collaborating with a team of 150 developers, 13 product managers, 9 designers and 2 design researchers – across 5 timezones, to reinvent a product initially built for a technical user.

My role

 

Lead Designer on the Watson Discovery v2.0 release, collaborating with a product team of 150 developers, 13 product managers, and 9 designers – across 5 timezones.

Timeline

 

10 months

Outcome highlights

  • New version of Watson Discovery was recognized as a Fast Company’s Innovation by Design Award Finalist for its user experience

  • Recognized with 2 IBM awards for excelling at collaborating at the leadership and squad level

  • Influenced product management and engineering leadership to shift their mindset to prioritize design, as proven by the fact that they made user experience be the measure of our success for the following year’s OKRs

 

Skill highlights

  • Design leadership - led collaborative design sessions that produced a user-centered, end-to-end flows and way of thinking that informed the well-rounded base of our product

  • Collaboration - cultivated an environment where design was tightly collaborating with product management, engineering and design research along every step of the way – from defining the vision, to sketching, to creating flowcharts, to writing epics to writing code so that we could remain aligned and take advantage of our unique perspectives

  • Problem solving + innovation - my design leadership and tactical design work transformed a historically technical user experience into a tool that an everyday business user could use, differentiating our product in the market

 

The challenge

 

The product team for IBM Watson Discovery took on a moonshot mission for our users. Our 3 high level goals of this work were:

  • evolve our broad portfolio to consolidate key functionality into a single product

  • bring in net-new AI capabilities

  • most importantly, revamp the user experience for an everyday business user rather than a technical expert

We had a handful of challenges to conquer along the way:

  • Challenge 1: AI for business users

    Revamping the workflow for the business user meant rethinking how machine learning models are built and evaluated, along with the metrics that should be laid out to measure "success" of the AI towards their business goals.

  • Challenge 2: Experiences, not features

    Since the team was combining established products, Design had to break people out of a feature-minded approach. Design needed to help identify an MVP that balanced this approach with table-stakes functionality from our existing products.

  • Challenge 3: Tight timeframe

    We had to move fast to deliver concrete value towards our vision within the year. We needed to make appropriate choices as a team, to balance technical and design debt incurred along the way such that it wouldn't lead to regressive experiences for any of our users, current or new.

  • Challenge 4: Team dynamics

    We were working across a huge team that was geographically and culturally distributed. We were tackling varying priorities across stakeholders coming into this effort, due to fragmented offerings & individual roadmaps in the portfolio

  • Challenge 5: Innovation

    We had to feed cutting-edge technology from research labs into the product in ways that would feel native to the experience.

Design process

 
 
design-process-workshop.jpg

Workshop to define goals

The project kicked off with a design-facilitated workshop, bringing together various teams in a single location. Together, representatives from engineering, product management and design generated user stories, empathy maps and journey maps. This approach helped us identify and align as a group on parallels in the product functionality and target use cases.

design-process-sketches.jpg

Collaborative visioning

Next, Design helped everyone step back from what's existing already, towards a vision of what's possible. Multiple rounds of visioning using sketches and storytelling helped us facilitate this. The first round was blue sky ideation for 4 days, to reset possibilities.

design-process-disco-concept.png

Produce concept

I took the ideas from the visioning sessions and boiled it down to something more realistic, and that told a story. It came to life in the form of a concept car (a design technique borrowed from the auto industry) that tells the story of how we can solve a customer’s problem, without getting hung up on technical constraints. With each iteration of the concept car, we shared with stakeholders and evaluated the concepts with users.

design-process-disco-epics.png

Partner on E2E epic creation

Design paired with product management and development to write all epics from a user-centered point of view, instead of falling back into the habit of creating feature-based epics. The squads were revamped to drive “end-to-end” objectives and deliverables. They each owned a piece of the user experience in its entirety which covers all layers of the product stack.

design-process-flowchart.jpg

Collaborative flowcharts

Next, I led the team in collaborative flowchart sessions, which were instrumental in shifting the team's mindset to think about user needs first. This helped us facilitate conversations around system behavior and constraints. And enabled us to establish common ground on a path for these high level concepts to come to reality.

design-process-MVE.jpg

Define an MVE

Our design team always pushes for an MVE (minimum valuable experience) rather than an MVP. With an MVE, the focus is shifted from features to the user’s experience. We sliced the story map in a thin, horizontal way that would enable a user to achieve their end goal, without going too deep in any one area.

design-process-disco-wireframes.png

Design wireframes

Now that we had the MVE flow laid our for our user, it was time to start breaking down the experience further to solve more specific interaction problems. More design team members joined us to jump in and iterate on bringing small pieces of the flow to life by building wireframes.

unsplash-image-DUmFLtMeAbQ.jpg

Validate user experience

We paired with design research to conduct a series of user research sessions with a cross-section of existing customers to validate new concepts for their needs.

Any issues or gaps identified through either of the above mentioned avenues were triaged together with other technical issues being sorted out for the release. This ensured that we not only had a documented design debt backlog for future, but also that user experience was given an equal weight and priority in the software development lifecycle on the team.

design-process-disco-hifi.png

Oversee final UI design

I worked with several designers to undertake the massive effort to design the visual and motion design needed for a complicated and UI-heavy product. At this time, IBM was releasing a revamped design language that we implemented into the new Watson Discovery. We were the first product to GA in the IBM Data and AI division utilizing the new design language, becoming an example for the rest of the products to follow.

Results

 

The evolved Watson Discovery product GA-ed by the end of the year. We successfully accomplished the three overarching user outcomes that we outlined when we kicked off the project.

Our success was largely due to the high degree of trust within each squad as well as across squads. Each squad feels comfortable speaking for the other, and challenging each other to go to the next level. The team is invested in our new squad structure – actively strengthening our squad relationships and cross collaboration by retrospecting on the past 10 months of work and spending time on tangible action items.

Next
Next

Enabling non-technical users to visually train Watson on their data