AI Application
Microsoft & Chat GPT Application
How Microsoft might offer AI Assistance to its customers ?
When Microsoft made a major investment in Open AI, it had a plan to integrate the LLM across its platform and offer it to its customers. Microsoft had worked with Indigo Slate before on a number of projects for its cutomers including product concepts, envisioning, strategy and my team, HXI was for the most part in charge of such projects.
This time the ask was to dive deep into design-research and explore applications of Chat-GPT across various markets Microsoft serves including: Healthcare, Manufacturing, Entertainment, Retail, etc.
Initial (hot) market: Healthcare
In this year-long project, I worked with Microsoft to develop AI use cases and applications for its customers. The initial market was the Healthcare system, where the need was more urgent and palpable.
This case concerns Optum (United Health Care) and how the application of an AI assistant can help Optum better serve its customers.
Coordinating a collaborative effort
HXI Studio was in charge of the project. Due to the new-ness nature of the project, we formed a collaboration between my team, Microsoft and SMEs. The project involved a tight cadence of meetings and workshops. From my team, I had two Sr. Strategist in addition to collaborating with the Creative, and Technical Writing and Editing (TWE).
Microsoft
Client+Collaborator
Sr. Directors, Directors and Experts from MSFT
HXI
My team
I also invited the Creative team to work with us on this project
SMEs
Expert consultants
We consulted with AI experts to explore AI use cases for customer service
Summary of use case discoveries
Entry recogniution
Intelligent routing
Intelligent routing
Member authetication
Biometric authentication
Automated call classification
Caller recognition
Optum Health Customer Success Reimagined
Microsoft's goal was to offer a new solution to Optum. And for that, Optum needed to see clear value in the application of the AI. To reimagine the customer success experience across the Optum platform, we needed to have a deep dive into the current process and learn about how things work. What works? And what does not work.
Designing a new research process
To maximize potential outcomes, I designed the process so that each of the three teams would offer the result of their brainstorming independently from each other. This ensured that we captured everybody's ideas before the teams started collaborating, before teams meeting and start influencing others' thoughts in the collaboration process.
Then, the exchange of ideas started. We had a wealth of application use cases that could have been applied in a number of contexts and situations.
Pre-Call
AI Assist routes the member to the right advocate with all the summary and action items
Call
AI Assist keep generating guidelines for the advocate and navigates them through the conversation
Post-call
AI Assist provides feedback to the advocate and generates new knowledge for the system
Kep touchpoints
We started by understanding customers' current state experience when they need help and contact Optum, learning about their pain points and celebration points. For this, we held workshops with all stakeholders to learn about the journey map.
Contextualizing the research: Optum
We discovered the key touchpoints that make the current experience stressful and perceived as negative throughout the process. In the case of the customer (the member), Katherin, a new mom, there were a few touchpoints that were certainly low-point experiences that could have been enhanced by an AI Assistant.
AI as Adocate Asssitant
Various use case scenarios: we build a wide number of scenarios that required AI assistance, and some of the key findings included entity recognition. For example, understanding "benfeits" and relating that to the membership benefits.
Below shows some key moments where the AI Advocat Assist genrates guidance for Celest. Key functions include real-time transcription handoff, entity recogition and business process automation
AI Assisant
A major part of our research on figuring out possible sceanrios that AI Assist could help the process, was during the call. We found out that the call expereice is the most intense, brain power consuming and stressful phase of the entire journey, both for the member and for the advocate.
We came up with some subcategories that the LLM model could assist this part of the journey. This included ready-to-act items in the dashboard, live collaboration, and providing insights for the next steps, providing visibility to the advocate to claim visibility.
Persona Research & Journey Map
At the company level, I worked with our Executive Directors, Founder, and Account Directors on proposals and meetings with potential clients. At the project level, the HXI Studio regularly partnered with BizTech, Creative, and Motion teams. I led HXI during project conception, resource planning, and collaboration with these cross-functional teams.
I also collaborated closely with Indigo Slate’s Executive Directors, Founder, and Zensar executives on ongoing projects and new business opportunities. I played a key role in securing new business globally for both Zensar and Indigo Slate.
Empowered Advocates
Reduced Call Volume
Faster Time to Resolution
The end result… :)
I also collaborated closely with Indigo Slate’s Executive Directors, Founder, and Zensar executives on ongoing projects and new business opportunities. I played a key role in securing new business globally for both Zensar and Indigo Slate.