AI for Domain Experts

Learning objectives

• Separate the myths of AI from its realistic promises, and say what AI is and is not.
• Name the spectrum of what AI can do and the right tool for each job.
• Audit your own working day for where AI genuinely helps.

Key focus

• What AI is and is not: myths versus promises
• The spectrum of possibilities, each grounded in a specific tool
• Chat vs assistant vs agent vs workspace
• Free, paid, and team plans, and the confidentiality line
• Tokens and the token economy

You produce

A working-day opportunity list and a personal tool map.

Target win

Identify 5 or more AI-suitable tasks and 3 or more hours a week of addressable grind.

Learning objectives

• Diagnose why a prompt underperforms and fix it.
• Use the role, context, task, constraints, format pattern reliably.
• Steer output, and improve a draft with self-critique.

Key focus

• The five-part prompt pattern
• Steering with examples; controlling tone, length, and format
• Making the AI interview you before it answers
• Reason-then-critique for a stronger draft

You produce

Three rebuilt, reusable prompts for your real tasks.

Target win

First drafts you edit, not rewrite: about half the correction rounds.

Learning objectives

• Turn a recurring task into a reusable, parameterised master prompt.
• Organise prompts so they are one click away.
• Start a shared team prompt library.

Key focus

• What a master prompt is and why it beats one-off asks
• Variables and fill-in-the-blank templates
• The four homes: custom instructions, saved prompts, workspaces, team library
• Naming, versioning, and pruning a library

You produce

A personal prompt library of 3 to 5 master prompts.

Target win

Recurring tasks start from a saved prompt in seconds, not from scratch.

Learning objectives

• Configure an assistant with custom instructions and memory.
• Build a reusable workspace over your own documents.
• Get grounded answers with citations from your own files.

Key focus

• Custom instructions and memory
• Projects and Custom GPTs as reusable mini-assistants
• A knowledge base the assistant always references
• Connectors to email, calendar, and files (awareness)

You produce

A configured assistant and one working workspace over your own files.

Target win

Stop re-uploading and re-explaining: a grounded, cited answer in one step.

Learning objectives

• Explain, as a user, what an agent, a skill, and MCP are.
• Package a repeatable workflow, and set an act-versus-approve rule.
• Describe no-code automation and where it fits, with no programming.

Key focus

• From answering to acting: agents
• Skills: packaging a repeatable workflow
• MCP explained for an AI user (the USB port for AI)
• No-code automation: trigger, AI step, action
• When to let AI act versus approve first

You produce

A packaged skill or automation and a written autonomy rule.

Target win

A multi-step routine runs from one prompt, or on a trigger, instead of by hand.

Learning objectives

• Catch hallucinations and verify output before relying on it.
• Apply clear rules on confidentiality and safe tool choice.
• Assemble a personal AI playbook that ties the week together.

Key focus

• Verification: AI drafts, you are accountable
• Confidentiality and the plan that makes a tool safe
• Data controls, memory, and privacy settings
• Your profession's guardrails, in brief

You produce

Your completed personal AI playbook.

Target win

A Monday-ready system, with safe and verified use as the default.

Learning objectives

• Generate a first draft of a document from a brief, grounded in your own templates.
• Steer drafting with a clause library to match house style.
• Iterate to filing quality while keeping control of the substance.

Key focus

• A brief to a grounded first draft
• Grounding in your precedents and clause library
• The drafting master prompt, marking any invented content
• Iterate: self-critique, verify defined terms and any cited authority

You produce

A Legal Drafter workspace and a drafting master prompt.

Target win

A house-style first draft from a brief in under 30 minutes.

Learning objectives

• Run a structured review that surfaces risks and missing protections.
• Compare an incoming draft against your standard position.
• Produce a redline-ready list of changes, each with a reason.

Key focus

• The review master prompt: summary, obligations, risks, missing items
• Deviation analysis against your standard
• From review to a redline-ready list
• Flags are a first pass to verify, not a verdict

You produce

A Contract Reviewer workspace and a review master prompt.

Target win

A structured review plus a redline-ready list in under 30 minutes.

Learning objectives

• Use grounded tools for current, source-cited answers.
• Summarise long judgments and statutes accurately, with pin-cites.
• Verify research and avoid the fabricated-citation trap.

Key focus

• Two jobs, two tools: summarise what you have versus research the law
• Pin-cited summaries of long documents
• Grounded, citation-first research
• The fabricated-citation trap and verification

You produce

A research master prompt and a personal verification checklist.

Target win

A pin-cited summary of a long judgment in minutes, with every authority verified.

Learning objectives

• Query a large document set in natural language, grounded.
• Extract obligations, dates, parties, and red flags across many files.
• Build a repeatable diligence workflow that scales beyond manual review.

Key focus

• Querying a whole corpus with file and page citations
• Extracting a structured findings table
• A repeatable standard question set
• Respecting coverage limits: AI narrows, you review what matters

You produce

A diligence question set and a findings-table format.

Target win

A data room narrowed to the pages that matter, in a fraction of manual time.

Learning objectives

• Turn dense legal material into clear client-facing explanations.
• Draft client updates and advice notes quickly and consistently.
• Adapt tone and complexity to the audience.

Key focus

• The plain-language explanation, accurate and audience-tuned
• Routine correspondence in house voice
• Plain does not mean inaccurate: preserving the material nuance

You produce

A client-explanation master prompt and a client-email template.

Target win

An accurate plain-language explanation or client letter in minutes.

Learning objectives

• Assemble a complete, grounded Legal AI workspace over your own materials.
• Apply privilege, confidentiality, jurisdiction, and the duty to verify.
• Leave with a Monday-ready legal AI setup.

Key focus

• Assembling the workspace from the track's components
• Guardrails: privilege and confidentiality
• Guardrails: jurisdiction and fabricated authority
• A personal ethics-and-verification checklist

You produce

A Legal AI workspace and a personal ethics-and-verification checklist.

Target win

One matter run end to end, and a reusable, Monday-ready legal setup.

Learning objectives

• Summarise long financial documents into decision-ready briefs.
• Extract and sanity-check key figures, trends, and ratios.
• Frame the right questions to ask of a set of accounts.

Key focus

• A long report into a decision brief with page cites
• The financial-summary master prompt
• Extracting a key-figures table
• Every number is verified: AI never certifies a figure

You produce

A financial-summary master prompt and a key-figures extraction format.

Target win

A 100-page report into a decision brief in under 30 minutes.

Learning objectives

• Draft MIS narratives, board notes, and commentary from figures.
• Turn a variance table into a clear, structured explanation.
• Keep recurring reports consistent period to period.

Key focus

• Turning numbers into narrative
• A workspace with your report templates and prior notes
• House-structure commentary and board notes
• Reconcile every cited figure to source

You produce

An MIS and board-note workspace and a commentary master prompt.

Target win

A first-draft management commentary from a variance table in under 30 minutes.

Learning objectives

• Compare datasets and surface mismatches and outliers.
• Get likely causes for anomalies to investigate.
• Understand the limits: AI assists the eye, it does not replace the tie-out.

Key focus

• Comparing two sets into a categorised difference table
• Flagging anomalies and testing the reasons
• AI as anomaly-spotter, not the reconciliation of record
• Verifying every computed number

You produce

A reconciliation and anomaly master prompt and a verification note.

Target win

Anomalies across thousands of rows surfaced in minutes, with each material item verified.

Learning objectives

• Summarise and interpret provisions, circulars, and standards.
• Track and explain what changed in an updated regulation.
• Verify regulatory answers against primary sources.

Key focus

• A circular into a practitioner brief
• Old-versus-new change summaries
• Current-rule research with every source opened
• Judgment and the final position stay yours

You produce

A regulation-summary master prompt and a change-tracking format.

Target win

A practitioner brief plus an old-versus-new change summary in minutes.

Learning objectives

• Draft and structure working-paper documentation faster.
• Build test checklists and document procedures and findings.
• Keep audit documentation consistent and review-ready.

Key focus

• Drafting procedure narratives and findings
• First-cut test checklists from the applicable standard
• Consistency and review-readiness
• Conclusions stay the auditor's; verify against the work performed

You produce

A working-paper drafting workspace and a checklist master prompt.

Target win

Review-ready working-paper narratives in a fraction of the time.

Learning objectives

• Assemble a complete Finance AI workspace over your templates and data.
• Apply figure verification, confidentiality, and control over material information.
• Leave with a Monday-ready finance AI setup.

Key focus

• Assembling the workspace from the track's components
• Guardrail: every figure verified
• Guardrails: price-sensitive data, audit trail, and retention
• A controls-and-verification checklist

You produce

A Finance AI workspace and a personal controls-and-verification checklist.

Target win

A Monday-ready finance setup, with verified figures as the default.

Learning objectives

• Summarise and interpret policy wordings quickly and accurately.
• Compare coverage across policies and against a standard wording.
• Extract exclusions, conditions, and limits into a usable structure.

Key focus

• A structured cover breakdown with clause references
• Coverage questions answered from the wording
• Comparing wordings and deviation from your standard
• The wording, not the AI paraphrase, governs

You produce

A policy-analysis master prompt and a cover-breakdown format.

Target win

A structured cover breakdown of a wording in minutes.

Learning objectives

• Extract and structure the key facts from an underwriting submission.
• Summarise risk and surface gaps and red flags for a referral.
• Speed the read of a submission without ceding the underwriting call.

Key focus

• A submission pile into an underwriting summary
• Key exposures, loss history, and missing information
• Referral flags
• Judgment, pricing, and terms stay the underwriter's

You produce

An underwriting-summary master prompt and a submission-intake format.

Target win

An underwriting summary and missing-information list in under 30 minutes.

Learning objectives

• Summarise a claim file and its correspondence into a clear status.
• Triage and categorise incoming claims for routing.
• Draft claims documentation and updates consistently.

Key focus

• A claim file into a status note
• Triage by type and severity
• Consistent status notes and update letters
• The coverage decision stays the handler's

You produce

A claims-summary and triage master prompt and a status-note format.

Target win

A claim file into a status note, and a batch triaged, in minutes.

Learning objectives

• Surface inconsistencies and possible fraud indicators in a file.
• Produce structured, defensible case notes for investigation.
• Understand the limits: signals to investigate, never a determination.

Key focus

• Reviewing a file for internal inconsistencies
• Structured case notes: observation, source, why it is unusual
• Signal, not finding: a human decides on the evidence
• Fairness and data protection govern

You produce

A case-review master prompt and a case-note format.

Target win

A whole claim file reviewed for inconsistencies in minutes.

Learning objectives

• Draft customer letters, proposals, and renewal communications quickly.
• Explain cover and decisions to customers in plain language.
• Keep customer communication consistent and on-brand.

Key focus

• Renewal letters that explain a premium change
• Plain-language decision explanations
• Proposals from the risk and terms
• Accuracy, required disclosures, and fairness before it goes out

You produce

A customer-correspondence master prompt and letter templates.

Target win

A renewal letter or decision explanation in minutes, checked for fairness.

Learning objectives

• Assemble a complete Insurance AI workspace over your wordings and templates.
• Apply fairness, disclosure, personal-data protection, and the human decision.
• Leave with a Monday-ready insurance AI setup.

Key focus

• Assembling the workspace from the track's components
• Guardrail: fairness, never the sole basis for an adverse decision
• Guardrails: disclosure duties and personal-data protection
• A fairness-and-verification checklist

You produce

An Insurance AI workspace and a personal fairness-and-verification checklist.

Target win

A Monday-ready insurance setup, with fairness and verification as the default.