Index / Selected Work Case 01 · signalcast.tv · Ongoing

SignalCast

Free digital signage that turns any smart TV into a sign. No media player, no IT ticket. Install a TV app, pair a code, and update every screen by Telegram message, Alexa, or AI agent. I'm designing the whole thing: the on-screen experience, the control model, and the apps across seven TV platforms.

Role

Lead Product Designer

For

ITsutra

Tools

Google Stitch · Google AI Studio

Timeline

2026 · Live & ongoing

SignalCast - Mission Control dashboard Fig. 01 - SignalCast Mission Control
The bet

Digital signage without the price tag, or the IT ticket.

Traditional digital signage is quietly painful: $200–$700 media players, $15–$30 per screen every month, hours of setup, and an IT team for every change. SignalCast's bet is that none of that is necessary: the TV you already own, plus a message, is enough.

My job is to make that bet feel real: to take a category that's always been enterprise-heavy and make it something a café owner or office manager can run from their phone, exactly the kind of complex system I like making feel simple.

The challenge

One product, seven TVs, and a phone, all in sync.

The surface area is unusually wide. The same experience has to feel native on Roku, Fire TV, Samsung Tizen, LG webOS, Apple TV, Android TV, and the web (each with its own remote, constraints, and conventions) while staying coherent end to end:

  • A 10-foot, lean-back screen designed to be read across a room, not tapped.
  • A control surface (phone, chat, voice, dashboard) that's always one message away.
  • Multi-zone layouts: main content plus ticker, sidebar, or clock, without clutter.
  • Resilience designed in: screens keep playing from cache when the network drops.
SignalCast Display Fleet: every paired screen with status, location and what's now playing Fig. 02 · The display fleet, every screen at a glance
The control model

"Say it or text it." No CMS to learn.

Instead of a heavy content-management system, SignalCast is controlled with the tools people already use: a Telegram message (natural language, even voice notes), an Alexa command, the web dashboard, or, fittingly for how I build, AI agents over MCP, with tools any model can call.

The design problem there is trust: when someone types "show lunch specials on the lobby TV," they need instant, unambiguous proof the screen actually changed. So confirmation, status, and the sub-two-second update loop are as much a part of the design as the screen itself.

SignalCast Bot Integrations: Telegram, WhatsApp and Teams with a chat command reference and activity log Fig. 03 · Control by chat, with a command reference anyone can read
Onboarding

Live on screen in under ten minutes.

Install the app, enter the pairing code on the dashboard, send your first piece of content. Done. Designing that first run so it genuinely takes minutes, with no training and no IT, is the moment the whole promise either lands or doesn't.

SignalCast Scene Builder: drag-ordered content sequence with durations, transitions and live preview Fig. 04 · Scene builder
SignalCast Asset Vault: a centralized media library of images, videos, PDFs and widgets Fig. 05 · Asset vault
My process

From research to screens, the AI-native way.

I started with research. Audience research to understand who actually runs signage, café owners, office managers, front-desk staff, and market research into where the existing tools get it wrong. That fed the core user flows: how someone gets from a blank TV to a live screen with as few steps as possible.

From the flows, I used Google Stitch to turn the structure into real screens quickly, then moved into Google AI Studio to refine the visual design and the interactions. Once the direction held together, I handed the work to the development team to build out across the seven platforms.

Designing this way, AI-native from flow to screen, is a big part of why a product this broad can keep moving with one designer leading it.

Where it's at

Live, and still shipping.

SignalCast is my active flagship, so this case study grows as the product does. The clearest way to see the current state is to open it live →, or get in touch for a walkthrough.

Next case · 02 Orchestra →
All work