Introducing the nvisia AI Lab
Where curiosity becomes clarity - for the organizations navigating what's next.
This is your introduction to how we explore AI, what we've been building, and how it can benefit you.
Why the AI Lab Exists
AI is moving faster than most organizations can confidently invest. Our role has never been to chase trends — but to help our clients see clearly through them.
Market Signals We're Tracking
- AI adoption pressure is outpacing organizational readiness
- Investment urgency is rising, but confidence in direction remains low
- The gap between vendor promises and proven outcomes is widening
- Fear of the wrong investment is slowing more decisions than lack of interest
Our Response
nvisia created the AI Lab as a buffer between uncertainty and commitment
We are not selling speed. We are offering discernment through learned experience.
What the AI Lab Is — and What It Is Not
A dedicated environment for exploration, testing, and learning — not a product, not a promise.
What it is:
- ✅ A dedicated environment for exploration, testing, and learning
- ✅ A place where nvisia invests ahead of client demand
- ✅ A proving ground for ideas, tools, and approaches
- ✅ A portfolio of real experiments — not theoretical slides
What it is not:
- 🚫 A product
- 🚫 A packaged service
- 🚫 A hype engine
- 🚫 A guarantee of outcomes
The AI Lab isn't something we sell. It's something we invest in — so our clients don't have to… until they're ready.
35+ years of helping organizations see around the corner — AI is simply the next chapter.
Our Approach: Curiosity Before Commitment
For over 35 years, nvisia has helped clients anticipate technology shifts, prepare teams and systems for change, and avoid reactive, fear-based decisions.
How we carry the curiosity:
- nvisia absorbs early risk so you don't have to
- We explore platforms, hardware, and models firsthand
- You see how things actually behave — not just hear vendor claims
- Learnings are shared transparently, not locked away
What this means for you:
Reduced uncertainty before you invest
Better questions before committing capital
Faster alignment when you're ready to move
Confidence rooted in observation, not assumption
We don't chase the horizon — we walk toward it with intention, so our clients can follow when the ground feels solid.
What We're Exploring
Real experiments, not theoretical promises — practical AI questions explored before there's pressure to invest.
These aren't theoretical questions — each one is being actively explored through a prototype you can see at the Symposium.
Understanding Complex Systems Faster
Helping teams understand legacy systems, processes, and decisions more quickly without risky shortcuts.
Secure, Local AI Use
Exploring how AI can operate safely when data privacy, compliance, or cost are critical concerns.
AI as a Teammate, Not a Black Box
Designing AI systems where humans remain in control, with visibility and confidence in outcomes.
Spec-Driven Development in Practice
What does it look like to apply a structured, spec-first methodology to building an AI-powered tool from scratch — and what breaks down when the problem is still being discovered?
Autonomous Code Generation
Exploring what it looks like when a coordinated swarm of AI agents takes requirements all the way to production-quality code — without human review between phases.
Legacy Modernization
Using AI agents to understand large, aging codebases — surfacing what the system actually does before designing what it should become.
AI as a Personal Productivity Layer
Exploring how AI can work quietly in the background to capture, organize, and surface what matters — without creating new privacy or compliance risks.
Agentic UI (AGUI)
What happens when the interface itself is generated by AI — not just the content inside it? Exploring what it looks like when lazy loading is replaced by live context, and every user sees a different experience.
The challenge isn't AI itself. It's knowing when, where, and how to engage responsibly — and we're doing that work now, on your behalf.
How You Can Engage With the Lab
Flexible ways to explore what's possible — no obligation, no pressure.
You can engage with the AI Lab in the way that best fits your curiosity, timing, and comfort level. There's no single path.
1
Events & Showcases
Experience AI Lab demos at nvisia events like Tech Showcase & ProdCon
2
Virtual Lab Experience
Explore demos and prototypes remotely — ideal for distributed teams
3
Lunch & Learns
In-person or virtual sessions where we share what we're learning, what's changing, and what to watch for next
4
Pre-Project Exploration
Collaborate with us on an idea you're curious about — before it becomes a formal initiative
How we explore AI — responsibly.
Our Principles & Guardrails
Human-First, Always
AI supports people — it does not replace judgment, accountability, or expertise.
Security & Privacy by Design
We assume data sensitivity by default and explore approaches that respect it.
No Black Boxes
If we can't explain how something works or why it made a decision, we slow down.
Predictability Over Hype
We prioritize approaches that are testable, repeatable, and grounded — not flashy demos.
Experiment, Then Share Honestly
We share what works and what doesn't — no cherry-picking.
Trust is built not by claiming certainty — but by demonstrating care, humility, and readiness.
Explore Our Prototypes
Each prototype below represents a real experiment — built by nvisia consultants exploring specific AI questions. Scan the QR code at each booth to go deeper, or explore them here before you arrive.
Dark Factory · Kevin Quon
What happens when AI writes, tests, and ships code — without a human in the loop?
A fully autonomous L4 code generation platform. Drop in requirements, get back production-quality code, tests, and validation reports — no human intervention between phases.
Legacy Modernization — Smart Agent · Mark Panthofer
What if you could ask a 20-year-old codebase anything — and get a trustworthy answer?
A containerized code intelligence system that deeply analyzes legacy Java codebases and generates clean OpenSpec specifications for a modern replacement — with humans reviewing every step.
We built a working Dark Factory in 8 days. Not a mockup. Not a slide deck. A running system.
Shadow Notes · Mark Panthofer
What if your calendar already knew everything that happened in your meetings — privately and automatically?
A Windows desktop app that transcribes and summarizes meetings entirely on-device, then places structured notes directly on your Outlook calendar. No cloud recording. No data leaving your device.
Sherlock — Legacy Code Intelligence · Ruben Rotteveel
When the code has been running for 30 years and the person who wrote it is gone — how do you figure out what it does?
An AI agent system that reverse-engineers requirements from legacy codebases — COBOL, Java, batch files — and produces the documentation needed to rebuild them.
Legacy Navigator · Justin Montgomery
What if you could ask a legacy codebase anything — and actually trust the answer?
A locally-run CLI tool that ingests a codebase, builds a semantic understanding using vector embeddings, and lets developers navigate unfamiliar code through natural language questions. Built as a hands-on experiment in spec-driven development using Java, Spring AI, and the Embabel agentic framework — with a model-as-judge evaluation loop built in.
Personalized Retail Recommender · Kevin Quon & Michael Arce
What if your shopping app knew you — and rebuilt itself every time you opened it?
A multi-agent retail recommendation system that learns from your behavior, explains every suggestion it makes, and generates its own UI on the fly using agent-user interaction (AG-UI) protocol. Backed by vector embeddings, persistent memory, and a closed feedback loop.
More prototypes coming — visit us at the AI Symposium · Chicago · May 7, 2025
Intrigued by what you've seen? Reach out to the nvisia AI Lab — we'd love to talk.