AI
AI that does real work.
Reading the paperwork nobody wants to type, answering questions from your own data, drafting the first version for a person to finish. Built into the software you use every day.
01 What we use it for
Reading, sorting and first drafts.
The work AI does well is the reading, sorting and drafting that takes up hours of a working week.
-
01 Reading paperwork
Supplier bills, price books, plans and forms turned into structured data, ready to check and save.
-
02 Answers from your own data
Ask in plain English and get an answer from your own records, with the records it came from.
-
03 First drafts
Learning stories, quotes, emails and reports drafted from what you already have, for a person to finish.
-
04 Sorting and routing
Incoming email, submissions and requests labelled and sent to the right person.
02 Ways we build it in
Where it lives is up to you.
An agent of its own, a panel in a system you already use, a chat on your website, or quiet work in the background. Often more than one.
- ✓Checked stock levels
- ✓Found 3 items below minimum
- ✓Compared prices from 2 suppliers
- ✓Drafted purchase order to Carters
- ●Waiting for your approval Approve
Do you take Saturday bookings?
Yes, 8am to midday. Would you like me to book one?
Talk to a person03 In our own products
AI in Stringline, PISTON and Oversite.
-
Trades
Stringline
Supplier bills, price books and plans read into structured data, so a quote starts from real prices.
-
Automotive workshops
PISTON
An assistant that answers from the workshop’s own jobs, customers and vehicles.
-
Early childhood education
Oversite
A learning journal where teachers note what they saw, and a first draft of the story is ready to finish.
04 How we use it
The rules we work to.
The same four on every project that touches AI.
-
01
A person stays in charge
Nothing is sent, paid or published until a person has checked it.
-
02
It shows its working
Answers link back to the record or document they came from, so they can be checked in a click.
-
03
Your data stays yours
We use providers whose business terms keep your data out of model training, and send only what a task needs.
-
04
Only where it helps
If a form, a rule or a search does the job better, we use that. We try it on your own documents before anyone relies on it.