Designing a Community @ IBM
10-week internship, cohort of 20 interns
My team comprised of 2 designers + 2 developers
Collaborated with IBM designers + product managers
Designed and pitched two experiences that re-imagine existing IBM Watson and IBM Cloud products
25 user interviews
10+ formal / informal presentations
Was asked to feature on IBM Design’s website
This summer, I had the opportunity to intern as a UX designer at IBM's flagship design studios. Coming from a startup of just two designers, it was a dramatic shift to be in a studio space that housed 400+ designers. I was inspired by the work around me, but also challenged to push my designs further.
Given real IBM business challenges, we used design thinking methods to imagine new possibilities for IBM Watson and IBM Cloud. We worked closely with IBM product teams to balance user needs and business needs. My team gave over ten presentations over the course of ten weeks, to align our stakeholders and communicate our designs and design process.
While IBM design thinking provided a framework of workshop-style activities, we spent a lot of time and energy designing our design process, as a small, nimble team working on large business challenges.
The problem involves superheroes…
Synthesis from 20+ interviews
We found that within IBM Cloud, there is a subset of expert users that are intrinsically motivated to help novice users, but lack the tools to do so effectively.
An example of these expert users is Jon (see below), who is an IBM retiree and has to-date answered over 800 questions on the IBM support forums.
In our research, we found that the missing piece for a lot of our clients was efficient channels of support and communication.
An opportunity to bridge two groups…
Has someone ever helped you troubleshoot an issue, and you thought - if only I could publish this conversation or share this knowledge in some way, it might help so many other people as well? What if useful conversations could be preserved?
Within IBM Cloud, we heard from a subset of users that helping people is intrinsically motivating, except it doesn’t scale. You help someone out on a Slack channel, but it doesn’t show up on a Google search for the next user to find; it isn’t archivable.
We identified within IBM Cloud an opportunity to bridge the gap between expert and novice users of the platform.
How to create a community? We diverged…
Live chat with a developer community (inspired by gamer chat applications)
Refer an expert to a specific question / topic or “pass the baton”
View an AI-generated answer that has been approved by expert users
Browse questions to curate areas of interest (inspired by Reddit)
Matchmaker between novices, experts
(None of these are good enough, keep ideating?)
…and came to one key flow.
What happened is that we talked to actual IBM expert users, showed them multiple ideas (sketched wireframes and digital), and found that they were excited about the notion of community, and that it was important that it:
Felt informal / like a chat
Had a reward / incentive system in place
Only sent relevant notifications
Forum vs. chat…
A challenge we faced was balancing the existing IBM forum-style format, with our new chat-style format. Originally, we had a public chat that could contain / branch into forum posts. We also briefly considered having forum posts contain their own chat rooms. We talked to one of our users who said “would it be possible to put all streams of communication on a single timeline?”
In the end, the product we created not only leveraged existing IBM products, but also made use of an existing user base of expert users who answer community questions in order to gain “reputation points”.
We use the concept of the expert-novice relationship, and created a tool to connect the two parties within IBM Cloud.
An optimistic outlook…
In testing with users of IBM Cloud, and iterating round after round on their feedback, we were able to put a product in front of them that they could see a future with.
Aligned the team, made sure all ideas were heard -
Visually, through whiteboard diagrams
In documentation, by keeping all our notes and files organized in Box
Conducted user interviews
Conducted user tests
Contextualized our work, in terms of -
Users - audited existing user research / personas / past attempts to solve the same problem
Market - looked at competitor products both within IBM and external to IBM, looked at sales wins / losses
Business - business model canvas + metrics
User Experience Design
Ideated design concepts that led to the final product
Created user flows
Used accessibility best practices (color contrast, readability, etc)
Created wireframes in Sketch to handoff to visual design
Created multiple iterations of presentation slide deck (content) and script
IBM Watson AI
In our research, we found that people had questions around Watson AI’s recommendations, which prevented emotional investment and trust. How could we create opportunities for more natural relationships with Watson AI? I explored a feedback loop for training the AI as well as an area to visibly view data sources which the AI draws from.