Whatever your organisation does, part of your people’s week goes to work that comes round again and again. Here’s how we give that time back, and what it turns into.
Most organisations have handed staff an AI tool and left them to work it out. It doesn’t know your format, your standards or your rules, so every draft needs rewriting.
We set it up for the work itself. You pick the three pieces of recurring work that take the most of your team’s time, and we set each one up from two or three of your best past examples, inside the AI your team already uses. Your people paste in their notes and get back a draft that reads like your organisation wrote it, ready to check and send. The first is working on real work in about three weeks, and all three within 90 days.
How you’ll know it worked. Before we start, we agree what the freed time is for: more billable work, more people supported, an outside cost you can stop. At day 90 you get that number in dollars, from your own records.
After 90 days, it keeps improving. Your team brings in new work as they need to, with no limit on how many people use it. Every month we check that each one still works, keep the rules it relies on current, and fix anything the AI’s own updates break. Every quarter you see what the freed time turned into.
Underneath it all, everything draws on one shared picture of your organisation: who you are, who you serve and how you sound. That’s why the second and third pieces of work come faster than the first, and why the quality holds.
Some questions come up every week, and the answer waits on the one person who knows the rules. Can this site take that building? Can we claim this, and at what price limit? Until they’re free, the work stops.
Our rules-checking tools answer from the rulebook itself, in plain words, so your expert checks an answer instead of hunting for one. Every answer quotes the rule it rests on. Anything the tool hasn’t been given is refused, not guessed. A person signs off before anything is used.
Describe a proposed planning change, such as six-storey housing near light rail. It reports where that runs into the Planning Act, the Territory Plan and the National Capital Plan, what would need to change, and how similar matters have fared at the tribunal.
Built from the ACT’s public planning documents, and we’re testing it with planning professionals now. Its reports are reviewed drafts, not legal advice.
Describe the support and the participant’s plan. It reports whether it can be claimed, under which item, at what price limit, and what evidence the claim needs, citing the Pricing Arrangements line by line.
We’re testing it with providers now.
Your rulebook next. If your sector runs on a body of rules many organisations share, tell us which one.
What you want from a grant application is the grant. Every hour your team spends wrestling the guidelines is an hour not spent on the work the grant is for. So everything we do is aimed at giving your application the strongest chance of being funded, with less of your team’s time.
We call it Assessor Ready. That’s the test every application has to pass before it goes in: will the person scoring it find everything they’re looking for?
A funder asks the same four questions any stranger does: is this for us, what does it get us, why should we believe it, and what happens next. Assessor Ready makes sure your application answers all four.
We can write the application with you, or set your team up to run all four steps themselves.
What changes: no one can promise a funder’s decision. Assessor Ready makes sure your application isn’t lost on anything you could have controlled. Your past applications become the starting point for every round after, so the second application starts further along than the first.
Right now someone is typing your problem into Google or ChatGPT, in their own words. If your website describes your work only in your sector’s words, they get sent somewhere else. In the paper’s test of 84 organisations, asked in people’s own words, ChatGPT recommended the organisation set up to help first only about a third of the time.
We find the words the people you serve actually use, starting with whoever answers your phone. Then we rewrite your website so the first page each kind of visitor lands on answers their four questions in those words: is this for someone like me, what does it get me, why should I believe it, and what do I do next. We check what Google and AI assistants say about you before we start and after we finish.
What changes: the parent, the carer or the business owner with the problem you solve finds you, recognises themselves on the first page, and knows what to do next.
One motor trades association in the test was recommended first when ChatGPT was asked for its own directory of repairers who follow a code of conduct. Asked “How do I find a mechanic I can trust?”, the way a car owner asks, it didn’t come up at all. The fix is to say both.
Buying AI is a workflow decision, not a technology one. The question is where your people lose hours to work the right set-up could carry.
A general AI tool leaves your people to work out what to ask and whether the answer’s any good. Set up for one piece of work, it already knows your format, your standards and your rules. A licence is a pantry full of ingredients; what changes a team’s week is the meal kit.
Hours saved are a measure. What they turn into, in your budget holder’s own terms, is the result.
Your team signs off anything drafted in your organisation’s name, and public-facing tools show their sources.
The method maps to the six practices in the Australian Government’s Guidance for AI Adoption, and we flag anything near the new privacy transparency line before it’s built.
Training on your own work, a named owner for each piece, and a check at 90 days. If something isn’t working at 90 days, we work out why with you and fix what’s ours to fix.
Five of the ten leaders in the paper raised resourcing, and several run small teams. Something your team can keep current does more than something big nobody maintains.
The map is yours to keep, whether or not we ever work together.