We have successfully kickstarted “An Agentic Journey” today.

Thank you to our 100+ attendees for joining us for 90 minutes of sharings, laughs and insights. This is the beginning of pushing our boundaries and responsibly harnessing Gen AI to move Singapore forward.

A few key sharings:

• Visual-heavy documents should be processed by a Vision-capable tool/model first

• Always inspect what the LLM is actually “seeing”, and does it make sense?

• Whatever I am serving to the LLM, is the LLM seeing what I want it to see?

• Granularity of a task matters. Am I planning the storyline? the outline of slides? or per-slide content? or a prompt to generate an image in a slide?

• Design with AI-ready design systems

• Be a curious cat; Leverage on the experts; Explore the many resources out there

• SDD vs Plan Mode vs Direct Prompts requires consideration of complexity

• MCPs are powerful and dangerous

• Do we really need an LLM? Is there a better, faster or cheaper way?

• How much LLM is required? More importantly… at WHAT STAGE is an LLM required?

Thanks Janice Tan, Jing Xuan Lam, Puay Hiang Goh, Reynaldo Wijaya Hendry, Workplace Connect, SWE Practice!