Project Overview

Title: Gen AI ERP Service

Project Type: Discovery Sprint

My Role: UX designer

Duration: 3 Months (2024)

Objective: Ideating and proposing how to inject AI into an Enterprise Resource Planning (ERP) system to enhance user efficiency and productivity

Aims

Context: An existing ERP system can manage many business process and is used among a range of different industry players. This versatile system is often used to distribute business communication, or to facilitate conferences with hundreds of attendees.

Target Audience: Business admins who manage communication distribution.

Key Goals: To identify areas that could be enhanced through the use if generative AI (GenAI) to provide a better user experience and improved output. To them produce and end to end prototype of the identified area with the AI enhancements ready for build.

Discovery

• This project was a 4-6 week intensive discovery. This project was kicked off by the research stage which we all supported on.
• I specifically looked into AI design how competitors have facilitated AI to human interaction. This was particularly interesting as I hadn’t design for AI before so I was keen to see trends in this field. Lots of brands take the approach of ‘chat bot’ style interaction with conversational style AI.
• After the initial research we moved onto ideating a direction/flow to focus on.
• We concluded narrowing down on content generation and comms distribution would be the most valuable areas to continue with.
• Once this was agreed with the client I began rounds of design iteration with user testing running parallel.
• I started creating some user journeys, mapping the AI supported flow through the comms generation and delivery process.
• This lead straight into building basic wireframes so get some early visuals.
• I then began many rounds of design iterations, making adaptations to the AI interaction, adding personification and personality, consideration over tone and communication style etc.
• We finally get to a place were we had incorporated all the necessary features and pleased the client with the design approach, concluding our discovery sprint

Personified AI Assistant

I created this animation to represent the AI assistant with a combination of HTML, CSS and Javascript.

I used Anime.js to help create the animations so bring the character to life.

After testing with some users the general feedback was much more positive once the character had been introduced. Users made comments such as “approachable”, “cute”, “eager helper but not too over the top”

It seemed by adding an identity to the AI users where much more willing to engage and trust it.

Deliverable and outcome

• Our team manage to produce an extensive amount of work if this short time. And collated that all into a powerpoint which was presented and the circulated after.
• The intention was for the client to take this to their stakeholder for approval to continue the work.
• I also handed over the figma file with all the iterations and final prototype, which would inform developers in the next stage of the project.
• Due to this project having limited initial design resources I felt it was appropriate to build out a basic visual guide to inform any future work, particularly around the AI which we had given an identity and personality that would have to be consistent going forward.
• Ultimately this project was very successful and there are plans to continue with both building on the work thats complete and looking into other potential areas to incorporate GenAI.

Client Feedback

“Hey team, representing your work today was an absolute pleasure. We collectively amazed the local executives !! We had great discussions as to where we could take this innovation. I have worked with amazing teams over so many years, your innovation, collaboration and passion is second to none !! Thank you very much !!”

– Key Senior Stakeholder

Skill call-out

• Proficient Figma Skill
• Visual design
• End to end design process
• Iterative design
• Design for AI

Reflection

• This was a really interesting project to be apart of. It was the first time I designed for AI and it was more complex that expected. It was interesting discovering all the extra factors AI introduces, such as the extent of possibilities around how a user might interact with it, assumptions around what its capabilities are in different contexts. More specifically what level of justification does the AI need to provide? How can we catch error scenarios when the AI is incorrect?
• This project was also a great experience in having owner of my work and having junior support. I was able to dictate my design process and delegate tasks to effectively progress.
• As this project progresses into its development stage, I will be supporting on the development of AI, ensuring it is consistant with the personable establish.