UX Beers - A Story about product management at uman.ai - Jasper Verplanken
Sep. 2, 2019•0 likes
0 likes
Be the first to like this
Show More
•214 views
views
Total views
0
On Slideshare
0
From embeds
0
Number of embeds
0
Download to read offline
Report
Design
After a consulting and agency career as a UX architect, Jasper is now involved in uman.ai, an AI startup involved in corporate learning and competency management. He focuses on building the product and the product marketing that comes with it.
Hi,
I’m Jasper
‣ Civil Engineering - Computer Science
@Ghent University
‣ Started in Information Management
Consulting
‣ Earlier: digital strategist and UX
architect @Wijs
‣ Co-founder and Head of Product
@uman.aijasper@uman.ai
www.uman.ai
The actual scenario
Knowledge Sharing Reality
Interest: “Why would I read this, it
doesn’t seem interesting to me”❌
❌
Time: “I don’t have time to read this right now”
The actual scenario
Knowledge Sharing Reality
Interest: “Why would I read this, it
doesn’t seem interesting to me”❌
❌
Time: “I don’t have time to read this right now”
Relevance: “This is too heavy for me
now, it might be relevant later though.”
❌
The actual scenario
Knowledge Sharing Reality
Interest: “Why would I read this, it
doesn’t seem interesting to me”❌
❌
Time: “I don’t have time to read this right now”
Relevance: “This is too heavy for me
now, it might be relevant later though.”
❌
Transparency: “I don’t know who and
what can help me with this subject.”
❓
❓
❓
Where we want to bring you
1. Where & when you need it In the flow of work
2. To your skills and level Personalised & dynamic
3. Better project matching Real-time talent insights
Connected knowledge automated for your teams
AI layer
Knowledge
sharing platform
Talent insights
module
In The Flow
of Work
Intelligent knowledge sharing, in the flow of work
Knowledge Efficiency
Knowledge Retention
Competency Mapping
knowledge,
content &
resources
UX at an early-stage startup
It’s not pretty.
Unlearn “methodologies”
Unlearn “processes”
Unlearn “best practices”
Setting new standards
There’s very cost-sensitive trade-off between efficiency and flexibility
Design Systems
Hi-Fi
Prototyping
Fine-grained
product backlog
Adjusting standards
There’s very cost-sensitive trade-off between efficiency and flexibility
Design Systems
Hi-Fi
Prototyping
Fine-grained
product backlog
The experience is highly dynamic
How can you prototype a highly dynamic experience?
Prototyping dynamic experiences with static designs is a huge investment.
How we work today
Identify needs Qualify need Draft UX flow
Develop first
version
Test it in early
adopter program
PRODUCT STRATEGY EARLY ENGINEERING
The trade-off
Design & test
static designs
Develop
dynamic
experience
Draft static
experience
Develop & test
dynamic
experience
Testing anticipatory design
The main challenge is that we’re testing anticipatory cues:
▸ Suggestions for new learning resources
▸ Suggesting tags for learnings
▸ Showing a notification “this might be interesting” when visiting a relevant URL
▸ etc.
How to test anticipatory experiences
Step 1
Based on:
● Earlier user feedback
● Usage data
● Strategic product decisions
Define an assumption and draft a UX flow that may validate this.
How to test anticipatory experiences
Step 2
Build it, so that it ‘kinda works’
Example:
● Start with a very easy to implement algorithm, rather than a fully fledged graph-based recommender
system
Most of all, reuse as much components as possible.
This is crucial to allow fast iteration in development
How to test anticipatory experiences
Step 3
Be mindful of the ‘type of experience feedback’
● Case #1 - the proposed solution isn’t adequate
● The suggestions aren’t that personal to me → Iterate on the algorithm
● Case #2 - there’s no need fulfillment
● “I don’t get why I need suggestions, I know what I want” → Iterate on the flow
Example
“I want to fill this in before I
complete an article.”
“The skill I want isn’t in there.”
“I want to edit the skills before
and after I have completed it.”
How we prototype for AI is different
(even as a designtool-nerd)
It became clear that static design tools are lacking for AI products
Conclusion
Spend more time upstream, stay flexible downstream
Try iterating faster during engineering, spending less time in
high-fidelity design tools
1
Try iterating faster during engineering, spending less time in
high-fidelity design tools
Just like wireframes in UX, you have similar ‘early prototyping
methods’ during development
1
2
Try iterating faster during engineering, spending less time in
high-fidelity design tools
Just like wireframes in UX, you have similar ‘early prototyping
methods’ during development
Experiment with design thinking techniques to identify actual
needs as a solution validation method
1
2
3
Try iterating faster during engineering, spending less time in
high-fidelity design tools
Just like wireframes in UX, you have similar ‘early prototyping
methods’ during development
Experiment with design thinking techniques to identify actual
needs as a solution validation method
1
2
3
talk to one of our experts
Thank you
Charles Boutens
charles@uman.ai
Thomas Verschuere
thomas@uman.ai
Jasper Verplanken
jasper@uman.ai
AI layer
Knowledge
sharing platform
Talent insights
module
In The Flow
of Work
Intelligent knowledge sharing, in the flow of work
Knowledge Efficiency
Knowledge Retention
Competency Mapping
knowledge,
content &
resources
In the flow of work
● Your knowledge communicates w/ you
● You push knowledge to uman.ai
🌍Learn and share while working
Hyper intelligent
● Attaches skills to knowledge resource
● Measures expertise level of knowledge
resource
🧠Knows relation between +10.000 skills
As approved by
Real time talent insights
📊Real time competency data
● Uncover employee and team skill
profiles
● Match innovative projects with
people’s growth paths
As approved by