​The path towards 100% accuracy & repeatability

​Working with enterprise data demands perfection. When analyzing sales forecasts, processing financial transactions, or handling customer data, approximations and inconsistencies aren't just inconvenient – they're deal-breakers. A single incorrect calculation can shatter trust and make production deployment impossible.

​Join us for the first instalment of AI Data Agents, an event exploring how to achieve complete reliability when deploying AI systems on enterprise data. As organizations integrate retrieval-augmented generation (RAG) and tool calling capabilities, they face a critical challenge: maintaining perfect accuracy throughout complex multi-step tasks. We'll tackle head-on the problem of "goal drift" – where agents lose sight of their original objectives as context grows – and explore architectures that enable rigorous validation of every intermediate step.

​This event is aimed at bringing together AI change agents - technology leaders and architects as well as business leaders who are driving AI adoption themselves - who refuse to compromise on accuracy. We'll examine breakthrough approaches for achieving 100% consistent, reproducible results when operationalizing AI systems with private data, backed by real-world implementations.

​Expect deep dives into:

• ​Real-world implementations of AI agents in production environments, with focus on maintaining accuracy across long-running tasks

• ​Architectural patterns for secure private data integration that enable intermediate state validation and user steering

• ​Strategies for enhancing agent accuracy and repeatability with enterprise data, including methods to manage context growth and prevent goal drift

• ​Data preparation techniques and ETL workflows optimized for LLMs, designed for iterative processing and data aggregation

• ​Emerging tools and frameworks in the AI data preparation space that support robust retrieval and transformation pipelines

• ​Practical approaches to balance automation with user control in complex multi-step agent workflows

• ​Adoption challenges for Agents connected to enterprise data

​Agenda:

​5:30pm - Doors open

​6:15-7:00pm PST: Fireside chat with Amit Butala (AI/ML Tech leadership at Rippling) on building accurate data agents

​7:00 - 7:45: The Enterprise Data Benchmark preview - a 650 question benchmark of real world business tasks with a performance review of the top models (GPT 4o, o1, Claude Sonnet 3.5, Deepseek r1) - AI Lab at PromptQL Hasura

​7:45-8:00pm: Demos [submit here]

​8:00 - 8:30pm: Networking

​8:30pm: Door Close

If you'd like to speak at this meetup, you can submit your talk here.