moodflo

Year

2025

Industry

Tech Health

Scope of work

Founder • Research • UX • UI • Prototyping • System Architecture • Pilot Strategy

Timeline

6 months

Problem

You don’t always hear when someone is struggling.

It hides in digital silence.

It smiles in meetings.

It keeps going… until it can’t.

Modern work created visibility for performance —

but not for pressure, stress or quiet quitting.

You don’t always hear when someone is struggling.

It hides in digital silence.

It smiles in meetings.

It keeps going… until it can’t.

Modern work created visibility for performance —

but not for pressure, stress or quiet quitting.

You don’t always hear when someone is struggling.

It hides in digital silence. It smiles in meetings. It keeps going… until it can’t.

Modern work created visibility for performance —but not for pressure, stress or quiet quitting.

Solution

Build Moodflo: an AI meeting layer that detects emotional tone and team energy in real time, transforming invisible pressure into actionable mental health insight.

Outcome

Moodflo moved beyond concept into traction.

• Three enterprise pilots secured (Tesco, MetLife, Soldo)

• Engineer building for google meets/zoom plug in

• 2 x investor interest

• Bootstrapping currently to retain control

What began as a design exploration became an operational product with enterprise credibility - proving the problem is real, the signal is measurable, and the market is ready.

Moodflo moved beyond concept into traction.

• Three enterprise pilots secured (Tesco, MetLife, Soldo)

• Engineer building for google meets/zoom plug in

• 2 x investor interest

• Bootstrapping currently to retain control

What began as a design exploration became an operational product with enterprise credibility - proving the problem is real, the signal is measurable, and the market is ready.

Moodflo moved beyond concept into traction.

• Three enterprise pilots secured (Tesco, MetLife, Soldo)

• Engineer building for google meets/zoom plug in

• 2 x investor interest

• Bootstrapping currently to retain control

What began as a design exploration became an operational product with enterprise credibility - proving the problem is real, the signal is measurable, and the market is ready.

Most meetings hide how people actually feel.

Cameras go off, mics go on mute, and the only time you “see” a problem is when work slips or someone burns out.

Surveys don’t fix that—by the time a form goes out, the moment has passed, emotion has gone and the answers are filtered, polite, or forgotten.

Moodflo targets the live moment: it reads silence, tone and participation as the meeting happens and gives one clear room‑level signal, so leaders can act before pressure become visible.

Most meetings hide how people actually feel.

Cameras go off, mics go on mute, and the only time you “see” a problem is when work slips or someone burns out.

Surveys don’t fix that. By the time a form goes out, the moment has passed, emotion has gone and the answers are filtered, polite, or simply forgotten.

Moodflo targets the live moment: it reads silence, tone and participation as the meeting happens and gives one clear room‑level signal, so leaders can act before pressure become visible.

Phase One * Define

First things first; I needed to know if anyone actually wanted this. I ran a survey to 50 managers and leaders.

Most meetings hide how people actually feel.

Cameras go off, mics go on mute, and the only time you “see” a problem is when work slips or someone burns out.

Surveys don’t fix that—by the time a form goes out, the moment has passed, emotion has gone and the answers are filtered, polite, or forgotten.

Moodflo targets the live moment: it reads silence, tone and participation as the meeting happens and gives one clear room‑level signal, so leaders can act before pressure become visible.

First things first; I needed to know if anyone actually wanted this. I ran a survey to 50 managers and leaders.

Phase One * Define

First things first; I needed to know if anyone actually wanted this. I ran a survey to 50 managers and leaders.

Survey Outcome

76% Market potential. That was enough to move from “nice idea” to “this could actually change how people run meetings.”

From there I researched with claude possisble current uses for such a tool around the world. Nothing in my chosen market was currently active.

Further research led me to ask: “If this really lived inside Zoom or Teams, what would you expect to see?”

Those prompts surfaced lists of metrics, states and layouts that I then cut back to what mattered: a handful of room‑level signals. I then created a lo-fi wire and design for feedback.

76% Market potential. That was enough to move from “nice idea” to “this could actually change how people run meetings.”

From there I iterated prompts around a single question: “If this really lived inside Zoom or Teams, what would you expect to see?”

Those prompts surfaced raw lists of metrics, states and layouts that I then cut back to what mattered.

76% Market potential. That was enough to move from “nice idea” to “this could actually change how people run meetings.”

Those prompts surfaced lists of metrics, states and layouts that I then cut back to what mattered: a handful of room‑level signals.

I then created a lo-fi wire and design for feedback.

Concepts to Clarity

I defined measurable signals already possible with AI: tone, sentiment, silence, participation, vocal energy, mood, volatility, atmosphere, and pauses.

I then shared it with three willing participants for testing from tesco, MetLife and Soldo.

“What do these scores actually mean?”
“There’s too much information.”
“Overwhelming.”

The feedback made one thing clear: it wasn’t intuitive. It was confusing.

I defined measurable signals already possible with AI: tone, sentiment, silence, participation, vocal energy, mood, volatility, atmosphere, and pauses.

Each signal is tracked during the call, then combined into simple room-level states — no individual scoring, no data wall — just a clear overall signal for the facilitator. I then shared it with three willing participants for testing from tesco, MetLife and Soldo.

“What do these scores actually mean?”

“There’s too much information.”

“Overwhelming.”

The feedback made one thing clear: it wasn’t intuitive. It was confusing.

Moving the needle

Next, I amped up the visuals to feel more like a real product and reduced the interface to five core panels: Energy, Participation, Volatility, a live emotion graph, and an overall meeting room summary.

I added a Zoom‑style view and a Coaching panel with practical tips for keeping the conversation healthy. I also introduced Export JSON for comparing rooms over time and Export PDF for sharing how the meeting turned out and suggested well-being team tips for the next meeting.

This static version worked much better. The clear winner was the post-meeting summary and tips.

After multiple prototypes in Lovable and parallel tests in Base44 confirmed the newer concept, I brought in a developer to build a functional version that could be tested in the wild.

Moving the needle.

Next, I amped up the visuals to feel more like a real product. I added a Zoom‑style view and a Coaching panel with practical tips for keeping the conversation healthy.

I also introduced Export JSON for comparing rooms over time and Export PDF for sharing how the meeting turned out and suggested well-being team tips for the next meeting.

Challenges

Building a live plugin that converts surface audio into measurable signals requires considerable backend infrastructure (and cost) and for a testable MVP, i wasnt totally confident with the design and overall feel of the dashboard.

However, I had momentum and i didnt want to put the brakes on. Quickly another solution presented itself.

Workaround

Instead of relying on live call input, I pivoted to uploaded pre-recorded meeting content. The experience could still pull data except this time it was from a video and not 'live'.

I created the 'live' environment locally with terminal which saved on backend ai infrastructure and uploaded a video into Moodflo.

Visually, it lacked energy. The shifts in room tone weren’t rewarding to watch, and the interface didn’t create any meaningful sense of movement or feedback.

After more user testing, feedback was confident but the design was lacking/needed improvement. One comment was it 'dominated' rather than augmented the meeting room experience.

After a few hours of back and forth, the design fell into place.

Instead of relying on live call input, I pivoted to uploaded pre-recorded meeting content. The experience could still pull data except this time it was from a video and not 'live'.

I created the 'live' environment locally with terminal which saved on backend ai infrastructure and uploaded a video into Moodflo.

This put me in a different place completed. After a few hours of back and forth, it fell into place.

Smaller interface, great impact

The slide-tray approach felt right from the very start. It keeps the meeting central while still allowing me to show the key panels: energy over time, sentiment balance, and room tone/participation and flow suggestions.

These compact panels give quick cues about overall engagement and mood without covering faces or shared content, reducing distraction compared with the full‑screen analytics view. I still had all the functionality but now real estate had been reduced to 20% view with a tab only view should you wish to see the whole room.

I created several views with figma-make. This allowed me to build interaction and motion into the experience which further expanded the design.

After a few hours of prompting and tweaking, I had my MVP:

https://www.figma.com/make/ZA4R1cCBbENlhaPPvgHg5U/Right-Hand-Slide-Out-Drawer-UI?t=sIQSFxBZWcPApNB3-1

The slide-tray approach felt right from the very start.