Implementation & training for RoboSystems
We stand up RoboLedger on your books and train the people who already sign off, so you close, report and plan on books you own. Investors get their portfolio companies’ filed reports as data, not PDFs. AI drafts. Your controller approves. The platform is RoboSystems: open source, and yours.
Not ready for a call? and we’ll come back within one business day.
Run it across your client book, put your own team on it, or get your portfolio reporting as data. Whichever door, your people hold the pen and we teach them the platform.
Run RoboLedger across your whole client book. We implement the first client with you and train your staff; you run close and reporting for every client after that. Your clients stay yours.
A controller or accountant who already signs off on the books. We implement RoboLedger through one real period with your people on the buttons, then train them to run it. We do not become your controller.
Funds, family offices and holding companies with an interest in many companies. Each company keeps its own books on RoboLedger, run by its own accountant, and their filed reports land in your graph as data, not PDFs. The investor side is early, and we say so: it is where reports land and positions live, not yet where they are valued. The first funds to do this do it with us.
A live demo first, at no cost. Then fixed-scope work that ends with your people running the platform and us out of your graph.
Thirty minutes on the real system: a complete set of books, closed and reported, walked through end to end. Costs nothing and touches none of your data. This is how we start.
Optional and fixed-fee, once a demo says it’s a fit. We take apart your current close and write down where the platform fits and where it doesn’t, so the implementation is pointed. Platform or not, you keep the plan.
RoboLedger on your graph: QuickBooks connected, your chart of accounts mapped to the reporting vocabulary, and one real period closed by your person with us beside them. Fixed fee per graph, stepped for the second client onward.
Close, Report and Plan for the people who sign off; the platform inside the chat they already use; the close playbook. Per cohort for a practice, per seat for a team.
We leave the graph. Your people hold every role, our access ends, and nobody from Harbinger keeps a key. The second period is yours alone, and it should be.
When it is the job: adapters (QuickBooks today; bank feeds, ERP, billing, CRM and custom APIs land as engagements), bespoke reporting flows, and deployment in your own AWS account. Through the public API, with a key you issue and can revoke.
If targeted tooling solves it, we won’t sell you a platform. If you don’t need one at all, we’ll say so.
How engagements work
A live demo comes first and costs nothing. The ladder starts once we both think it’s a fit.
The platform on a full set of books. Free.
Optional, fixed-fee: your current close, taken apart and written up
One real period on RoboLedger, your people on the buttons
Your controllers and client-facing staff run Close, Report and Plan
We leave the graph. Build work stays available; a controller on retainer does not.
Fixed fees, stated up front. No hourly billing, and no retainer for the close.
Most AI-accounting firms sell a black box and promise autonomy they can’t deliver. We do the opposite - because when it’s your books, trust is the product.
See the platform on GitHubThe platform is open source. Read the code. See exactly how every number is produced, with provenance back to the policy behind it.
Open formats and a forkable platform. Take it in-house and self-host whenever you like. You will never be stranded by a vendor shutting down.
AI generates, your people approve. We don’t sell an autonomous accountant that quietly breaks on the hard cases, and we don’t sell a controller on retainer.
Nobody from Harbinger posts, approves or closes a period on your books. Sign-off is yours, by name, and our access ends at handover.
Run fully managed on our service, or in your own AWS account. Same open platform either way. When it is yours, we stay out of it - support and integration work happen through the public API, with a key you issue and revoke.
If targeted tooling solves it, we won’t sell you a platform. We’re accountants who deploy AI to fit the problem - and we’ll tell you what you don’t need.
Harbinger FinLab is led by Joseph T. French, an inactive CPA who’s spent his career at the forefront of financial data. At Workiva, he was a principal engineer building innovative financial systems - and an entrepreneur who’s founded multiple companies and served on boards.
So this isn’t a chatbot bolted onto accounting. We build on RoboSystems - our own AI-native financial platform - and run our own books on it: an AI-drafted close, a human controller’s sign-off, GAAP-grade statements. It’s how we run our business, and it’s what we teach.
Proven beyond accounting
Supply FlowQC - an AI supply-chain resiliency product for healthcare - runs on the same RoboSystems platform that runs our books, proving it models any domain, not just accounting.
Start with a live demo - a complete set of books, closed and reported, walked through end to end. If it’s a fit, implementation is next, and then you run it.
Not ready for a call? and we’ll come back within one business day.