Making sense. Making things happen.

I work where things are unclear and messy, connecting people, systems, and ideas to find a way forward.

My work spans hands-on AI development, platform strategy, and consumer products. Pick a story to explore how I think.

Name
Seetha Val
Role
AI, platform and technical product leader
Based in
Montreal, Quebec
US hiring
Through my company, Productions GamesAd Inc.

Beacon

Investigating the unknown

How I shaped an AI product for exploring data and finding what’s wrong.

Open the storyClose the story

Sketch of the Beacon workspace: a conversation on the left, charts opening beside it, saved dashboards below, and a spike circled as what is wrong.

Situation

We saw an opportunity to help customers investigate their operational data themselves. Our hypothesis was that discovering problems would help them identify what they wanted automated. Beacon began as an internal analytics agent and evolved into a customer-facing product.

Task

My role was to shape and deliver the product end to end: understand users’ needs, define the experience and agent capabilities, and establish how we would evaluate and improve it.

Action

I ran customer and stakeholder interviews, translated findings into product direction, and worked hands-on across prototyping, agent behavior, and delivery.

I gave Beacon a dedicated workspace with visualizations opening alongside the conversation. I anticipated that users would need to revisit findings and view related charts together. The experience grew to include saved visualizations and dashboards with refresh controls.

Beyond the interface, I shaped the multi-agent architecture and introduced evaluation workflows to test changes to agent behavior. I used traces, product analytics, and user feedback to investigate quality and performance issues and guide improvements. I also defined product metrics and partnered with product marketing on positioning.

Result

Beacon evolved into a self-service experience where customers can investigate data, generate visualizations, and assemble dashboards they can revisit and refresh. In interviews, users described going to Beacon to “find what’s wrong.”

That feedback supports its role in problem discovery. Connecting those discoveries to automations remains the next part of the original hypothesis to validate.

Working through something similar? I’d like to hear about it.

Email sv@eccy.orgLinkedIn

Let’s talk.

Have a messy product problem, an interesting idea, or an opportunity to work together? I’d like to hear about it.

Email sv@eccy.orgLinkedIn

Or explore the experience behind these stories.