Auto-fix Suggestions (Safe, Suggested, Bulk)

Spend less time correcting documents, more time closing books.

🧠 How Auto-Fix Works

TaxLayer doesn’t just point out problems — it helps fix them. The auto-fix engine looks at your documents, compares them with vendor/client profiles, myDATA rules, and your past approval patterns, then suggests or applies corrections.

There are three levels of autofix:

  • 🟢 Safe Fixes → Automatically corrected (no risk).

  • 💡 Suggested Fixes → High-confidence changes you approve.

  • 📦 Bulk Fixes → Apply corrections across dozens of documents at once.

Every change is logged, reversible, and validated again — so you stay in control.


🟢 Safe Fixes (Auto-Apply)

These are the “no brainer” corrections applied instantly:

  • Date formats (DD/MM/YYYY ↔ YYYY-MM-DD)

  • Cleaning up text (extra spaces, wrong capitalisation)

  • Known defaults (EUR, common tax rates, saved VAT numbers)

  • OCR misreads (Greek Ο vs zero, Ι vs 1)

  • Obvious field mappings (invoice number → invoiceNo field)

🔎 Pain point solved: 60–80% of common formatting errors disappear before you even see them.

Where you see them:

  • Batch overview → “Auto-fix summary”

  • Document detail → “Auto-applied fixes” section

Controls:

  • Adjust aggressiveness (Conservative / Standard / Aggressive) in Settings.

  • Rollback any auto-fix for up to 30 days.


💡 Suggested Fixes (Your Review Needed)

When the system is confident, but not 100%, it asks for your approval.

Examples:

  • “Issuer VAT missing → use vendor profile VAT [999123456]?”

  • “Totals don’t add up (€1,234 vs €1,244) → adjust VAT line?”

  • “Classification missing → apply ‘category 1.1’ based on vendor type?”

Your options:

  • Approve once (fix this document).

  • Apply to similar docs (pattern-based).

  • Multi-select and apply in bulk.

🔎 Pain point solved: Fix clusters of errors in seconds instead of hunting document by document.


📦 Bulk Fixes (Clear Errors in One Sweep)

When patterns appear, TaxLayer groups them and suggests a bulk fix.

Example flow:

  1. Detection → “15 invoices missing VAT classification.”

  2. Suggestion → “Apply Materials 1.1 to all?”

  3. Preview → Shows which documents will change.

  4. Apply → Fixes all in one go + auto re-validation.

  5. Results → Success/failure breakdown per document.

Bulk categories covered:

  • Vendor-specific patterns (missing VAT, known categories).

  • Layout issues (same PDF template → same fix).

  • Compliance fixes (VAT rates, required fields, myDATA codes).

Advanced options:

  • Confidence thresholds (“apply only if >90% sure”).

  • Staged rollout (try 5 docs, then the rest).

  • Bulk rollback (undo the entire operation if needed).

🔎 Pain point solved: Hours of manual corrections reduced to minutes.


🎚 Confidence & Control

Every suggestion has a confidence score:

  • 95–100% → Auto-applied.

  • 80–94% → Suggested for approval.

  • 60–79% → Medium confidence, flagged for careful review.

  • <60% → Manual only.

Safety nets:

  • Rollback anytime.

  • Full before/after logs for audit.

  • Manual override always available.


🧭 Real Example — Vendor Classification Fix

  • Scenario: 50 invoices from your office supplier missing classification.

  • Detection: System flags “Materials 1.1” as likely category.

  • Confidence: 92% (vendor profile + line items like “pens, paper, supplies”).

  • Suggestion: “Apply Materials 1.1 to all 50 documents?”

  • You approve → System applies and revalidates → All 50 green in 30 seconds.


💡 Pro Tips

  • Start conservative with autofix thresholds → loosen as you build trust.

  • Focus on high-volume vendors first → biggest payoff.

  • Always revalidate after bulk operations.

  • Export logs before major changes for compliance comfort.

  • Use AI Chat with autofix for “why” + “how to fix” explanations.


👉 With auto-fix, validation goes from “error hunting” to “error clearing.” You stay in control, but the system does the heavy lifting — so month-end closes are faster, cleaner, and less stressful.

Last updated