SOFTWARE IN THE AGE OF AI
We combine AI with software engineering to build new kinds of products, intelligent business systems and AI agents for our clients, and we keep creating products of our own.
We care about more than whether something can be built. We care whether an idea becomes a product people use. These case studies come from our own product work; client projects will be added as they can be published.
Case study · Web security
Making HTTPS certificates a non-event.
A certificate service that issues free HTTPS certificates with one DNS record and renews them before they expire, for people and for AI assistants.
Read the case study CertverNew certificate
shop.example.com
Case study · File transfer
File transfer with no server in the middle.
Browser-to-browser file transfer: no app, no account, encrypted end to end, and no copy of the file ever stored on our side.
Read the case study BeeDropCase study · Practice management
A back office built around confidentiality.
Practice management for counseling studios: scheduling, client records, consent, payments and assessments, in one system designed for sensitive data.
Read the case study EmpathiaThis week
New certificate
shop.example.com
Useful AI goes into the product, understands context, connects to tools and data, and does real work. We combine model capability, systems engineering and product design so that AI moves from a single conversation to a capability that runs every day.
Claude · Certver MCP server
Is the certificate for shop.example.com OK?
Opportunities in software come from clients, and also from our own exploration of new technology, new interfaces and new ways of working. We build and run our own products, test ideas in real use, and bring what we learn back to every client project.
AI is changing how software gets built, but good products still need clear judgment, solid engineering and an understanding of real needs. We bring AI into how we build, and engineering standards into every stage of a product's life.
Users, goals and constraints first. We agree in writing what success looks like, and whether AI is the right tool for it.
We use AI to explore and build faster, and test the riskiest assumption early with a working prototype, so decisions rest on evidence.
Security, permissions, reliability, cost and evaluation are designed in before launch, not patched in after it.
After release we monitor, measure and improve, as the models, the data and the business change.
Not every problem needs AI. We start with the users, the business goal and the constraints, then decide between models, automation, conventional software or a mix of them.
AI takes the work that suits automation; people keep the judgment, the oversight and the decisions that matter. We design that division of work on purpose.
Software is not finished at launch. It is a product that changes with the technology and the business, so we build it to be changed.
A good demo is not the goal. We plan for data security, access control, reliability, cost, evaluation and maintenance, including interfaces other AI systems can read, like this site's llms.txt.
Exploring an AI product, redesigning an existing business system, or looking for a way to bring AI into real work? Talk to us, and we will help turn what is possible into a product people can use.