Services / Custom AI engineering
We design, build and run AI tools that package your method for your team and your clients. They’re in production, with real client teams, today.
What we do
Your method, frameworks and playbooks are your edge. We engineer AI tools around them, for your team and for your clients.
Assistants, agents and tools designed, built and run in production, with monitoring and cost controls.
An expert that answers from your frameworks and each client’s own documents, available whenever they need it.
Your playbooks, processes and plans made searchable by AI, with each client’s data kept separate.
Software that reads the inbox, matches bookings, orders and forms, and flags only what a human needs to check.
AI that researches a prospect and scores them against your ideal client before a salesperson picks up the phone.
Tutorials and walkthroughs recorded, captioned and voiced automatically, so they stay up to date.
Case study · SalesStar Global · Ask Penny
SalesStar’s coaches can’t sit in on every rep’s week. Ask Penny can. It knows SalesStar’s frameworks and each client’s own playbooks, lives inside the CRM the reps already use, and answers from their documents, not the open internet.
The full SalesStar story →In production with client sales teams in several countries
Hosted in Frankfurt for GDPR
Data-isolation tests passed between client knowledge bases
Tutorial videos recorded with an automated pipeline
The work itself
Tap a video for sound.
More results
SalesStar Global · Sales coaching · five markets
Every new lead is checked against SalesStar’s ideal-client rules within minutes, then researched and briefed by AI before a coach picks up the phone.
Leads enriched in the first two weeks, 0 failed, and all 67 ideal-client leads delivered to the CRM
Language Solutions Aotearoa · Interpreting & translation · NZ
We brought bookings, interpreters, calendars and paperwork together in one operations hub. An AI process reads and matches the information as it arrives, so the team can spend its time on the jobs that need a person, and GoHighLevel reaches out to new clients.
Our process
We start with the process that costs the most hours or holds the most IP, and agree how we’ll measure it.
What the AI needs to know, where it lives, and who is allowed to see it.
Tested against real questions, including whether one client could ever see another’s data, with an evaluation designed by our research partner.
Inside the tools they already use, with tutorial videos so people get value in their first session.
Monitoring, cost controls and usage reporting, then the next workflow.
We start with the single process that costs you the most hours, ship it, and measure it before building the next one.
We host in the region your clients need: Frankfurt for EU/GDPR work, New Zealand or Australia for local data. Sensitive records never go to a model that shouldn’t see them.
Every build ships with tests for the things that matter, such as whether one client can ever see another’s data.
Monitoring, cost controls and handover documentation are part of the build, not an afterthought.
Who does the work
Founder · Strategy, AI & growth
Founded AlphaBean in 2017, straight after a degree in computer science and statistics. Jono has spent more than ten years in digital marketing, with a statistician's eye for what the data actually says. He is completing a master's in artificial intelligence at the University of Auckland while building the business. He scopes every engagement himself and builds the ROI model before anything is proposed.
Full-stack engineer
Builds the application code behind our AI products, including Ask Penny's API, agent and data layers. If it has to run in production and stay up, Aron builds it.
BI developer · Power BI & Snowflake
Igor is a BI developer and data professional specialising in data engineering and visualisation with Power BI and Snowflake. He builds intuitive dashboards and models that turn complex data into actionable insights, with experience across transport, government, and supply chain industries delivering scalable, data-driven solutions.
Founder and chief scientist of Alethic Innovations and a former adjunct professor of psychology at Columbia University. He holds an interdisciplinary PhD in psychology, philosophy and political studies from Stanford, and wrote the book on how accurate human judgment is and how to improve it. With us he designs evaluations: how we test whether an AI tool, or a programme, is actually working.
AI coaches built on your method, private knowledge bases, operations hubs that read the inbox and match bookings, lead research and scoring, and automated training videos. We design, build and run them in production.
Each client’s data is kept separate, and we test that one client can never see another’s. We host in the region your clients need; Ask Penny, the AI coach we built for SalesStar, is hosted in Frankfurt for GDPR.
A real product. Ask Penny is in production with client sales teams in several countries, with monitoring, cost controls and usage reporting.
With one workflow: the process that costs the most hours or holds the most IP. We agree how we will measure it, build and test it, and launch it inside the tools your team already uses.
Start the conversation
A few lines about your business and what you want to grow. We'll come back within one business day to set up a 30-minute discovery call. No pitch until we understand the problem.