Turn AI capability into business capability.
Carve Lab works with established companies to find where AI materially changes the business, design the products and systems worth building, and lead teams from strategy through production.
The problem
AI capability is advancing faster than most organizations can absorb it.
Access to capable models is no longer the constraint. Nearly everyone has it, and nearly everyone has a pile of experiments to prove it. The constraint is judgment: knowing which of the things now possible are worth doing, and being able to carry one of them all the way into production.
The questions that decide whether an AI program produces anything are rarely about the model.
- Where does AI actually create economic leverage in this business?
- What should we build, and what should we refuse to build?
- How does the product or the workflow change, not just get faster?
- What belongs to a model, what belongs to deterministic software, and what belongs to a person?
- How will we know whether it works?
- How do we get from a prototype that demos well to a system the business runs on?
- How does the organization get better at this without us?
Carve Lab exists to answer those questions and then lead the execution.
Three levels of opportunity
Automation is only one form of AI value, and usually the smallest.
Most AI programs are evaluated entirely at the first level, because it is the easiest to justify and the easiest to measure. The opportunities that change a company's position are almost always above it.
01
Efficiency
Do what you already do, faster and more reliably.
Real value, and the easiest to measure. It is also where almost every AI program stops, because it requires no one to change their mind about anything.
Bounded by the cost of the work you are already doing.
02
Capability
Do things that were previously impractical.
Reviewing every contract instead of a sample. Answering a question that used to take a week. Giving a junior employee the judgment of your most experienced one. The work changes shape rather than getting cheaper.
Bounded by what your organization can absorb and operate.
03
Product & Business Model
Offer something that was not previously economically possible.
New products, new services, new customer experiences, or a different way of making money. This is where competitive position actually moves, and where the work stops being an IT project.
Bounded by imagination and execution, which is the point.
What we do
Five capabilities, held by the same people.
Not a services catalog. These are the things that have to stay coherent with each other, which is why they are not handed between firms.
01
AI Strategy & Opportunity Architecture
Where does AI actually create economic leverage here?
Understand the business well enough to tell a real opportunity from an interesting one, then rank them by value rather than by novelty.
02
AI Product & System Design
What should we actually build?
Turn a promising opportunity into a specific system: what it does, who it serves, where the model ends and deterministic software begins, and where a human stays in the loop.
03
Incubation & Validation
Does this actually work, and is it worth building?
Build enough to remove the uncertainty that matters, and establish real evaluation before anything scales.
04
Production & Engineering Leadership
Who gets this into production and keeps it there?
Stay through execution. Lead your engineers, work alongside them, or bring the team, depending on what you have.
05
AI Organizational Capability
How do we get good at this without us?
Leave the organization able to find and execute the next opportunity without an outside partner.
How we work
Discover, design, prove, productionize, scale.
Most companies do not arrive at the beginning of this. If you already have prototypes and no route to production, you are starting at Prove, and that is a normal place to start.
Enter wherever you are
01
Discover
Understand the business, the workflows, and where value actually concentrates.
What you have at the end: A ranked portfolio of opportunities with value and feasibility attached.
02
Design
Decide what should exist: the product, the workflow, the system, and its boundaries.
What you have at the end: A specific system design, with evaluation criteria defined before the build.
03
Prove
Build enough to remove the uncertainty that would otherwise sink the investment.
What you have at the end: A working prototype and evidence about reliability, behavior, and economics.
04
Productionize
Engineer for reality: reliability, permissions, escalation, cost, and operations.
What you have at the end: A system the business depends on, with the telemetry to know it is working.
05
Scale
Turn one delivered system into organizational capability and the next initiative.
What you have at the end: Patterns, practice, and teams that can do the next one without us.
Strategy through production
Strategy should survive contact with engineering.
The usual failure is not a bad strategy. It is a good strategy that loses its meaning at every handoff: to the product team, to architecture, to the engineers, to whoever ends up operating it.
We work at every one of these layers, which is a different claim than being expert at all of them. It means the reasoning at the top is still intact at the bottom.
- BusinessThe problem, the economics, the constraint that actually binds.
- StrategyWhere to spend, what to ignore, in what order.
- ProductWhat should exist, and for whom.
- WorkflowHow the work changes, and who does what now.
- SystemBoundaries, data, context, integrations.
- AIWhere a model belongs, and where it does not.
- EngineeringThe build, and the decisions inside it.
- ProductionReliability, permissions, cost, escalation.
- EvaluationWhether it works, on evidence rather than opinion.
- Business outcomeWhich changes the business, which changes the strategy, which starts again.
Delivery
Three ways to staff it, depending on what you already have.
Lead your team
We supply the product, AI, and engineering leadership. Your engineers build.
Your teams keep the context, the domain knowledge, and the system afterward. What is missing is usually not engineering capacity but someone who has taken this class of system to production before and can make the architecture and product calls with authority.
Organizations with real internal capability and no senior AI direction.
Hybrid
Your teams and Carve Lab specialists work as one group.
Usually the gap is narrow and specific: evaluation infrastructure, retrieval and context architecture, agent and tool design, or the production hardening that nobody on the team has done before. Filling it alongside your engineers transfers the capability rather than renting it.
Teams that are strong in the domain and thin in a specific area.
Carve Lab-led
We assemble the product, design, and engineering capability around the outcome.
The team is assembled around the specific problem and dissolved when it is delivered, with the system, the documentation, and the operational knowledge handed to whoever will own it. The goal is a working system you can run, not a dependency.
Operationally strong companies without the delivery capacity to build it.
Where this usually applies
Problems we are built to work on.
These are categories, not case studies. Carve Lab is early enough that publishable client results do not exist yet, and inventing them would defeat the purpose of a site like this.
Operational Intelligence
Capability
Turn fragmented operational information into decisions, recommendations, and actions at the moment they are needed.
Looks like: An operations team reconciling five systems by hand to answer a question the business asks daily.
Knowledge Systems
Capability
Make what the organization already knows available in the context where decisions actually get made.
Looks like: Twenty years of institutional judgment that lives with nine people, three of whom are retiring.
AI-Native Products
Product
Redesign an existing product, or define a new one, around reasoning, generation, and tool use rather than bolting an assistant onto the side.
Looks like: A product whose competitors are all shipping the same chat panel, and none of them have changed what the product does.
Customer Operations
Efficiency
Move past deflection metrics into genuine triage, context gathering, resolution, and clean escalation.
Looks like: A support organization where the bot handles the easy third and makes the hard two thirds worse.
Document-Heavy Workflows
Efficiency
Turn contracts, reports, forms, and correspondence into structured decisions rather than structured data nobody uses.
Looks like: Underwriting, claims, intake, or compliance review where the bottleneck is reading, not deciding.
Engineering & Product Systems
Capability
Help a product and engineering organization adopt reliable patterns for building AI-native software.
Looks like: Six teams independently inventing six different approaches to evaluation, none of them good.
Prototype to production
The prototype proves possibility. Production proves value.
A demo is built to succeed on the path you walk it down. A production system has to behave on the paths nobody anticipated, at volume, for people who did not build it and will not forgive it. Most of the engineering lives in that gap, and most AI initiatives stop at its edge.
- ReliabilityWhat happens on the bad tenth of a percent.
- EvaluationEvidence it works, on data that looks like reality.
- PermissionsWhat the system may do, and on whose authority.
- Human reviewWhere a person stays in the loop, and why there.
- SecurityWhat it can reach, and what it must never reach.
- ObservabilityKnowing it is degrading before a customer tells you.
- CostUnit economics that survive real volume.
- LatencyFast enough that people actually use it.
- FeedbackThe system gets better because it ran, not because we rebuilt it.
- OperationsWho owns it on a Tuesday afternoon in eight months.
A compelling demo is not evidence of a reliable AI product. How we think about evaluation
Who you work with
You are not handed to an implementation team.
Carve Lab is led by Hyunmin "Max" Kim, a software engineering and AI leader with more than fifteen years building and leading production systems. He currently leads AI and software initiatives at one of the largest technology companies in the world, and has built and led systems at every scale below that, from pre-Series A startups through post-IPO public companies.
The work has centered increasingly on the place where business problems, systems design, engineering leadership, and AI meet. Founder-led strategic and technical involvement is not a sales arrangement here; it is the thing being offered.
More about Carve LabHave an AI opportunity, or a business problem that might become one?
You do not need a defined AI project to start. We can begin with the business problem and work out whether AI belongs in the solution at all.
Based in Folsom, California. Working with organizations regionally and beyond.