A focused AI workshop that helped Nyra Health's engineers ship faster, at their signature high bar.

Nyra Health is a leader in digital therapy software for neurological rehabilitation, a category where trust, quality, and engineering discipline matter. Its leadership saw a practical AI opportunity: help an already excellent engineering team use modern coding tools more deliberately, so the company could ship even more improvements while keeping the standards its users rely on.

What the session produced

1. Gave an experienced, discerning engineering team hands-on exposure to the current generation of AI-assisted development, refreshing views formed on earlier tools, in a single four-hour session.
2. Gave developers practical exposure to modern workflows with tools such as Cursor and Claude Code, and a clearer sense of when to delegate to AI and when to apply engineering judgment.
3. Reframed AI as a way to ship more product improvements with the same team size, not as a replacement for engineering judgment.
4. Established a workshop format Nyra Health can extend beyond engineering as it builds AI nativeness across product and the wider organisation.

Specialty Tokens was not asking engineers to suspend their judgment; it was helping them update it for the current generation of AI tools.

Why the shift stuck

1. A focused workshop shifted posture: developers left testing tools and approaches they had not used before.
2. The session tested modern coding harnesses against real developer concerns (trust, code quality, review burden, unfamiliar patterns) rather than abstract AI evangelism.
3. Leadership aligned the effort around one clear standard: move faster while staying proud of what gets shipped.

The Specialty Tokens workshop format

Strong teams don't need AI awareness; they need current answers. Models and coding harnesses change every few months; a tool that was weak in a given language six months ago may now deserve another look. The workshop format is designed to close that gap fast.

• Step 01, name the open questions. The session starts by surfacing the team's real questions about AI tooling: not generic concerns, but the specific ones this team holds about their codebase, their tools, their review culture.
• Step 02, test them live. Rather than presenting benchmarks, the workshop puts current tools in front of the team on realistic tasks, so engineers can compare today's capabilities against their inherited impressions.
• Step 03, keep judgment central. The premise throughout: the human manages the AI. The goal is not blind trust; it is better leverage, with the engineer making the final call.

The format is deliberately short. Four focused hours are enough to change posture; the compounding happens afterward, in daily work.

A strong team deserves current answers

For many teams, AI adoption starts with basic awareness. For developers, the starting point is different. Engineers have worked with AI coding tools longer than most knowledge workers, so their views are grounded in real experience with earlier generations of models, and the tools move faster than anyone's impressions can.

Nyra Health's team brought exactly that grounded experience, paired with deep care for product quality, regulatory context, and craftsmanship. That high bar made the adoption question more specific, and more interesting: how can developers use AI to increase output while keeping full control of the code and product experience?

Re-testing what modern AI coding tools can do

Following the format, the session opened by naming the questions experienced developers rightly ask of AI tooling:

"AI is bad at my specific codebase, language, or framework."

"I need to understand and review every tiny change before I can move forward."

"The AI may have introduced a better pattern, but I do not trust it until I study the whole thing myself."

These are the right questions in a serious engineering environment. The workshop created room to answer them against current tools rather than earlier impressions.

Nyra Health's team was already familiar with tools such as Cursor and Claude Code; some developers were using Claude Code with multiple models to avoid throughput limits. Specialty Tokens helped turn that tool access into a more intentional working practice: when should AI explore, draft, refactor, or compare approaches, and when should the engineer slow down and apply judgment?

High standards and openness, at the same time

The engagement fit because Nyra Health combines two traits sometimes treated as opposites: high standards and openness to new technology. Leadership wanted the team AI-forward, but not casual about quality. The product still had to meet the standards of a health-related software environment, and developers still needed to understand the system and make the final engineering calls.

That framing made the workshop credible. Specialty Tokens was not asking engineers to suspend their judgment; it was helping them update it for the current generation of AI tools. By the end of the session, developers were trying tools and approaches they had not used before. The most important result was not a single model preference or workflow template. It was the change in posture: actively testing what modern AI-assisted development makes possible, with impressions kept as current as the tools themselves.

Building an AI-native organisation

Engineering is where AI tooling is most mature, and where the workshop delivers the fastest, most measurable shift. But the underlying question Nyra Health is answering applies to every function: how does a team become AI-native while keeping the judgment that makes it great?

The same format extends naturally to product teams: naming each discipline's real questions, testing current tools against real work, and keeping human judgment in charge of the outcome. For a leader like Nyra Health, the engineering workshop is the first step in that broader direction: proof that AI nativeness can be built deliberately, one team at a time, with the bar staying high everywhere.

More leverage for a high-bar team

The workshop supported a broader goal: ship more product improvements with the same team size while preserving the care that makes the product trustworthy. Speed alone is not the win. The useful shift is an engineer opening Claude Code or Cursor for a real task, testing the output against the product's constraints, and keeping the final call with the human who understands the system. That is the practical edge of AI for healthtech.