By  Melp Meredith, Jay Davis, Rachhana Saish Baliga, Samikshya Satpathy

Meet TED

 Ted is a smart, portable speaking coach designed to support users as they practice public speaking. 

It listens while you speak, offers subtle guidance in the moment, and provides a clear performance summary afterward, all without judgement or pressure. Ted is especially helpful for  students, early career professional, and non-native speakers who want to build their confidence in a low-stress way

That single quote captured a gap we saw across our survey and diary study responses that people need feedback while they're practicing, not just after. But they also need it to feel calm, helpful, and non-disruptive. 

Ted fills that gap by providing gentle real-time guidance and clear post-session insights, creating a supportive practice experience that fits easily into everyday life.

Why We Built This 

Public speaking is one of the most common sources of anxiety, especially for students, young professionals, and non-native English speakers. While there are tools that offer feedback after a presentation, they often come too late to be helpful in the moment. Users told us they wanted something different:

"I wish something could tell me how I'm doing while I'm still talking so I could fix myself on the spot."

"I only get feedback after the presentation. I want help while I'm still speaking."

In fact, over 90% of the people we surveyed said they would find real-time feedback helpful but many also mentioned that current tools feel clunky, overly formal, or require practicing in front of others.

“I wish I could get feedback, but I don’t want to practice in front of people.”
“More user-friendly.”

We realized what people really need is a system that gives them gentle, live support while they speak  and helps them grow over time  without adding pressure or judgment. That insight became the foundation for Ted’s design.

"How might we support people during speech practice by giving them real-time, gentle feedback without breaking their flow?"

Idea and Prototype

Ideation

As a group, we knew we wanted to create something small and cute. We brainstormed a few options, like an anthropomorphic flower or animals like bunnies, cats, and dogs. We ultimately landed on a bear. The choice was a bit arbitrary since TED could really take any form, but it worked well that “Ted” could be short for “teddy bear.”

Design Decisions

Once we settled on the bear concept, we started sketching ideas. This ended up being the most challenging part. We debated whether TED should have an LED ring, a TFT screen, or both. We weren’t sure what would look right aesthetically or what would be feasible to implement.

Eventually, we chose to include both IoT components to make TED as accessible as possible. That led to another round of decisions: figuring out the size and placement of the NeoPixel ring and TFT screen. We spent several days narrowing down the dimensions and orientation of both devices. . Once we had a layout we liked, we moved into Figma to develop a wireframe. 

What Worked and What Changed

Once we assembled a rough version of TED, we realized the combination of the TFT screen and NeoPixel ring actually worked better together than we expected. The LED ring added a layer of emotion or mood that complemented the visuals on the screen. Getting both to physically fit and look cohesive on the bear was a challenge, but the final version felt balanced.

We didn’t do formal user testing due to time constraints, but even showing it around informally, people immediately recognized it as a bear and found it cute — which was one of our main goals. We were surprised at how much personality TED had once everything was assembled. If we had more time, we’d definitely want to test how people interact with it, what they interpret from the light and screen signals, and how they feel about using it in context.

How Ted Works

Ted is a portable, AI-supported speech coaching device that combines a built-in microphone, LED feedback ring, and small screen to deliver real-time visual feedback during practice, along with personalized reports afterward. Ted connects via Bluetooth to a companion app that tracks progress and lets users customize their coaching experience.

Our final concept is centered around three main goals:

• Reduce anxiety and distraction by avoiding verbal interruptions

• Offer timely, actionable feedback

• Support different types of users with flexible coaching modes

Key Features

Final Video

Bringing Ted to Life

In our current prototype, we simulate TED’s behavior manually using Wizard of Oz techniques. In the real product, these simulations would be replaced by fully functional hardware and software. TED would include:

• A high-sensitivity microphone array to capture the user’s speech clearly, even in noisy environments. These microphones would detect pacing, volume, filler words, and hesitation patterns.

• A compact wide-angle camera placed in TED’s forehead or chest area to record facial expressions and gestures, allowing TED to interpret nonverbal communication like posture, eye contact, and energy.

• An embedded processor capable of running lightweight machine learning models locally. This processor would handle real-time speech analysis, facial detection, and gesture recognition without needing to send data to the cloud.

• A NeoPixel LED ring around the bear’s face or collar area to provide soft color-based feedback in real time.

• A small TFT screen embedded in the belly or face to display text prompts or visual cues without being distracting.

• Bluetooth and Wi-Fi modules for syncing with a mobile or desktop application, where users can access their feedback reports and track progress over time.

TED would no longer rely on manual input or scripted responses. Instead, it would autonomously recognize speech patterns and gestures, and respond with immediate and personalized feedback.

Vision for the Final Product 

TED would operate through three main modes:

1. Practice Mode
   Users receive real-time guidance through the LED ring and subtle screen messages. These visual cues would signal when to slow down, speak up, or keep a steady pace. The goal is to help users self-correct naturally, without losing focus or being interrupted.

2. Performance Mode
   After finishing a speech, users receive a detailed session report through TED’s companion app. This report includes insights on clarity, pacing, tone, filler word frequency, and gesture use. Feedback is structured and nonjudgmental, encouraging reflection and growth without pressure.

3. Warm-Up Mode
   Before practice sessions, TED provides light vocal exercises and confidence-boosting phrases. This mode is designed to help users feel grounded and ready. TED can also track stress markers in the user’s voice, such as pitch variability or breathing patterns, and respond with calming encouragement.
   
   

A Supportive Everyday Tool

TED is built for real-life scenarios. It is ideal for students preparing for class presentations, non-native speakers practicing pronunciation, and professionals refining pitches or job interviews. TED is soft, portable, and approachable. It is easy to use anywhere—at home, in the office, or while traveling.

In future versions, TED could integrate with mobile speech apps, AR environments, or even smart home systems to expand its utility. It could become part of someone’s daily routine, offering encouragement and accountability with a personality that feels friendly and familiar.