evidoria

← Back to browse

Good practice Imported

Israel Tax Authority — AI cross-referencing of capital declarations to close the tax gap

Israel · Jerusalem · See the Israel profile · See the Jerusalem profile

Evidence: Descriptive / self-reported Top 66% 53/100 · Ask Evidence Copilot about this practice

Israel's Tax Authority deployed AI to cross-reference capital declarations against financial data, identifying NIS 20 billion (~$7 billion) in discrepancies from a 4% filing sample. Plans to scale to 100% of all returns.

20 NIS billion (~$7 billion)
Capital declaration discrepancies identified (as of June 2026)
4 %
Share of tax filings scanned by AI system (2026)
560 NIS billion (up from 540)
Revised 2026 national revenue forecast (2026)

Details

Maturity
Scaling
Promoter
Israel Tax Authority (Rashut haMisim be-Yisrael)
Period
2024-present
Keywords
tax gap analysis, AI compliance, capital declaration, cross-referencing, enforcement

Context

The Israel Tax Authority uses AI to cross-reference capital declarations filed by self-employed individuals and business owners against financial records already held by the Authority, including bank and investment data, in order to detect unexplained wealth gaps. The system currently scans about 4% of all filings, with the Authority stating an intention to move toward reviewing 100% of reports. The initiative is aimed at closing the tax gap without raising taxes on the middle class.

Objectives

Identify unexplained wealth by systematically comparing taxpayers' capital declarations with other financial data held by the Tax Authority, and expand this AI-based scanning from a small sample to full coverage of filings.

Activities

AI-based tools compare taxpayers' capital declarations against bank records, investment data and other financial reports held in Tax Authority systems to flag inconsistencies for further investigation.

Results

By June 2026, the AI cross-referencing exercise had identified discrepancies totalling NIS 20 billion (about $7 billion) from the roughly 4% of filings scanned so far. The Tax Authority partly credited anticipated enforcement yield from this and related efforts when it raised its 2026 national revenue forecast from NIS 540 billion to NIS 560 billion.

Conclusions

The identified discrepancies are investigative leads rather than confirmed additional revenue collected, and no independent technical evaluation or audit of the AI system has been published; the Authority's stated next step is to expand scanning coverage from 4% toward 100% of filings.

Implementation
Indicative cost
Medium (€50k–€500k) — No public budget figures found; described only as an internal AI capability within the Israel Tax Authority's existing systems.
Time to results
Medium (1–3 years) — Deployed and scanning ~4% of filings as of mid-2026, with a stated intention to expand to 100% coverage; no fixed end date given.
Staffing & skills
Israel Tax Authority enforcement/data-analytics teams (led by Director Shay Aharonovich); no detailed staffing breakdown published

Conditions for success

  • Access to integrated bank and investment data alongside capital declarations
  • Political commitment to expand scanning coverage without raising middle-class tax rates

Common failure modes

  • Discrepancies flagged by the system are unconfirmed leads and may overstate actual recoverable revenue
  • No independent technical audit exists to verify accuracy or false-positive rates

Where it fits

Governance type
national tax authority
Scale
national
Income level
high income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Similar practices you may find useful