Tau
AI for bankers · 11 lessons

AI for bankers: speed up client and credit work safely

Use AI to draft client briefs and credit write-ups, digest earnings calls and speed up customer checks, with controls your risk team will accept. For bankers, credit analysts and bank operations staff.

Who it’s for: Retail, commercial and investment bankers, credit and risk staff.

Banks already use AI across retail, commercial and investment banking, from fraud checks to research. Generative AI adds something new for individual bankers. It can draft a credit memo section, summarise an earnings call, or pull a client brief together from annual reports and your call notes. The catch is that every figure it writes must reconcile to the source.

This course starts with what data must never leave the bank. You'll classify customer data, know-your-customer files and material non-public information, and match each to approved tools. Then you'll prepare a one-page client meeting brief, draft a credit memo section and reconcile it to the spreads, and recompute covenant ratios using the facility agreement's own definitions.

Later lessons cover comps and precedent transactions for a pitch book, and faster KYC and anti-money-laundering reviews in which the analyst still makes the decision. They also cover model risk rules for AI. You'll also apply fair-lending duties, such as the US Equal Credit Opportunity Act and the EU AI Act's credit-scoring rules. The course ends with maker-checker controls that pass risk and audit review. Under these controls, a second person checks the work.

By the end you’ll be able to…

  • Draft a credit memo section from financials and verify every figure
  • Build a sourced first pass of comps and precedent transactions
  • Explain how model risk and fair-lending rules apply to AI where you work
  • Set maker-checker controls and an audit trail for AI-drafted credit work

Short on time? Start with the essentials

These 4 lessons cover the most useful ideas in this course. Each one takes about a minute.

Essential 1

See where AI already works inside banks

Describe the main AI uses across retail, commercial and investment banking and which ones touch customers or decisions.

Essential 2

Know what customer data must never leave the bank

Classify customer data, KYC files and MNPI, and decide which AI tools are approved for each.

Essential 3

Prep a one-page client meeting brief

Combine annual reports, recent news and your own call notes into a client brief with questions to ask.

Essential 4

Draft a credit memo section from the financials

Have AI draft the business and cash-flow analysis of a credit memo, then reconcile every figure to the spreads.

Every lesson in this course

This course has 11 video lessons of about a minute each, from the basics to expert habits. Each one ends with a question, three takeaways and its sources.

Foundations 2 lessons

  1. See where AI already works inside banksEssential

    Describe the main AI uses across retail, commercial and investment banking and which ones touch customers or decisions.

    “JPMorgan gave nearly 250,000 staff an in-house AI assistant. What does yours do?”

  2. Know what customer data must never leave the bankEssential

    Classify customer data, KYC files and MNPI, and decide which AI tools are approved for each.

    “Account numbers, KYC files, inside information: three things no public chatbot should see.”

Building skills 4 lessons

  1. Prep a one-page client meeting briefEssential

    Combine annual reports, recent news and your own call notes into a client brief with questions to ask.

    “Annual report, last call notes and news: one-page client brief before your coffee cools.”

  2. Draft a credit memo section from the financialsEssential

    Have AI draft the business and cash-flow analysis of a credit memo, then reconcile every figure to the spreads.

    “AI can draft the cash-flow section fast, and get EBITDA wrong just as fast. Catch it.”

  3. Digest an earnings call for credit and coverage

    Summarise an earnings call into changes in leverage, liquidity, guidance and covenant headroom, quoting the transcript.

    “A 90-minute earnings call → 5 lines on leverage, liquidity and guidance changes.”

  4. Check a covenant compliance certificate

    Have AI recompute covenant ratios from a borrower's compliance certificate and financials using the facility agreement's own definitions, then confirm each figure yourself.

    “Borrower says leverage is 2.9x. Recompute it with the agreement's own EBITDA definition.”

Advanced 4 lessons

  1. Build comps and precedents for a pitch book

    Produce a sourced first pass of trading comps and precedent transactions, then check each multiple against filings.

    “The all-nighter comps table, drafted with sources, then checked line by line.”

  2. Speed up KYC and AML case reviews

    Use AI to triage adverse-media hits and draft a case narrative while the disposition decision stays with the analyst.

    “200 adverse-media hits, 3 that matter. AI sorts; the analyst decides.”

  3. Apply model risk rules to AI at your bank

    Explain how model risk guidance scopes generative and agentic AI, using the US Federal Reserve's SR 26-2 (which replaced SR 11-7) and the UK PRA's SS1/23 as examples, and check which framework your bank follows.

    “In April 2026 the US Fed's SR 26-2 replaced SR 11-7 and scoped generative AI out.”

  4. Keep AI in credit decisions fair and explainable

    Apply adverse-action, anti-discrimination and high-risk AI duties to AI used in lending decisions, using examples such as the US Equal Credit Opportunity Act and the EU AI Act's credit-scoring rules, and check what applies in your market.

    “Under US rules, if AI helps decline a loan, the applicant is owed the specific reasons.”

Expert 1 lessons

  1. Set maker-checker controls for AI-drafted credit work

    Specify, for one AI-assisted credit or deal task, the checker step, the evidence logged and the error that triggers escalation, so it passes risk and audit review.

    “Audit asks who checked the AI-drafted memo, and how. Have the answer in the file.”

Apps we recommend

These are the tools this course uses, and what each one is best for. We have no partnership with any of them.

Rogo

Best for: Investment banking research, comps and deck prep

Built by ex-bankers for sell-side work; runs in Excel, PowerPoint and Word with licensed financial data. Clients include Lazard and Moelis.

Paid

AlphaSense

Best for: Searching filings, broker research and expert calls

Market-intelligence search with generative summaries across filings, transcripts and expert calls, citing the source passage.

Free trial

nCino

Best for: Commercial lending workflows and credit memos

Loan origination platform used by many banks; its Banking Advisor AI drafts credit memo narratives and answers credit-policy questions.

Paid
Hands-on project

Produce one AI-assisted credit or deal deliverable

Take a real task from your desk: a credit memo section, a covenant test, an earnings digest or a comps table. Do it in your bank's approved tool with public or anonymised data only, reconcile every figure to source and put it through your normal maker-checker review.

What you’ll make: The finished section or table plus a control note: sources, figures reconciled, AI steps used, checker and approval date.

Check your own work

  • Every figure ties to the spreads, filing or transcript, with its location noted
  • No customer data or MNPI entered a tool outside the bank's approved list
  • Ratios were recomputed independently using the governing definitions
  • A checker other than the maker reviewed it and signed off
  • AI use is recorded as your bank's model risk or AI policy requires

Try a ready-made prompt

Copy a better prompt for a real task, and see why it works better than the version most people type.

Further reading and viewing

We’ve chosen these articles, research papers, videos and guides from other trusted sources.

Guide · Federal Reserve Board · 30 min

SR 26-2: Revised Guidance on Model Risk Management

The April 2026 guidance that replaced SR 11-7, including how it treats generative and agentic AI.

Research · Financial Stability Board · 40 min

The Financial Stability Implications of Artificial Intelligence

How global regulators see AI risks in finance: third-party concentration, cyber, model risk.

Research · Bank of England and FCA · 25 min

Artificial intelligence in UK financial services - 2024

Survey data on how UK firms use AI, which use cases are growing and what risks they report.

Guide · Future of Life Institute (AI Act Explorer) · 10 min

EU AI Act Annex III: High-risk AI systems

Shows why creditworthiness and credit scoring systems count as high-risk under EU law.

Questions about this course

How is AI used in banking?

Banks use AI for fraud detection, customer service, credit analysis, research and compliance checks. Generative AI now helps bankers draft memos, client briefs and pitch materials, and summarise filings and earnings calls. Some uses touch customers or credit decisions directly, which brings stricter rules. This course starts by mapping these uses and where the risks sit.

Can bankers use ChatGPT at work?

Only in the way your bank allows. Many banks restrict public chatbots and provide approved internal tools instead. Customer data, KYC files and material non-public information should never go into an unapproved tool. Public information, such as published annual reports, is usually lower risk. Check your bank's AI policy, and when unsure, ask your risk or compliance team.

What rules apply to AI in banking?

It depends on your market. Model risk guidance, such as the UK PRA's SS1/23, can cover AI models. Fair-lending and anti-discrimination laws apply to AI in credit decisions, such as the Equal Credit Opportunity Act in the US. The EU AI Act treats credit scoring of individuals as high risk. Check which frameworks your bank follows.

Can AI write a credit memo?

It can draft sections, such as the business description and cash-flow analysis, from financials you provide. It can also misread a figure or calculate a ratio wrongly. Reconcile every number to the spreads, check its reasoning, and put the draft through your normal maker-checker review, where a second person checks the work. The credit judgement stays yours.

How long are Tau lessons?

About a minute: a short video, one question, then three takeaways and the sources. Many lessons include a realistic task to try with feedback from AI. A three-question skills check sets your starting level. The first 5 lessons are free, and Tau Pro gives you every course and lesson.

Get early access to this course

Tau opens soon. Join the waitlist and we’ll email you as soon as you can start this course.