sourcing automation 2.0

keelvar / product desIgn lead

Sourcing Automation is a no-code workflow configurator that enables users to define their organisation’s sourcing and supply chain workflows, then instruct Keelvar’s highly customisable intelligent software Bots to automate the workflow, eliminating repetitive tasks from users’ day-to-day processes.

This case study shows Sourcing Automation’s evolution from version 1 to version 2.

An exploded view of Sourcing Automation user interfaces.

the team

chris walsh
Product manager
nathan dolan
Lead product designer
karen crowley & glen mitchell
Product designers
kevin dunleavy
Technical lead
gwen burchell
Engineering manager
+6
Backend, Frontend and QA engineers

discover

problem statement

Sourcing Automation 1.0 has been designed to automate a very specific process for a niche group of users. This poses problems for onboarding new customers as the product is not flexible enough to cater for other use cases. It can take up to 12 weeks for new customers to be onboarded and existing users are not satisfied with existing functionality, usability, and the rate that administrative tasks are completed at. These issues are leading to low levels of adoption and a struggle to accommodate new customers use cases.

discovery goal

Decide how Sourcing Automation can be improved so that more customers can be onboarded and have a better overall experience.

discovery objectives

  • Full audit of existing Sourcing Automation from a UX and Engineering standpoint
  • In depth knowledge of the current onboarding process
  • Full review of customer support tickets
  • Understanding of our potential customers' tools
  • Breakdown of potential customers sourcing processes

customer interviews & review of existing customer support tickets

  • Reasons for low adoption of Sourcing Automation
  • Identify key usability issues
  • Understand customers varying sourcing processes
  • Learn about the other systems or tools used for sourcing

user journey maps

User onboarding is a process that needed investigation. Through interviews with internal team members and recently onboarded customers we were able to create a user journey map and identify areas of the process that cause issues or time delays.

review of existing userflows

By reviewing the userflows of Sourcing Automation version 1, and gathering data from user interviews and customer support tickets, we could identify pain points and communicate the need to fix these issues to the wider company.

discovery outcomes

how might we decrease the onboarding time for new customers?

We believe that if we enable customers to onboard without the input of Keelvar employees that we can onboard customers more quickly. We know this to be true when we see a significant increase in the number of customers onboarded over a 6-month period.

how might we allow sourcing automation to automate more diverse sourcing processes?

We believe that if we allow users to define their own sourcing process within our tool that we can remove the engineering effort in onboarding new customers and allow for the automation of more spend categories. We will know this to be true when we see users setting up Sourcing Automation workflows for new categories of automations without the input of the Keelvar team.

how might we improve the overall user experience and enable quicker improvements in the future?

We believe that by redesigning and rebuilding the Sourcing Automation user interface, we can improve the overall user experience without being limited by the existing Sourcing Automation technical debt and user flows. We will know this to be true when existing users report an increase in satisfaction with usability.

defIne

user personas

  • Created based on information gathered from user interviews
  • Feedback from our customer success team
  • Continuously updated as we interacted with more users / potential users

user stories

  • Written in collaboration with the product manager and engineering
  • Groomed with engineering team.
  • “As <persona>, I want <what?> so that <why?>.”
  • All user stories had Acceptance Criteria
  • Used to scope engineering work, plan project timelines and individual sprints.

project planning

The project was split into two major milestones.

1. alpha

Release a working product to production in order to allow for the first group of preselected users to be onboarded. Gather feedback and make continuous improvements.

  • Create a request (single item)
  • Monitor request progress
  • Execute workflow
  • Bot creation
  • Email notifications

2. general availability

Full release of Sourcing Automation 2.0. This will allow for fully self service bot creation and advanced configuration, full administrative settings and notifications available along with more advanced bot capabilities.

  • Create a request (single item)
  • Create a request (multiple items)
  • Monitor request progress
  • Execute workflow
  • Bot creation
  • Email notifications
  • In app notifications
  • User management
  • Edit bot details
  • Edit bot workflow
  • Edit request form

Develop

user flows

The following user flows were prioritised and identified as being essential to design and develop before users could be onboarded.

1. Bot creation
2. Create a request
3. Monitor request progress and execute workflow

The user flows were created based on previous discovery work and with input from various stakeholders and members of different teams.

1. bot creation
2. create a request
3. monitor request progress and execute workflow

wireframes

Visual representations were needed in order to establish alignment within the project team and to gather feedback from existing Sourcing Automation customers and potential customers.

1. bot creation
2. create a request
3. monitor request progress and execute workflow

high fidelity designs

Through continuous iterative design and gathering feedback from the necessary parties, final designs were proposed. These included the essential features planned for the Alpha release, along with many considered essential for General Availability.

These designs would be used as the “end goal”, it was up to the product and engineering teams to work together to refine the designs in order deliver a functional product for Alpha.

1. bot creation
2. create a request
3. monitor request progress and execute workflow

clickable prototypes

This clickable prototype was used to carry out usability testing, gather internal feedback and communicate designs to the development team.

This is one of many prototypes created using Figma.

Deliver

development
  • 2 week sprints following Agile methodology
  • Members of Engineering, Product, Design and QA participated in each sprint
  • Work tracked in Jira
  • UI implementation review with design team
  • Refinement of scope throughout development process
  • Regular project team sync up calls
  • Planning, grooming, retrospective calls for each sprint
  • Company wide update calls and demoes of progress when suitable
alpha release

Sourcing Automation 2.0 Alpha was deployed to production on  5th August 2022

4 customers were selected to be onboarded in the following weeks

These customers were selected based on the complexity of the automation workflows and spend categories

alpha customers onboarded

outcome

how might we decrease the onboarding time for new customers?

We believe that if we enable customers to onboard without the input of Keelvar employees that we can onboard customers more quickly. We know this to be true when we see a significant increase in the number of customers onboarded over a 6-month period.
12 weeks to 1 day
  • Redesigned and developed a new end-to-end onboarding flow
  • Highly configurable automation workflow creator
  • Used by Keelvar employees to configure workflows on behalf of customers
  • To be released to end users in coming months

how might we allow sourcing automation to automate more diverse sourcing processes?

We believe that if we allow users to define their own sourcing process within our tool that we can remove the engineering effort in onboarding new customers and allow for the automation of more spend categories. We will know this to be true when we see users setting up Sourcing Automation workflows for new categories of automations without the input of the Keelvar team.
new workflow configurator enables automation of new categories 
  • Nestlé automate their packaging sourcing process
  • Coca-Cola automate a new type of ocean transport sourcing process
  • These sourcing processes could not be automated using Sourcing Automation 1.0
how might we improve the overall user experience and enable quicker improvements in the future?
We believe that by redesigning and rebuilding the Sourcing Automation user interface, we can improve the overall user experience without being limited by the existing Sourcing Automation technical debt and user flows. We will know this to be true when existing users report an increase in satisfaction with usability.
tackled high priority improvement requests and validated improvements through usability testing
  • Search and filtering of requests
  • Improved transparency of automation process
  • UI consistent with style guide
  • Top level navigation
  • Settings and Admin Settings pages
  • Removed “chatbot window” UI element and it’s related UX issues
An exploded view of Sourcing Automation user interfaces.

conclusion

next steps

Alpha release is the first major release milestone. Work must continue in order to provide end users with necessary feature additions, usability improvements and a completely self service platform.

Usability testing, usage analytics, customer support requests and feedback from customers during onboarding and other interactions with existing and potential customers should be utilised in order to inform design and development efforts moving forward.

what did i learn?

  • The value of a well executed Discovery phase in order to de-risk a project
  • The importance of communication and collaboration between teams
  • How to design and aid engineering efforts in a fast paced software development process
  • The value of product managers and how to collaborate with them
  • How to lead a design team, plan design work, mentor and run design critiques effectively
  • How to set up and run usability testing and compile findings in order to inform design decisions

what would i do differently?

  • Push for inclusion of analytics tool in Sourcing Automation 1.0
  • Set quantitative outcome goals to measure the success of Sourcing Automation 2.0
contact me
tell me about your project
email Nathan