Articulate

The idea started with a piece of career advice from a senior engineering manager: strong technical people still need to communicate differently with customers, engineers, and executives.

I wanted a way to practise that skill deliberately rather than wait for a high-stakes meeting to discover where I struggled.

The problem I chose to explore was simple: how do you turn communication practice into a repeatable daily habit that gives useful feedback, not just more content to read?

I built Articulate end to end, including the personalised passage generator, recording and transcription flow, speech feedback, PWA behaviour and dyslexia-friendly reading experience.

Articulate is designed for professionals who want to become more confident and deliberate speakers, especially when communicating in work settings.

The experience assumes the user does not need a generic public-speaking course. They need short, relevant practice tied to their own role, goals, and the kinds of people they want to communicate with more effectively.

That is why the app personalises passages around job title, industry, goals, focus areas, and speakers the user admires.

Practice has to feel relevant. A random passage is easy to ignore. Content tied to a user's actual work makes the habit more purposeful.

Feedback needs to be specific. A score on its own does not tell someone what to improve, so the experience combines scoring with constructive feedback on delivery.

Accessibility cannot be cosmetic. While testing, I noticed someone struggling to read the content. That changed the project from adding an optional font to redesigning the reading experience itself.

Why personalise the passage?
The product is trying to build a habit. Relevant content gives the user a stronger reason to practise than a generic reading exercise.

Why score more than accuracy?
Reading every word correctly is only one part of effective delivery. Fluency and pace make the feedback closer to how speech actually feels to a listener.

Why use the Web Speech API?
It allowed the recording experience to provide real-time transcription in the browser and connect directly into the scoring loop.

Why redesign accessibility instead of adding a toggle?
The reading experience itself was the product. If the content remained difficult to process, changing the font alone would not solve the user's problem.

Articulate became a deployed full-stack PWA rather than a static prototype.

A user can create a personalised profile, receive a fresh passage, record themselves, get feedback on delivery, save passages, build a streak, choose between two daily flows, use optional reminders, and switch into a reading experience designed for dyslexic users.

The strongest outcome for me was not the number of features. It was seeing the product evolve because of a real usability observation and then changing both the interface and AI-generated content around that insight.

If I were evaluating it with real users, I would measure:

I started Articulate as a way to practise communication more deliberately. The project became more interesting once I stopped thinking only about the AI generation and started thinking about the whole user habit.

The biggest shift was the accessibility work. It reminded me that a feature can technically exist and still fail the person it is meant to help.

If I continued the product, I would spend more time testing the scoring feedback and daily habit loop with real users before adding more features. I would want to know whether the feedback actually changes how people speak over time, not simply whether they enjoy the interface.