Google AI - Kroger Partnership
Goal: provide feedback on Google's Gemini AI to provide the best item content for Kroger's items and test to see if Google could provide better item content than Kroger.
Outcome: succeeded in a comparable output of item content followed by provided recommendations for best practices of AI throughout Kroger's merchandising processes
Tools Used
The Story
~ learn the tech
~ introduce Kroger's item content & standards
~ evaluate the output along with analyzing merchandisers' reception to AI
~ provide Google Feedback
The Gallery
Definitions
Gemini AI
Google’s AI model that understands and processes through a variety of inputs. This is a scalable product that reasons inputs of texts, images, videos, and audio to serve the users’ purposes.
Regenerative AI
form of AI that learns, adapts, and improves outputs based off of human-input-prompting
Hallucinations
AI-generated incorrect or misleading outputs
Prompting
Human input questions and commands to achieve a generated output
Data-Driven Decisions
merchandisers become data experts and use data-based strategies to achieve successful business outcomes
Item Content
Item and product identity codes with attached Family Tree and attached attributes
My Personal Takeaways
AI is a touchy subject for a group of people who have essentially created a craft in their work. The consensus fear is job security, so I figured to lead conversations with the subject matter experts to instill trust and confidence that AI will work to enhance their data-driven decision making and not take their responsibilities away. This encouraged much more enthusiasm for AI. Additionally, I learned how to balance being Google’s client and the product designer.