PDI Technologies’ Steve Antonakakis shares how his company is implementing generative AI across their fuel and retail technology ecosystem through a practical, customer-focused approach using Amazon Bedrock.
Topics Include:
- PDI's COO/CTO discussing generative AI implementation
- Practical step-by-step approach to AI integration
- Testing in real-world settings with customer feedback
- AWS Bedrock and Nova models exceeded expectations
- Early adoption phase with huge potential
- Fuel/retail industry processes many in-person transactions
- PDI began in 1983 as ERP provider
- Grew through 33+ acquisitions
- Provides end-to-end fuel industry solutions
- Owns GasBuddy and Shell Fuel Rewards
- Processes millions of transactions daily
- Generative AI fits into their intelligence plane architecture
- AWS Bedrock integrates well with existing infrastructure
- Focus on trusted, controlled, accountable AI
- Productizing AI features harder than traditional features
- Created entrepreneurial structure alongside regular product teams
- Team designed to fail fast but stay customer-focused
- AI features can access databases without disrupting applications
- Customers want summarization across different business areas
- AI provides insights and actionable recommendations
- Conversational AI replaces traditional reporting limitations
- Working closely with customers to solve problems together
- Beyond prototyping phase, now in implementation
- AWS Nova provides excellent cost-to-value ratio
- Focus on measuring customer value over immediate profitability
- RFP use case saved half a million dollars
- Early prompts were massive, now more structured
- Setting realistic customer expectations is important
- Data security approach same as other applications
- Treating AI outputs with same data classification standards
Participants:
See how Amazon Web Services gives you the freedom to migrate, innovate, and scale your software company at https://aws.amazon/isv/