CallNote is an AI call-notes tool we built at Workvolve. It turns a client phone call into an accurate, compliant file note automatically, for Australian mortgage brokers, financial advisers and law firms who otherwise lose hours a week writing notes by hand. This is the short story of why we built it and how it works: a look at the kind of thing we build.

The problem we kept seeing

In regulated industries, every client call needs a file note: the record of what was discussed, what was advised, and what happens next. It is not optional, and it is not quick. Brokers and advisers taking ten or more calls a day were spending an hour or two just writing up notes, usually at the end of the day when the detail had already faded. Thin notes, lost time, and a quiet compliance risk. The same pattern showed up everywhere we looked.

What we built

CallNote takes the call and does the writing. The conversation is captured with consent, transcribed, and structured into a file note in the format the adviser already uses: discussion points, decisions, and follow-up actions. The adviser reviews and approves; the note files itself against the right client. The judgement stays with the human, the typing goes away.

The build had to clear a few real bars:

  • Accuracy. A file note that is wrong is worse than none. The structuring had to be reliable and reviewable.
  • Compliance. Consent captured at the start of every call, records stored securely, and the AI configured not to train on client data.
  • Fit. It had to slot into how brokers already work, not force a new way of working on them.

How we approached it

We built CallNote the way we build everything at Workvolve: start with the single most painful, most repetitive task, get that one thing working properly, and prove it before adding more. We tested it against real call workflows, tightened the accuracy, and got the consent and storage side right before anything else.

Ben built every part of it himself, from a blank page to a live multi-tenant product: Next.js, React and TypeScript on Supabase and PostgreSQL, authentication, Stripe subscription billing, connections to Dialpad, Aircall and Microsoft Teams, and a Claude-powered notes engine with a human review step. He built the marketing site and the onboarding too. Custom builds for clients follow the same approach, with one difference: CallNote is Ben's own product, while a client build runs in the client's own accounts and the client owns all of it, with no lock-in.

What we learned building it

A few things came out of the build that shape how we approach every project:

  • The boring task is the valuable one. Note-writing is not glamorous, but it was the thing costing brokers the most time. The unglamorous, repetitive jobs are almost always where the real return is.
  • Compliance comes first. In regulated work, getting consent and data handling right decides whether people can actually use the tool at all.
  • Keep the human in the loop. The adviser approving every note is the point: AI drafts, the human owns the record.

What it means for our clients

CallNote is a product now, live at callnote.com.au. The reason it matters here is what it shows: this is the standard of build we bring to custom work, from a single automation to full custom software. Whether it is call notes, quoting, invoicing, reporting or lead follow-up, the approach is the same: find the task that is quietly costing you, automate it carefully, keep the judgement human, and hand you something you own outright.

If you are a broker, adviser or law firm, take a look at CallNote directly. If you have a repetitive job of your own that is eating hours, that is exactly what we build, so book a free 30-minute call. We will look at how your week runs and tell you straight what is worth automating, and what is not.