No stale extractions
Extracted document data and research are cached with a version stamp, so re-running preparation neither redoes work nor reuses stale results.
The problem it solves
Re-running preparation should be quick, but never use outdated extraction.
How it works
- Run preparation.
- Extractions are cached with a version stamp.
- Re-runs reuse only current results.
What your practice gets
- Quick re-runs without stale data.
- Version-aware.
- Automatic.
Questions accountants ask
When is cached data reused?
Only when it matches the current version.
Can I force a fresh extraction?
Changing the source or version refreshes it.
Does it save AI usage?
Yes, by not redoing current work.
Related features
Features that work alongside no stale extractions.
- Workpapers from client documents
Build a workpaper from the data extracted and classified from the client's documents.
Video - Maintain workpapers from the interface
Add lines, bind a workpaper to a client and list workpapers from the interface as well as the chat.
- Workpaper sync with Fedix
Workpaper edits sync between your computer and Fedix, merging changes from both sides.
- Learns your workpaper habits
FediPilot records your edits and comments and recalls your habits next time.
- Prepare Workpapers
One command starts workpaper preparation for the active client and financial year.
- See the source of any figure
Click a number in the workpaper to open the original document it came from.
See it working in your practice.
Once you have subscribed, you can download FediPilot for Windows from your Fedix dashboard and sign in with your Fedix account. FediPilot runs on your own ChatGPT or Claude plan, which Fedix does not supply and does not mark up. Training sessions are included and can be booked at any time; they are not a precondition.
There is no free trial: you pay month one on the day you subscribe. Month one refunded if it isn't working after onboarding. Plans are month-to-month with no minimum term. Pay annually up front and get 2 months free. See pricing.