Selected work

Systems people use
every working day.

28 projects designed and built, from sales and finance to health and operations. Each one starts from a problem in a real process and ends in a system the team works in.

Client Zero

The Audit starts
at home.

I run the AI Decision Audit on my own companies first: the same inventory, the same kill, fix or scale verdicts, the same cut list.

  • A written list of about 60 AI tools, wins and problems across my companies.
  • A shared workspace where AI agents and people split tasks and documents every day.
  • A blind test, 48 human votes on 24 pairs of model answers, picked the model behind a sales assistant I built. Demos do not get a vote.

Selected builds

Eight of the 28,
by sector.

Proposal editor on the left and a live preview of the client-facing proposal page on the right.

Professional services · Sales

From the conversation to a reviewed proposal

Writing every proposal from scratch eats time. AI drafts the text from the conversation, and the team checks the fields and the final page before anything goes out.

Monthly statement showing October revenue, amounts due, tax set-asides by quarter and a total split across two offices.

Accounting · Operations and data

Invoices become a month-end close you can check

Documents sit in different files and have to become consistent calculations. The system imports them, computes the month, and leaves pending items and statements for review.

Finance dashboard on a phone showing what to set aside, what can be transferred today, how it splits between two partners and the month's billing.

Healthcare · Financial management

A practice's finances, on one phone screen

Revenue, taxes and upcoming payments in an experience built for the phone, fed by the same calculation the team uses.

Request board with five stages on a desktop and a phone inbox showing three notices.

Healthcare · Client technology hub

Requests and answers in one place

Tools, materials and requests stop living in separate channels. The first version is a queue with one conversation per request.

Visualization connecting audience sources, patterns and messaging.

Professional services · Market research

Market research with AI and real sources

Videos and comments feed an analysis of language, objections and themes, so the message starts from what the audience actually says.

Payment planner with sliders for down payment, annual installments and term, and a breakdown of the monthly amount.

Construction · Real estate sales

A payment plan the buyer can try out

Down payment, annual installments and monthly payments change on the screen while the visitor explores payment scenarios.

Lead inbox with an open record, next step and history, next to the log-contact form on a phone.

Sales · Lead hub

Leads with a source, a next step and an alert

Leads from several businesses in one place, with origin, segmentation and next step, and a notice when a new one arrives.

Company search screen with filters for industry, state, size and revenue, and a results list with three companies selected.

B2B intelligence · Prospecting

Prospecting by company and profile

Filters for activity, size and location find the right companies and organize their contacts in one list.

What the projects show

AI, data and systems
inside the daily work.

Each case is a problem and the system built to solve it.

  • Multi-company dashboards with access roles
  • Web applications with authentication and per-user data isolation
  • Installable mobile apps that work offline
  • AI agents for market research and customer listening
  • Citation checks done in code
  • CRM and ERP integrations
  • Data engineering, analytics and semantic search
  • Document processing and financial close
  • Interactive financial simulators

Sectors

Where the work has been done.

  • Healthcare
  • Accounting
  • Professional services
  • B2B software and sales
  • Food and e-commerce
  • Real estate and construction
  • Education
  • Personal finance
  • Fitness and wellness
  • Automotive

Next step

Which process should
run in production next?

Tell me where the work stalls and which systems your team uses. We start from that.

Request a 30-minute call