Recently I had a chance to lead a team at the IBM WatsonX Hackathon and I must say it was a great experience. My team chose track 2 and we built a total of 8 intelligent assistants using prompt engineering techniques.
The use case was about increasing profitability for a fictitious retail client that we customized to match a real client scenario.
Here are the brief steps we took.
1. Customized the case study to match a real client (without using client confidential or sensitive data)
2. Extracted the client pain points
3. Mapped the pain points to KPIs to better monitor and manage them
4. Identified intelligent assistants that could be created
5. Created and tested those assistants
6. Identified ROI with respect to improvement in cost and time savings vis. a vis. manual effort
As a final step we created a stand up pitch to showcase what we had done.
While I do not expect to win anything as there were thousands of great solutions submitted by IBMers around the globe, the whole experience led to a lot of learning and confidence building.
As employees, we got first hand experience working with LLMs.
No comments:
Post a Comment