Hong Kong Valuation Guide

Recommended practice for using AI in valuation

AI can improve workflow, but it should assist—not replace—professional judgement, evidence verification, confidentiality controls and accountability for the conclusion.

Useful applications

  • Drafting information-request lists and interview questions
  • Organising industry material and summarising lengthy agreements
  • Checking model formulas, links and internal consistency
  • Generating scenario ideas and sensitivity tables
  • Improving report structure and plain-language explanations
  • Flagging unusual items for further professional review

High-risk or inappropriate applications

  • Accepting unverified comparable companies, transactions or market data
  • Using invented citations, discount-rate inputs or legal/accounting conclusions
  • Uploading identifiable client data to uncontrolled public systems
  • Producing a value or fee range without purpose and case details
  • Allowing AI to select the final method, make key judgements or sign the report
  • Hiding material automation so that the work cannot be traced or reproduced

A practical control framework for Hong Kong valuation work

Define permitted uses and risk levels

Separate administrative assistance, research organisation, calculation support and professional judgement. Higher-risk uses need stronger approval, testing and review.

Protect confidentiality and personal data

Review the provider’s data-processing, retention, training and cross-border arrangements. Do not enter client names, personal data, non-public transaction information or complete data-room contents into unapproved tools.

Verify every material input and source

Trace market data, contractual terms, industry information, accounting requirements and technical claims to reliable primary sources, recording source dates and versions.

Validate models and calculations

Use independent calculations, samples, extreme scenarios and version comparisons to check formulas, code, units, signs, time periods and output logic.

Document human review and challenge

A suitably competent professional should review methods, inputs, assumptions, bias and conclusions. The valuer remains responsible for professional judgement and scepticism.

Retain an audit trail and disclose appropriately

Record tools, versions, significant prompts, input limitations, human amendments and quality controls. Where AI use is material, consider client, employer, auditor, regulator and professional requirements.

Why “AI says the fee should be…” is unreliable

AI may combine different purposes, jurisdictions, deliverables and time periods. Unless it receives the valuation purpose, company scale, industry and business model, data condition, instrument or asset complexity, reporting requirements, deadline and reviewer expectations, the resulting fee range is little more than a guess.

Recommended practice: use AI to help organise the seven-line background brief—not to replace formal scoping and a professional quotation.

Questions a client can ask the valuation firm

  • Which AI or automated tools may be used on my engagement?
  • Will confidential information leave the firm’s controlled environment?
  • How are market data, comparable companies and model outputs verified?
  • Who reviews AI-assisted work, and who remains accountable for the conclusion?
  • Can the report and workpapers trace sources, judgements and changes clearly?

Primary references

Prepare a clearer enquiry

Help the valuer understand your case

Use the checklist to summarise the purpose, target, business model, financial scale, valuation date, deadline and expected reviewers.

Open enquiry checklist