Describe a complaint. The finder compares it with published AFCA determinations and shows how similar complaints were decided, and why.
Describe a complaint in plain English. FincomplaintIQ compares it with 1,397 published AFCA determinations and shows how similar complaints were decided, the likely outcomes and awards, and why.
Designed for complaints, IDR and legal teams at financial firms. Created by Serhat Saylak.
A customer disputes a storm claim declined as “wear and tear”.
Each analysis simulates AFCA deciding the complaint 100 times, using how similar determinations were actually decided and what AFCA actually awarded, with its compensation caps applied.
The site is in private preview while it's being tested, and access is by invitation. If you've been sent an access link, open it, or enter your code here.
Contact the creator, Serhat Saylak.
LinkedIn ↗No. It's general information and research based on published determinations. Decisions on a complaint remain with the firm, under its own IDR process, with legal advice where needed.
No. It's an independent project and isn't operated or endorsed by the Australian Financial Complaints Authority. It works from the determinations AFCA publishes.
No. AFCA decides each complaint on what's fair in all the circumstances and isn't bound by earlier decisions. The finder shows how AFCA has approached similar complaints and the range of outcomes that suggests.
The description is sent to Anthropic's Claude AI to produce the analysis and isn't stored on this site's server. Saved analyses stay in the user's own browser. Customer names, account numbers and other personal information should be left out.
Include the product, what the firm did or didn't do, dates, amounts, what you've asked for, and anything the firm has said in response.
Each determination tested uses one analysis' worth of your Claude usage.
Pick a topic. The chart shows how the determinations in the library were decided each year. Then have Claude compare the reasoning in older and newer determinations, including FOS and CIO decisions from before AFCA began in November 2018.
Which financial firms appear most in the library, and how often AFCA sided with the complainant in full or in part. Click a firm to open its profile.
These figures only cover the determinations in the library, not every complaint made to AFCA. Most complaints are resolved before a determination, and bigger firms have more customers, so a higher count doesn't mean a firm treats customers worse. Rates based on only a few determinations can swing a lot; the tooltip shows a likely range. For official complaint volumes by firm, see AFCA's published complaint data.
Drop in determination PDFs, or a whole folder of them. There's no practical limit: the library holds tens of thousands. Each PDF is summarised into a searchable record with its product, issues, outcome, remedy, deciding factors and key paragraphs quoted word for word. Download PDFs from AFCA's published decisions search.
Each PDF uses some of your Claude usage, so very large batches may pause at your usage limit. Choose Resume to carry on. Files already in the library are skipped. Add files in one tab at a time.
Saving determinations from AFCA's site: open the determination, press Ctrl+P, and set the printer to Save as PDF, not “Microsoft Print to PDF”. That keeps the text, so it's faster and quoted paragraphs can be checked. PDFs saved as pictures still work: the page reads them itself, which takes a few extra seconds per page. You can also save the page with Ctrl+S (.html), or copy and paste it below.
Anyone who opens this site can search the library and run analyses. Your own analyses and the facts you type are kept only in this browser.
FincomplaintIQ compares the facts of a financial complaint with published determinations of the Australian Financial Complaints Authority (AFCA), and summarises how AFCA has approached similar complaints: which way they went, the factors that decided them, and what AFCA ordered.
This site is an early version that's still being built and tested. That means:
While it's being tested, the tools are open by invitation. If you'd like to try it, ask for an access code on LinkedIn.
Feedback on anything that looks wrong is very welcome. See About the creator below.
This site was created by Serhat Saylak as an independent project. Feedback, corrections and suggestions are welcome.
Last updated 26 September 2026.