Statistical, finance and insurance clerks
Almost every task can be done with generative AI, with little uncertainty. This is where the change is biggest — and where switching makes most sense.
Exposure is not job loss: it shows how much of the work generative AI could do today. Whether that turns into transformation, productivity gains or cuts depends on employers, regulation — and on you.
0 = today’s generative AI cannot perform any of the job’s tasks, 100 = it could do all of them fully. The pale band is the spread across tasks.
Work out your own score ↓What can you do?
🤖 Written with AIThis text was written by artificial intelligence from the data points shown; it is built on verified figures but may contain mistakes.· 11/10/2026How is it made?
With an exposure score of 64/100, clerks face substantial AI integration in claims, statistical calculation, and transaction records. However, steady replacement demand in Europe indicates enduring needs for nuanced policy interpretation, complex claims review, and client-facing oversight.
What changes
- Automating routine insurance enrolments, policy updates, claims processing, and standard premium transactions.
- Compiling statistical and actuarial data automatically from distributed routine sources into structured datasets.
- Generating routine tables, calculating averages and totals, and reconciling standard securities transaction records.
What stays human
- Evaluating ambiguous, contested, or potentially fraudulent insurance claims requiring contextual human judgment.
- Validating actuarial and statistical anomalies that automated calculation engines misinterpret or overlook.
- Communicating sensitive policy changes and financial discrepancies directly to distressed clients or brokers.
What to learn
- Auditing automated financial tables and statistical models for data integrity and regulatory compliance.
- Mastering specialized insurance workflow platforms and AI-assisted claims triage tools.
- Developing advanced client advisory communication for disputed transactions and complex coverage terms.
Task by task
The ILO scored every task of the occupation separately, as the average of two AI models (GPT-4o and Gemini). The pale band is the range between the two models' scores, the dot the ILO's final score (for a few tasks expert review placed it outside the band).
- Processing insurance enrolments, cancellations, claims transactions, policy changes and payments67
- Calculating totals, averages, percentages and other details and presenting them in the required tabular form65
- Obtaining and compiling statistical or actuarial data based on routine or special sources of information65
- Maintaining records of bonds, shares and other securities bought or sold on behalf of clients or employer.64
- Preparing financial documents, and calculating interest or brokerage charges and stamp duties payable60
How much does it affect you?
Every job is different: mark which of these tasks you do and how much, and we'll compute your own, time-weighted score.
For each task: I don't do it · sometimes · most of my job.
- Processing insurance enrolments, cancellations, claims transactions, policy changes and payments(67)
- Calculating totals, averages, percentages and other details and presenting them in the required tabular form(65)
- Obtaining and compiling statistical or actuarial data based on routine or special sources of information(65)
- Maintaining records of bonds, shares and other securities bought or sold on behalf of clients or employer.(64)
- Preparing financial documents, and calculating interest or brokerage charges and stamp duties payable(60)
You are exactly at the occupation average. (The occupation's score: 64.)
About 100% of your working time goes to highly exposed tasks (score ≥ 50).
- • Processing insurance enrolments, cancellations, claims transactions, policy changes and payments
- • Calculating totals, averages, percentages and other details and presenting them in the required tabular form
- • Preparing financial documents, and calculating interest or brokerage charges and stamp duties payable
- • Maintaining records of bonds, shares and other securities bought or sold on behalf of clients or employer.
An estimate from the ILO task scores. Everything is computed in your browser — nothing is sent or stored.
How many people work in this occupational group?
Share of the Numerical and material recording clerks group in total employment, by country (Eurostat, ISCO 2-digit — 4-digit breakdowns are not public by country).
Table view
| Country | Numerical and material recording clerks |
|---|---|
| Belgium | 4.6% · 238 thousand · 2025 |
| Croatia | 4.4% · 75 thousand · 2025 |
| Slovakia | 4.3% · 112 thousand · 2025 |
| Netherlands | 3.8% · 372 thousand · 2025 |
| Germany | 3.6% · 1,504 thousand · 2025 |
| Estonia | 3.5% · 25 thousand · 2025 |
| Czechia | 3.1% · 164 thousand · 2025 |
| Slovenia | 3.1% · 31 thousand · 2025 |
| Sweden | 3.0% · 160 thousand · 2025 |
| Poland | 3.0% · 519 thousand · 2025 |
| Spain | 3.0% · 666 thousand · 2025 |
| Italy | 3.0% · 719 thousand · 2025 |
| Switzerland | 2.9% · 143 thousand · 2025 |
| Cyprus | 2.9% · 15 thousand · 2025 |
| Austria | 2.9% · 128 thousand · 2025 |
| Norway | 2.8% · 81 thousand · 2025 |
| France | 2.4% · 707 thousand · 2025 |
| Malta | 2.4% · 8 thousand · 2025 |
| Bulgaria | 2.3% · 67 thousand · 2025 |
| Hungary | 2.3% · 106 thousand · 2025 |
| Latvia | 2.2% · 19 thousand · 2025 |
| Ireland | 2.2% · 61 thousand · 2025 |
| Denmark | 2.1% · 64 thousand · 2025 |
| Lithuania | 2.0% · 30 thousand · 2025 |
| United Kingdom | 2.0% · 669 thousand · 2025 |
| Luxembourg | 1.9% · 6 thousand · 2025 |
| Portugal | 1.8% · 94 thousand · 2025 |
| Romania | 1.8% · 137 thousand · 2025 |
| Greece | 1.6% · 70 thousand · 2025 |
| Finland | 1.4% · 35 thousand · 2025 |
| Iceland | 0.5% · 1 thousand · 2025 |
What do they earn?
Mean gross monthly earnings in the major occupational group (Clerical support workers). Eurostat's earnings survey only publishes pay by occupation at 1 digit, so this is the job's broader group. Enterprises with 10+ employees, 2022.
By country, at purchasing power parity (PPS/month), so countries with different price levels can be compared.
- Switzerland4,326
- Luxembourg3,358
- Denmark3,338
- Belgium3,291
- Norway3,160
- Germany2,927
- Ireland2,900
- Iceland2,825
- Austria2,768
- Netherlands2,680
- EU272,590
- Italy2,521
- Sweden2,451
- Finland2,386
- Spain2,121
- Slovenia2,111
- Malta2,089
- Lithuania2,049
- France2,018
- Croatia1,847
- Greece1,837
- Poland1,796
- Romania1,755
- Estonia1,740
- Czechia1,725
- Cyprus1,685
- Hungary1,682
- Latvia1,617
- Slovakia1,583
- Portugal1,421
- Bulgaria1,203
Source: Eurostat Structure of Earnings Survey (earn_ses22_21), every four years. The United Kingdom is not covered by the survey.
Outlook to 2035
Employment in the Numerical and material recording clerks group according to Cedefop's 2026 forecast.
thousand · Cedefop Skills Forecast 2026 · 2026–2035: projection
Replacement demand is the jobs left by people retiring or changing careers — even a shrinking occupation can have many openings.
Related occupations
Sources: ILO WP140 (2025) · Eurostat lfsa_egai2d · Cedefop Skills Forecast 2026 · WEF Future of Jobs 2025 · ESCO. The exposure score measures how much of an occupation's tasks generative AI could perform — not whether the job disappears. For information only, not career advice. How do we calculate this?