As the Product Owner, I led the research, strategy, and execution of AI-driven solutions to enhance productivity across a large-scale Agile Release Train. This included shaping the product vision, conducting user research, and delivering AI-powered tools that streamlined internal team and sales associate workflows.
My Role
I led the cross-functional Data Science & AI team consisting of one designer, two full-stack engineers, and two ML engineers. My role encompassed:
- Owning the Enterprise AI adoption roadmap
- Conducting user research to assess AI readiness across the organisation
- Managing delivery day-to-day of POCs and MVPs within a 300+ person Agile Release Train
- Partnering with external business stakeholders to fund and build AI applications
The Problem Space
Prior to the formation of the AI initiative, the retailer’s Agile Release Train (ART) faced several operational inefficiencies. Leadership identified the need to explore AI-driven solutions to enhance productivity across a 300+ member organisation, but lacked clear direction on execution. They set an ambitious goal – drive a 20% productivity uplift through AI-enabled workflows, but faced several roadblocks:
DATA SECURITY RISKS Employees unknowingly shared confidential data with AI tools, exposing the organisation to potential breaches.
INEFFECTIVE USE CASES AI was misapplied to tasks where it underperformed, leading to decreased efficiency instead of gains.
AI LITERACY GAPS Teams had varied levels of AI proficiency, leading to inconsistent adoption and underutilisation of the full potential of AI at work.
SOLUTION PROMPT
How might we develop a framework that guides AI adoption, ensuring teams use purpose-built applications effectively while fostering a culture of responsible and impactful AI integration?
The Solution Space
To create a stronger culture of using and implementing Enterprise AI across the client team, we:
- conducted two formative research campaigns for our Agile Companion MVP,
- adopted a collaborative and iterative approach to agile process automation, and
- established a dedicated team to offer AI enablers across the client business.
RESEARCHING AI READINESS
Launching a successful Agile Companion MVP relied on understanding where AI could fit into current workflows. We spent lots of time with individuals across the Agile Release Train to empathise with their daily activities, perceptions and experience of AI tools, and identify where a purpose-built language model could offer new opportunities for collaboration.
Our formative research phase was split into three core activities:
- Mental model and AI perception study – qualitative discussion to understand perceptions, past experiences, and concerns about AI in the workplace
- Unmoderated exploratory usability testing – observing natural user behaviours and affordance recognition within our UX prototypes
- Scenario-based usability testing – observing users complete realistic, goal-oriented tasks to measure efficiency, success rate, cognitive load, and trust in AI
AUGMENTING AGILE WORKFLOWS
With a clearer understanding of AI literacy gaps within the ART, we needed to ensure that our AI solutions were not only intuitive and informative but also supported a culture of adoption and collaboration. We approached this through three key strategies:
- Intuitive and informative design: Purpose-built workflows allowed new users to integrate AI into their processes seamlessly from when they first open the app.
- Internal marketing and engagement: We fostered a culture where AI adoption was community-driven, with team members actively sharing their use cases and learnings.
- Collaborative development: Through platform analytics, we could see prompt usage patterns and feedback, gaining valuable insights that informed future improvements like integrations with meetings, product documentation, and project tracking databases.
AUTOMATING BUSINESS PROCESSES
With leadership support, the new dedicated Data Science & AI team was able to continue funding our Enterprise AI roadmap and expand our reach across the business to new use cases.
We built a Large Order Automation service that could offer time savings and accuracy improvements for the sales associates who were previously processing these orders entirely manually. This offered more time for the associates to strengthen the core of our client’s business by offering thoughtful sales advice to their professional customers.
The Outcomes
1) LAUNCHED PUBLICIS SAPIENT’S FIRST ENTERPRISE GENAI CHAT
Our Agile Companion tool was the first of its kind sold by Publicis Sapient and deepened our client partnership by demonstrating practical GenAI applications for enterprise workflows. This project’s success also positioned us as trusted advisors in AI strategy and adoption.
2) SECURED FUNDING FOR OUR DEDICATED AI SERVICES TEAM
The success of the AI tools we developed led to a 4x increase in AI investment, securing leadership support for continued innovation through the establishment of the Data Science & AI team and commitment to new initiatives in the Enterprise AI roadmap.
3) MEASURABLE EFFICIENCY GAINS IN ORDER PROCESSING
Our AI adoption framework positioned the company for scalable, responsible AI integration across multiple business functions. This included our development of a Large Order Automation tool which reduced order processing time by 50%.
Want to know more?
If you have any questions about this project, reach out and let’s set up a call.
- Email me at: edwardfraser@berkeley.edu
- Connect on LinkedIn: linkedin.com/in/edward-fraser/