We learn how your team actually works, then design and improve the systems around it.
Examples of product development, systems work and technical advice. Client identities and operational details are kept private.
We examined feasibility, integration requirements and delivery options for a proposed platform. We compared possible approaches and identified the dependencies and open questions that would affect the work.
The engagement delivered architecture, staffing and cost plans, giving the team a basis for deciding what to fund and build next.
We built a learning prototype that combined a visual editor with an AI tutor. The work covered the interface and supporting services, with keyboard and screen-reader interaction included in the design.
After trials with learners, we handed over a working prototype with internal administration and evaluation tools. The team could continue development and tune the AI tutor's performance, with ongoing technical guidance from us.
Working with the founders and users, we helped develop an early prototype into an operational product. The work connected interfaces, APIs and data models with the systems around them. Close attention to daily tasks shaped what people saw and how they used it; investigating inconsistent data also meant looking beyond the application itself.
Our responsibility extended through production releases, onboarding and technical support, alongside forecasting and AI-assisted data tools.
We developed and maintained a customer portal, supporting the business as its needs evolved.
We then built a broader self-service portal, giving the team the tools to manage ordering across multiple clients and expand its use at their own pace.
Software, systems and technical advice, grounded in the way your organisation works.
AI-assisted engineering. We use coding agents throughout research, development and testing. Knowing how to direct them, check their assumptions and refine their output is part of our engineering practice. How we work with agents →
We design and develop software around the people who use it and the work it needs to support. Product decisions, interface design and engineering belong together. We work across the application and its supporting services, with attention to how the whole system behaves in use.
Some problems span several tools, unreliable data or a process that depends on manual work. We trace the problem, connect systems and design changes around the operation. Where the route is uncertain, we investigate the options and develop an architecture or delivery plan that accounts for cost, security and maintenance.
We design and build agents for your products and operations. We develop evaluations to test how AI systems perform on the tasks that matter to you. We connect them with existing software, machine learning models and deterministic tools, with permissions and review suited to the work.
We work closely with you, taking responsibility for the technical work and making the decisions understandable.
Start with the operation and the people involved. Understand what is working, where the difficulties lie and which questions need answering.
Agree the approach and scope, then keep the work visible as it develops. Test decisions against the conditions the software needs to serve.
For software changes, take care of release and check the result in use. Agree what follows, whether that is ongoing development, maintenance or a handover to your team.
The first conversation establishes fit. Scope, timing and cost are agreed before work begins.
Founder & Principal Engineer
Ben works across product strategy, software architecture, data and applied AI. He works directly with clients to understand the problem, choose an approach and carry the work through implementation.
The practice combines technical advice with hands-on engineering. Ben collaborates with client teams and brings in specialist support when an engagement needs it.
More about Ben and his writing →Notes on understanding the work, building useful software, and choosing where AI helps.
How understanding, clear decisions and short feedback loops help us work effectively with coding agents.
How observing daily work shapes interfaces, data models and useful product changes.
Choose where a capability belongs, then design for clear actions, prompt feedback and interrupted work.
We’ll listen first and tell you honestly whether we can help.
You can also email ben@nscribe.ai.
