What we build
- Field data entry
- Time tracking
- Back-office software a spreadsheet can't handle
- Internal tools that replaced a spreadsheet or paper
- Web, iOS and Android
Owners who know exactly how their work should flow, and may have tried building it themselves with AI or a no-code tool. It worked until more people used it, the data started to matter or something broke, and nobody had time to fix it.
AI built in
Your tools, with AI that does the typing.
This is the part a weekend AI project doesn't reach: AI inside your own system, connected to your data and your rules. It reads paper, takes voice entries, checks the numbers and writes the reports, and your team has the final say.
Moisture 18%
Pieces counted 42
3 damaged, set aside
Reads paper and photos
Snap a form, an invoice or a delivery slip. The fields fill themselves, and your team checks instead of retyping.
Voice entry in the field
Say the numbers out loud with gloves on. The AI puts each one in the right field and reads it back.
Catches mistakes as they happen
Entries that don't fit the usual pattern are flagged on the spot, not weeks later in a report.
Reports that write themselves
Daily and weekly summaries in plain English, sent to whoever needs them, with the numbers behind every sentence.
Ask your data in plain English
"Which sites ran over budget this quarter?" answered from your own records, no new spreadsheet needed.
Moves your old spreadsheets in
AI maps messy columns and duplicates into the new system and shows you every change before anything is saved.
Agents that run the routine
An agent is AI that works through a whole job by itself, step by step, and reports back. Ours run on a schedule inside your system, with your rules.
Every Friday afternoon
- Collects the week's field entries
- Checks the totals against last month
- Writes the report and flags anything odd
- Sends it to the office
Run it from ChatGPT or Claude
We give your system an MCP connection, the standard way AI assistants plug into software. Then you run your business by asking the assistant you already use.
- “Which sites haven't reported this week?”
- “Add Tuesday's delivery slip to job 214.”
- “Draft the monthly summary for the owners.”
You stay in chargeYou decide what the AI may do on its own and what waits for a person to approve.
Your data stays yoursWe use AI providers whose business terms don't train their models on your data, and send only what each task needs.
Costs you can seeAI usage is metered and capped, so a busy month never turns into a surprise bill.
Built for real businesses
For Richmond Systems, a software company in British Columbia, we're building field data entry software used to record measurements outdoors: fast, one-handed entry with checks as you type, so mistakes are caught on site instead of back at the office.
How a project runs
- 01
Understand the work
We map how your business runs today: who books, who does the job, who gets paid, and where it breaks.
- 02
Design the screens
You see and click through the screens before any code is written, and change what doesn't fit.
- 03
Build
Web, iOS and Android, with the dashboards and back office behind them.
- 04
Test
We test on real phones and with real staff before anything reaches your customers.
- 05
Launch
Domains, app store listings, payments and accounts set up and handed over to you.
- 06
Look after it
We keep it running and keep improving it. Wheel Easy's system has run every day for over two years.
What clients say
I appreciated their innovative spirit as they did not hesitate to offer suggestions and options to improve my product.
Questions
Why not build it myself with AI?
You can get a working demo with AI in a weekend, and for some things that's enough. The hard part comes after: payments, logins, data that must not be lost, app store review, and someone to fix it when it breaks on a Saturday morning. We use AI to build faster, and we take responsibility for it working, for years.
We already started building it ourselves. Can you take it over?
Yes. We look at what you have, keep what works and rebuild what won't hold up, so you don't start from zero.
How long does it take?
A first version usually takes 6 to 12 weeks, depending on what it needs to do.
Who owns the software?
You do, once the project is paid in full: the code, the accounts and any app store listings.
Can it connect to the systems we already use?
Usually, yes. We can import your spreadsheets and connect to systems that offer an export or an API. We'll tell you on the first call.
What are AI agents and MCP, in plain terms?
An agent is AI that carries out a multi-step job by itself, like confirming tomorrow's bookings, and reports back. MCP is the standard way AI assistants such as ChatGPT and Claude connect to software; with it, you can run your own system by asking your assistant.
Can you add AI to software we already have?
Often, yes. If your system has an export or an API, we can connect an AI assistant to it. We'll tell you what's possible after a look at it.
Is our data safe with AI?
We use AI providers whose business terms don't use your data to train their models, send only what each task needs, and keep the records in your own system.
Which AI do you use?
We pick the model that fits the job, usually from Anthropic (Claude) or OpenAI, and can switch when a better one comes along.
What happens after launch?
We look after hosting, updates and fixes for a monthly fee, and quote new features separately.
Do you work with businesses outside Ontario?
Yes. We're based in London and Barrie, Ontario, and we're building for a client in British Columbia now.