Lucy · Meeting prepBeat 1 of 12
Current goalMeeting prep · Acme renewalRead-only context · human review on output
AI coach for life and work
Lucy is the missing link between your ideas and the solutions you can build with AI.
Lucy helps you turn your own context, creativity, lived experience, and domain knowledge into AI-supported solutions for work and daily life tasks already in front of you.
She spots useful opportunities, connects them to the right AI-based tools and patterns, and coaches you through building a solution that fits your task, respects your company's policies and approved tools, and keeps you in control.
Closing the gap between AI investment and human capability — one workflow at a time.
Mon Apr 27 · 9:42 AMLucy lives on your desktop as an app that is cloud-backed, alongside the tools you already use, mapping your tasks to the latest tools, practical AI skills, and knowledge so you stay current as AI changes the way work gets done.
Lucy turns the people closest to the work into architects who change how the work actually gets done.
Lucy is a coach who is contextually aware of what you are working on. When she spots a task or workflow where an AI-shaped solution could help, she walks you through building the AI-supported solution for it.
Lucy
Coaches, never acts.
- coaches
- explains
- suggests
- nudges
- teaches
Proactively in the moment, and sits back when you are in the flow. She never acts on your data or systems.
You
Decide, design, own.
- decide
- design
- approve
- own
- improve
You are the strategist and gatekeeper for what gets built.
What YOU build
Executes within the guardrails.
- agent
- workflow
- script
- system
Executes against your data and systems, doing whatever you designed it to do within the company's guardrails — leveraging Lucy's coaching on best practices.
The skills and solutions you build are yours, and stay with you.
Core distinction
Tasks that happen to leverage AI tools are not the same thing as solutions. A task is one instance; a solution creates the desired state. Lucy helps with both, depending on your preference.
For the full product mechanism — the Coaching Loop, the Evolutionary Framework spine, and the capability archetypes Lucy helps build — see How Lucy Works →
The cost of turning ideas into solutions is collapsing.
Most of us aren't limited by ideas or ambition; we are limited by the cost of turning ideas into solutions — slow, lossy, and expensive. That cost is now collapsing.
Code generation, multimodal AI, verification loops, and agent runtimes all crossed practical thresholds in 2024–25 — what once needed a development team and months can now be a few-hour Lucy session with the right coaching.
Personalized AI coaching at workforce scale wasn't economically possible until now — desktop AI capable of reasoning about specific work, in real time, in the flow, is what makes Lucy possible at all.
When information is free and models are commodities, the scarce resource becomes the human capability to apply them.
Read the deeper economic thesis for the full case.
AI does not transform businesses. People transform businesses with AI.
The gap is the same for everyone — domain experts have the context for the AI-supported solutions their work needs, but not the design discipline to turn that context into working agents. Lucy closes that gap.
Lucy's coaching mechanism is identical for a Madrid analyst and a Texas mechanic; the leverage scales with the audience and the budget. The audience cuts below are lenses on one product, not separate products.
For businesses
The 80% of work that creates competitive advantage is the 80% that's specific to your company, your customers, your operating model. General AI tools cover the shared 20%. Lucy coaches the people doing the 80% to build the AI-supported solutions for their own work — and the capability stays with the company because the people who built the solutions are still doing the work.
For the person doing the work
You know what's repetitive in your week. You know where the meaningful judgment lives. Lucy helps you spot the AI-shaped task, build the small agent for it, and decide when the agent runs. The capability stays with you — the skill of knowing what to automate and how to design it compounds, and what you build goes with you to your next role, your next employer, and the next model generation.
For life admin
The same coaching loop works on the work that isn't a job — managing a household, a chronic condition, eldercare, immigration paperwork, the invisible labor that compounds without ever getting redesigned. The same gap applies; Lucy coaches the same way.
General tools solve the shared 20%. Lucy is for your 80%.
The unique 80% is unbuildable as an off-the-shelf product because it depends on your processes, your data, your governance, and your exceptions. It only gets solved when the people who hold that context build the solutions for it.
Solution building means wiring tools and agents into full workflows that respect your company's data, governance, and approval rules.
The question isn't which tool to switch to — it's how to layer them. Lucy coaches you on the routing literacy: how to match the shape of the work to the shape of the tool.
The thesis. General tools solve the shared 20%. Lucy helps with your 80%.
vs. ChatGPT. ChatGPT lives in another tab. You carry context in by hand and answers out. Lucy lives in the flow of your work and already knows your context.
vs. Copilot. Copilot makes existing processes faster. Lucy helps your team redesign and/or automate them.
vs. Training. Training measures knowledge. Lucy measures results. Training pulls you away from the work to teach you; Lucy coaches you while you're doing the work.
vs. Consultants. Consultants leave. Lucy stays. Consultants leave you with a deliverable; Lucy leaves you with people who can build the next one.
Full alternatives panel — including packaged enterprise platforms — on How Lucy Works.
See the full panel →Trust is architectural — Visibility, Confidence, Control, with refused capabilities at the contract level.
You stay in charge. Lucy coaches; you decide; the agents and workflows you build do the work.
Manager-aggregate-only with cohort-size gating: managers see patterns across teams; private prompts and individual rankings are architecturally refused.
V
Visibility
I can see.
Every action Lucy and the agents you build take is visible — to you, in the surface where the work happens.
C
Confidence
I know.
Provenance, source, and reasoning come with the output. Lucy shows the work, not just the answer.
C
Control
I decide.
You can pause, redact, delete, or leave — without losing what you built. Refused capabilities live in the architecture, not the policy.
From the employee's seat
Visibility
"Lucy observes my work, not my private life."
Confidence
"Lucy keeps what helps me; refuses what would harm."
Control
"I can pause, redact, delete, or leave — without losing what I built."
Capabilities we refuse
We will not sell Lucy to interrogation contexts. We will not sell Lucy to surveillance contexts. We will not sell Lucy where the architecture refuses what the customer expects.
Full V·C·C diagram, audience anchors (employee · manager · works council), public-sector EU and US sub-sections, and the refusal scene live on Trust →
Lucy is right in front of you — say hi.
Frequently asked, briefly answered.
What is Lucy Labs?
Lucy Labs is a Madrid-based AI company building Lucy, a cloud-backed desktop AI coach. Lucy helps the people doing the work — domain experts, people with lived experience — build the AI-supported solutions for their own work, decisions, and daily needs. The company is European-founded, with a US Delaware C-Corp parent and a Spain operating subsidiary, in design-partner development.
What does Lucy actually do?
Lucy lives in the flow of your work as a desktop application. Lucy spots AI-shaped opportunities in your real tasks, coaches you through building a small agent for each one, and the agent runs while you stay in control. The capability stays with the person who built it. The mechanism is the Coaching Loop on /how-lucy-works.
How is Lucy different from ChatGPT, Copilot, or training programs?
General AI tools (ChatGPT, Copilot) answer prompts. Training programs teach skills out of context. Consultants build solutions and leave. Lucy coaches you through building solutions for your real work, in flow, so the capability stays with you. The full alternatives panel is on /how-lucy-works.
Can I trust Lucy with my data?
Trust is architectural. Lucy uses the V·C·C model — Visibility, Confidence, Control — with refused capabilities at the contract level. Manager-side data is aggregate-only with cohort gating. The full architecture is on /trust, with audience anchors for employees, managers, works-council representatives, and public-sector buyers.
Is Lucy available now?
Lucy is in design-partner development. The product, the trust architecture, and the coaching loop are available. Additional surfaces are planned. Specific certifications and the manager-side admin surface are planned for GA. The current pilot path is on /design-partners.
How do I get started?
Two paths. Talk to Lucy directly using the chat widget in the lower-right corner of this page — talking to Lucy is the evaluation. Or apply to be a design partner. Both paths reach the same conversation; the chat is faster.