Think twice before you act on AI financial advice.
Paying in instalments when shopping online, money tips from influencers, answers from ChatGPT and similar tools: FinShield helps you check this kind of advice before you spend money, take out a loan or invest.
FinShield does not provide financial advice. It supports your own judgment: what is missing, what is assumed, what it costs, and when to ask a professional.
A ZHAW research project
FinShield is developed and scientifically evaluated at the ZHAW School of Management and Law, together with Schuldenberatung Kanton Zürich and iconomix, the Swiss National Bank’s education programme. The project is funded by the Digitalization Initiative of the Zurich Higher Education Institutions (DIZH).
- Institution
- ZHAW School of Management and Law
- Duration
- 12 months
- Funding
- DIZH Rapid Action Call 2026.1
- Practice partners
- Schuldenberatung Kanton Zürich, iconomix
Fluent, personal and possibly wrong
Financial decisions increasingly happen inside apps, chatbots and social feeds. AI tools produce guidance that sounds neutral and tailored, but often know very little about you: your income, your debts, your goals or how much risk you can afford.
An investing app…
…suggests putting your savings into risky shares without knowing that you are still paying off an expensive loan.
An online shop’s chatbot…
…assures you that you can easily afford to pay in instalments, without knowing about your other bills and commitments.
On top of that, we are all influenced by how an offer is presented, by wanting things right away, or by an answer that simply sounds competent. We also often trust computers more than we should. And money worries add stress, which makes it harder to weigh things up calmly.
Debt is common, and help comes late
of the Swiss population lived in a household with at least one form of debt in 2022.
BFS, 2024
lived in a household with at least one payment arrear.
BFS, 2024
new cases at Swiss debt counselling services in 2024. Almost half had been in debt for five years or more before seeking help.
Schuldenberatung Schweiz, 2025
of people in counselling said that feeling overwhelmed, for example by bills, forms and paperwork, contributed to their debt.
Schuldenberatung Schweiz, 2025
Two modes, one way of thinking
Most AI finance tools try to give faster answers. FinShield flips that role: it uses AI to help you interrogate the answers you get.
Practise with everyday examples
Practise in a safe space. Work through realistic situations and learn to spot what is missing, what is hidden and what to ask next. After each example you get feedback on what you did well and what you missed.
Pay in 4 interest-free instalments, only today!What are you not being told?
Help when it matters
Facing a real decision? Copy an offer, a piece of advice or an answer from ChatGPT and similar tools into FinShield. It walks you through the important questions step by step, and tells you when it is better to ask a professional.
What FinShield helps you check
- Missing information
- Hidden assumptions
- Costs & fees
- Risks
- Alternatives
- Who earns money from it
- Pressure and sales tricks
- The right follow-up questions
- When to seek professional help
🧩 Take it with you: checklists for AI chats
FinShield’s questions are also available as ready-made text templates (called “Financial-AI Judgment Skill Templates” in the project). Paste them into AI chats like ChatGPT so they answer your money questions more critically and completely.
Built for people facing real financial decisions
Young adults
People managing their own money for the first time quickly run into pay-later offers, loans via apps, subscriptions, money tips from influencers and AI answers, often before they have settled habits.
Households under pressure
Adults who are short of money, have a lot on their plate or lack easy access to advice, yet still have to judge online offers and advice.
Educators & prevention actors
Schools, debt counselling services and other organisations looking for realistic practice examples and teaching materials.
What users learn and what the project delivers
With FinShield you learn three things: spotting warning signs such as time pressure (“today only!”), “everyone is doing it” or hidden costs; asking the right questions, for example whether you can really afford something and what the risks are; and knowing when you need help. We study scientifically whether this works: participants solve practice tasks before and after using the tool. What counts is whether they decide better afterwards, not just whether they like the tool.
Web prototype
A user-tested, web-based training and reflection tool.
Practice examples
Realistic everyday situations, developed with debt counsellors and educators.
Checklists for AI chats
Text templates that let you use FinShield’s questions in AI chats like ChatGPT.
Documentation
Guides for teachers, counselling services and debt-prevention organisations.
Grounded in prevention and education
Our partners cover both ends of prevention: the realities of over-indebtedness and the financial education that comes before harmful decisions accumulate.
Schuldenberatung Kanton Zürich
Debt counselling & prevention
This non-profit organisation helps people in debt and works to stop debt problems before they start. It makes sure FinShield reflects real everyday situations, shares its knowledge of common mistakes and risks around budgeting, loans and debt, helps choose the examples, and checks that the tool is easy to understand and use without prior knowledge.
Visit website: Schuldenberatung Kanton Zürich →iconomix
Education programme of the Swiss National Bank
The Swiss National Bank’s programme for teaching economics and finance at upper secondary and vocational schools. iconomix already offers teaching material on loans, paying in instalments, manipulative website design and the psychology of money decisions. It helps make FinShield easy to learn with and advises on how to use it in the classroom.
Visit website: iconomix →Funded by the DIZH Rapid Action Call
FinShield is funded by the Digitalization Initiative of the Zurich Higher Education Institutions (DIZH) as part of the 5th Rapid Action Call “Strengthening resilient and democratic communities” (2026.1) of the DIZH Innovation Programme.
Financial vulnerability weakens people’s sense of control, social participation and even political engagement. By helping people stay active decision-makers in AI-mediated financial environments, FinShield contributes to more equitable financial resilience, for individuals and communities alike.
Who is behind FinShield

Prof. Dr. Kremena Bachmann
Institute for Personal Finance & Wealth Management, ZHAW School of Management and Law
Researches behavioural and household finance, financial decision making, financial advice and retirement decisions. Co-author of “Behavioural Finance for Private Banking”.

David Jaggi
Associate Researcher, Institute for Asset Management, ZHAW School of Management and Law
Works on asset and risk management, machine learning and data science.
Frequently asked questions
Does FinShield give financial advice?
No. FinShield does not recommend products or tell you what to do. It helps you check whether a recommendation, offer or AI-generated answer is trustworthy, complete, transparent and appropriate for your situation before you act on it.
What if I am already in financial difficulty?
FinShield points you to professional help when it matters, for example if you can no longer pay your bills on time, have expensive debts or face a decision with big long-term consequences. In the canton of Zurich, Schuldenberatung Kanton Zürich offers free and confidential counselling.
What does “buy now, pay later” mean?
Many online shops let you pay for a purchase later or in instalments. It feels convenient and often looks free. But with several such purchases running at once, or one missed payment, it is easy to lose track, and late fees can pile up.
What are the AI-chat checklists (“Financial-AI Judgment Skill Templates”)?
Ready-made text templates containing FinShield’s questions. You paste one into an AI chat like ChatGPT before asking a money question. The AI will then ask about what it does not yet know about your situation, and point out risks, costs and alternatives.
Can schools or counselling services use FinShield?
Yes, that is one of the goals. Practice examples, checklists and guides are designed for use in upper secondary and vocational schools, in debt counselling and in consumer protection.
When will the tool be available?
The project runs for 12 months. A first version of the tool is planned for months 3–5, and the tested, finished version with all materials for the end of the project.
Interested in collaborating or using FinShield in your organisation? Get in touch with Kremena Bachmann or David Jaggi.