Overview
Looking at the problems our users faced, the business was looking to solve their frustrations by speeding up the creation process using data linked to their workbooks as well as suggesting "best-of-business" charts and workbooks without the need to create multiple draft versions and pull in data manually.
Looking at customer feedback, over 60% of ivusers found this to be a real problem to get started linking their data to visualisations as well as having to create multiple rough drafts and wasting many hours and sometimes days. They were wanting an easier way to get started and this helped solve that for them.
Looking at customer feedback, over 60% of ivusers found this to be a real problem to get started linking their data to visualisations as well as having to create multiple rough drafts and wasting many hours and sometimes days. They were wanting an easier way to get started and this helped solve that for them.
Pains
• Visualisations are often fiddly to create and users create multiple draft versions of them before settling on a final version because they need to manually create and link this to their data in the product
• It takes a lot of time creating drafts of sheets and visualisations
Gains
• To speed up the creation process
• To use the power of AI to suggest appropriate visualisations based on data
As a busy Business Intelligence Expert, Joe needs recommendations on suggested dashboards/sheets/visualisations based on his data so that he can build his apps faster and deliver this to companies he works with.
Ideation
Looking at best-in-class experiences, combined with competitor research, I ideated with my cross-functional team including a product manager and engineers to design the best solution for implementation and to validate viability.
Design
I tested the low-fidelity prototypes with 6 of our users to refine the flows and to find any areas for improvement. I then iterated on the lo-fi prototypes and took the preferred flows across to high-fidelity prototypes and presented that to senior stakeholders along with my findings and feedback from users. I also presented to my product design department to test that the existing design patterns adhered to our design system and documentation.
"I find this very intuitive and would love for this to be ready for our customers,"
Randy,Customer Support
Randy,Customer Support
"This makes sense to me and I would find this helpful in speeding up my process without having to create so many drafts," Stephanie, Customer
Scenario 1: Dashboard generator
Once a user had uploaded their linked their data to Astrato, they could create a new workbook using the Astrato Insights selection from the wizard.
View the prototype below.
View the prototype below.
Scenario 2: Generated AI visualisation
Once a user had uploaded their linked their data to Astrato and created a new workbook, they could add a new visualisation using the AI visualiser from the toolbar.
View the prototype below.
View the prototype below.
Findings and next steps
This proof of concept was presented at a Data Summit towards the end of 2023.
At that time, my contract had come to an end so I was not part of any further implementation or research on this. The feature has since been shipped and there has been lots of positive commentary on this on LinkedIn showcasing the feature.
At that time, my contract had come to an end so I was not part of any further implementation or research on this. The feature has since been shipped and there has been lots of positive commentary on this on LinkedIn showcasing the feature.
I would have suggested testing the high-fidelity prototypes once more with external users on platforms such as UserTesting.com and refining this with their feedback before handing over to my development team.