mobile logo


Key takeaways from the London AI Meetup: GenAI PoC to production


Prasad Prabhakaran

Head of AI practice




post hero image

esynergy proudly hosted another London AI meetup on April 25th, drawing over 60 AI practitioners from across the city. Attendees ranged from Northern Trust and Sky to HSBC, Sainsbury, BA, Crisil, and HMRC. The focus of this meetup was on moving Generative AI (GenAI) from Proof of Concept (PoC) to production, a theme resonating with AI enthusiasts, given that Forbes research indicates 90% of PoCs fail to make it to production.

Navigating the road to production

The meetup featured presentations from leaders in AI, including AWS, Snowflake, and DataStax. Vikesh Pandey from AWS addressed key hurdles to transitioning from PoC to production, such as:

  • Business buy-in: securing early support from stakeholders.
  • Stable models: ensuring models are robust and reliable. You don’t need a Ferrari to go to the supermarket!
  • Data observability: maintaining visibility into data processes.
  • Governance: establishing policies and guidelines.
  • Operations: efficiently managing operational workflows.

Suraj Rajan from Snowflake highlighted how Snowflake’s infrastructure, including Cortex, facilitates production deployments of experiments, particularly for governed data, ensuring both safety and ease of use.

Kiyu Gabriel, CTO of DataStax, emphasized the importance of vectorization and embedding strategies to create value-driven use cases with manageable TCO. He also showcased RAG orchestration for comprehensive checks and balances, ensuring secure Route-to-Live deployments.

Panel discussion: Overcoming GenAI challenges

The meetup concluded with a panel moderated by Prasad Prabhakaran, Head of AI at esynergy. The panel delved into:

  • Involving business: early engagement and defining KPIs for measuring success.
  • Data assets: leveraging structured and unstructured data securely for GenAI use cases. Importance of data ingestion and pipeline.
  • Governance: addressing key considerations to ensure compliance and data integrity.
  • Monitoring: providing visibility and control to manage risks associated with GenAI solutions.

Optimism for the future

Overall, the meetup showcased the growing optimism around GenAI’s potential to add significant value to businesses. At esynergy, we are helping organizations leverage large language models and internal data securely. Current use cases include:

  • Regulatory reporting
  • KYC
  • Business operations
  • Customer support
  • Personalization
  • Knowledge and information management

The meetup underscored the immense potential for GenAI to transform businesses across sectors, and esynergy is proud to be at the forefront of this journey.