30 AI Prompts for Legal & Compliance Teams
Copy-ready prompts that assume you will attach the contract, clause or policy — built to accelerate the first pass so your judgement goes where it counts. Always have a qualified lawyer review anything consequential.
Each prompt gives the AI a legal-analyst persona, references your attached document, applies a structured review method, and defines the output — with a clear instruction to escalate to counsel. These prompts are tool-agnostic and optimised to work across all major AI assistants — Claude, Microsoft Copilot and ChatGPT (and Google Gemini). Replace the [bracketed placeholders] with your own details, and always review AI output before you act on it.
Contract review
First-Pass Contract Risk Review
Doing an initial risk read of a contract before deciding whether to escalate to counsel
full signed/draft contract (PDF/Word)
- Summarize parties, purpose, effective date, and key obligations side by side
- Note term, renewal, and termination triggers plus the liability/indemnity position
- Flag the five clauses most favoring the counterparty, citing clause numbers for every point
- List anything material that must escalate to counsel
Structured review (obligations table, term/termination, liability, top-5 unfavorable clauses) plus a must-escalate list.
See example AI output
Parties: Serunai Ventures (Client) and Kelana Logistics (Vendor), 3-year logistics agreement from 1 Jan 2027. Client pays RM45,000/month (cl.5.1); Vendor guarantees 98% on-time delivery (cl.4.3). Either party may terminate on 90 days' notice (cl.12.1); auto-renews unless 60 days' notice given (cl.12.4). Vendor's liability capped at just 1 month's fees (cl.9.2). Top concerns: uncapped Client indemnity (cl.10.1), broad Vendor force majeure (cl.14), unilateral price review (cl.6.3), exclusivity lock-in (cl.3.2), governing law set to Singapore (cl.18). Escalate cl.9.2, 10.1, 18 to counsel.
Plain-English Clause Explainer
Explaining a confusing clause to a non-lawyer colleague before signing
no file - paste the clause text
- Explain what the clause means in plain language
- State who it favors and why
- Describe the realistic worst-case outcome if triggered
- Suggest one question to raise with counsel, and note this isn't legal advice
Short plain-English explanation, favors verdict, worst-case scenario, one question for counsel.
See example AI output
Clause 10.1 means Kedai Segar must cover the Vendor's losses, fines or legal costs from any claim linked to the products supplied - even if the Vendor was partly at fault. It favours the Vendor: no cap, no carve-out for the Vendor's own negligence. Worst case: a product-liability claim triggers a six-figure indemnity payout our insurance doesn't cover. Question for counsel: can we negotiate a cap and a carve-out for gross negligence or wilful misconduct? Not legal advice - confirm before signing.
Redline Change Impact Analysis
Assessing whether a counterparty's redlines help or hurt us before responding
no file - paste both clause versions
- Compare original and revised clause text word for word
- Identify exactly what changed, including deletions not just additions
- Assess whether each change helps or hurts our position and note the negotiation implication
Change table: Change, Effect (helps/hurts), Our position.
See example AI output
Payment clause (cl.5) changes: (1) '30 days' to '45 days' - hurts us; push back to 30 days or add 1.5%/month late interest. (2) Added 'subject to Client's satisfactory acceptance' - hurts us, creates a subjective payment trigger; require objective, time-bound acceptance (e.g. 5 business days' silent acceptance). (3) Deleted 'time is of the essence' - hurts us, weakens our right to terminate for late payment; reinstate it or add an explicit cure period with termination rights.
One-Sided Terms Risk Scan
Scanning a full draft contract for unusual or one-sided terms before counsel review
full draft contract
- Scan for unusual, one-sided, or missing terms across liability caps, auto-renewal, exclusivity, IP, termination, and governing law
- Rank each finding by risk level (high/medium/low) with a one-line rationale
- Compile a prioritized issues list for counsel - do not give a final verdict
Prioritized issues list (High/Medium/Low) each with a one-line rationale, for counsel review.
See example AI output
Amanah Teknologi SaaS agreement issues: HIGH - no liability cap for either party (cl.9), uncapped exposure on a RM2M contract. HIGH - auto-renews for 24-month terms unless 120 days' notice given (cl.12.2), easy to miss. MEDIUM - IP in customisations assigned to Vendor by default (cl.7.1). MEDIUM - no governing law clause, should specify Malaysia. LOW - exclusivity restricts competing vendors in Sabah/Sarawak only (cl.3.4). Recommend counsel review the HIGH items before signature.
Pre-Signing Due-Diligence Questions
Generating gap-closing questions for the counterparty before signing a near-final agreement
near-final agreement
- Read the agreement to identify gaps and ambiguities
- Generate sharp, deal-specific questions grouped under Commercial, Legal, and Operational headings
- Prioritize questions that close real gaps over generic boilerplate
Question list grouped under Commercial, Legal, and Operational headings.
See example AI output
Questions for Warisan Agro before signing with Ladang Emas: COMMERCIAL - Is the RM18/kg price fixed for 12 months or subject to review? What happens if volume falls below the 500-tonne minimum? LEGAL - Who owns IP in the customised packaging design (cl.8)? Is there a change-of-control clause if Ladang Emas is acquired? OPERATIONAL - What's the lead time for rejected-batch replacement? Who bears recall costs from a food-safety issue? Who is the escalation contact during Ramadan peak demand?
Drafting clauses
Draft Plain-Language Contract Clause
Producing a first-draft clause for counsel to finalize, e.g. a confidentiality clause
no file - specify clause type & purpose
- Identify the clause type and purpose
- Draft a balanced clause for a Malaysian commercial contract, not one-sided by default
- Mark every point needing counsel confirmation and explain the drafting choices made
Draft clause text plus a short note on the choices made.
See example AI output
Draft confidentiality clause for Sinar Kreatif's vendor agreement: 'Each party shall keep confidential all non-public information disclosed by the other party... except to employees/advisers on a need-to-know basis, or as required by law. Survives termination for 3 years. [COUNSEL: confirm carve-outs for independently developed information.]' Note: kept mutual, not one-sided, since both parties exchange pricing and product data; 3-year survival is a common Malaysian SME default, flagged for confirmation.
Rewrite Dense Clause for Clarity
Turning a dense legalese clause into language a business reader understands
no file - paste the dense clause
- Read the pasted clause carefully
- Rewrite it in plain business language while preserving the legal effect exactly
- Keep defined terms consistent and flag if the rewrite risks any meaning drift
Rewritten clause plus a one-line caution note on residual meaning-drift risk.
See example AI output
Original (cl.6.4): the Purchaser's remedies for non-conformance are limited to repair or replacement at Vendor's sole discretion, with no liability for consequential, indirect or special damages. Rewritten: 'If the goods don't meet spec, Timur Jaya (the Vendor) will either repair or replace them - Vendor chooses which. Vendor isn't responsible for any knock-on losses this causes us, like lost sales.' Caution: confirm with counsel whether 'consequential' also excludes loss of goodwill in our jurisdiction.
Draft First-Version NDA
Producing a first-draft NDA structure for counsel to finalize before a new deal
no file - specify parties & purpose
- Set out recitals, definitions, and the operative confidentiality clauses
- Add term, termination, and standard boilerplate provisions
- Mark every point requiring a counsel decision and list assumptions made
Full draft NDA structure with an assumptions list at the end.
See example AI output
Draft mutual NDA between Cahaya Digital ('Party A') and Beta Robotics ('Party B') for evaluating a joint robotics project. Structure: recitals; definitions ('Confidential Information', 'Representatives'); obligations (non-disclosure, permitted use, need-to-know access); exclusions (public domain, independently developed, legally compelled) [COUNSEL: confirm exclusions]; 2-year term surviving termination; governing law Malaysia [PLACEHOLDER]; boilerplate (notices, assignment, entire agreement). Assumptions: mutual disclosure; no IP transfer intended; 2-year term pending business confirmation.
Clause Negotiation Fallback Ladder
Preparing negotiation fallback positions before a clause discussion with a counterparty
no file - paste the clause under negotiation
- Identify our current ask on the clause
- List fallback positions from most to least favorable, with rationale for each
- Identify the realistic, market-consistent walk-away trigger
Tiered fallback table: Position, Rationale, Walk-away marker.
See example AI output
Fallback ladder for the liability cap clause with Nadi Fintech: TIER 1 (ideal) - cap at 12 months' fees, carve-out for gross negligence. TIER 2 - cap at 12 months' fees, no carve-out. TIER 3 - cap at 2x annual fees. TIER 4 (minimum) - cap at 2x annual fees, ceiling RM500,000. WALK AWAY if uncapped liability or a cap below 6 months' fees, given our RM1.2M annual exposure. Tiers 3-4 match typical Malaysian SaaS market terms.
Limitation-of-Liability Clause Options
Presenting counsel with multiple liability-cap options at different risk levels for a deal
no file - specify contract type & deal value
- Draft two or three limitation-of-liability options spanning different risk levels
- Explain the trade-off of each option and note the typical market position for this contract type
- Flag all options for counsel before use - none are legal advice
Options listed side by side with the trade-off explained for each.
See example AI output
Limitation-of-liability options for Zenith Manufacturing (as supplier) on a RM3M equipment contract: OPTION A (lowest risk to us) - cap at 12 months' fees, exclude consequential loss; buyer may resist. OPTION B (balanced) - cap at 1.5x annual value, carve-out for gross negligence and IP infringement; more likely accepted. OPTION C (buyer-friendly) - cap at 3x annual value, includes limited consequential loss for downtime; stronger negotiating position on price. Market norm sits near Option B. Flag all three for counsel sign-off.
Compliance and policy
Regulatory Compliance Checklist Builder
Building an actionable compliance checklist for a named regulation and industry
no file - specify regulation & industry
- Identify core obligations under the named regulation for the industry
- List the evidence needed to demonstrate compliance for each obligation
- Assign an owner and review frequency, tailored to a Malaysian SME context
Checklist table: Obligation, Evidence, Owner, Review frequency.
See example AI output
Compliance checklist for Rimba Fresh (F&B manufacturer): Obligation 1 - maintain valid MeSTI certification - Evidence: certificate + annual audit - Owner: QA Manager - Review: annually. Obligation 2 - label allergens correctly - Evidence: approved label proofs on file - Owner: Packaging Lead - Review: per SKU launch. Obligation 3 - temperature-log the cold chain - Evidence: daily logs, 12-month retention - Owner: Warehouse Supervisor - Review: monthly. [ADVISER: confirm current MeSTI renewal cycle and recent labelling amendments.]
PDPA Obligations Staff Summary
Turning a regulation's requirements into a one-page do/don't guide for non-legal staff
no file - specify the regulation, e.g. PDPA
- Identify key obligations under the named regulation for the business
- Translate them into concrete musts and must-nots for staff
- State the penalties for non-compliance and flag anything needing legal confirmation before publishing
One-page plain-language summary staff can act on.
See example AI output
PDPA one-pager for Meridian Retail staff: MUST - get clear consent before collecting NRIC or phone numbers; use customer data only for its stated purpose; respond to access requests within 21 days. MUST NOT - share customer databases with vendors without a data-processing agreement; keep data longer than needed; email customer lists over personal Gmail. Penalties: breaches can attract fines up to RM1 million and/or imprisonment for the company and responsible officers. [LEGAL: confirm against the latest PDPA amendment before distributing.]
PDPA-Aligned Privacy Notice Draft
Drafting a customer-facing privacy notice for a new data collection use case
no file - specify use case, e.g. checkout page
- Identify what personal data is collected and why
- State the legal basis, retention period, and any third-party sharing
- Draft the notice covering the individual's PDPA rights, marked for counsel review
Draft privacy/data-protection notice text.
See example AI output
Draft PDPA notice for Kedai Manis Online (checkout): 'We collect your name, delivery address, phone number and payment reference to process and deliver your order. Retained 3 years for warranty and tax purposes, then securely deleted. Shared only with our courier partner, Pos Ekspres, solely to fulfil delivery. Under the PDPA 2010, you may request access to or correction of your data at privacy@kedaimanis.my.' [COUNSEL: confirm retention period against tax requirements and courier data-sharing agreement.]
Policy-to-Staff Plain Summary
Translating a dense internal policy into a short employee-facing guide
no file - paste the internal policy text
- Read the pasted policy without adding rules not in the source
- Summarize what staff must and must not do
- Illustrate with three concrete, realistic examples
Short employee-facing summary with three worked examples.
See example AI output
Summary of Gemilang Holdings' Gift & Entertainment Policy: MUST - decline gifts worth over RM200; log any accepted gift with your manager within 5 days. MUST NOT - accept cash or cash-equivalents (vouchers, gift cards) from vendors. Examples: (1) A vendor sends a RM150 CNY hamper - acceptable, but log it. (2) A supplier offers a RM500 shopping voucher - decline and report. (3) A client invites you to an RM80 lunch during a project review - acceptable, no logging needed under the RM200 threshold.
Compliance Training Scenario Quiz
Building a scenario-based quiz to train staff on a compliance topic
no file - specify the compliance topic
- Identify realistic Malaysian workplace scenarios for the named topic
- Write ten scenario-based questions with one unambiguous correct answer each
- Provide a one-line explanation per answer, avoiding trick questions
Ten-question scenario quiz with answers and explanations.
See example AI output
Compliance quiz on anti-bribery for Kencana Engineering (sample of 3): Q1 - A supplier offers to 'fast-track' your invoice approval for recommending them for a tender. Accept? Answer: No - this is a bribe linked to a business decision; report to Compliance. Q2 - An officer hints a 'facilitation payment' would speed up your licence. Answer: Refuse and escalate; prohibited under the MACC Act 2009. Q3 - You receive an RM3,000 'thank you' after closing a deal. Answer: Decline and report - this is a bribe, not a gift.
Legal research and summaries
Plain-Language Legal Concept Explainer
Explaining an unfamiliar legal concept to a business team, e.g. force majeure
no file - specify the topic
- Explain what the concept is in plain language and why it matters in business
- List the typical considerations that apply
- End with points to confirm with counsel - keep general, no case-specific claims
Plain-language explainer plus a list of points to confirm with counsel.
See example AI output
Force majeure, explained for Pelangi Events: a clause that excuses a party from performing when an extraordinary event outside their control makes it impossible - e.g. natural disaster, war, government shutdown. Without it, you could be liable for damages even if the cause was outside your control. Considerations: whether the clause lists specific triggers or is broadly worded; whether notice must be given within a set timeframe; whether it excuses or only delays performance. Confirm with counsel: does our events contract cover pandemic-related closures specifically?
Long Legal Document One-Page Brief
Condensing a lengthy legal document into a one-page brief for decision-makers
full lengthy legal document
- Read the full document and identify purpose, key terms, and obligations
- Identify the risks and the decisions required, citing section numbers for every point
- Keep language plain for a business audience
One-page brief: Purpose, Key terms, Obligations, Risks, Decisions required.
See example AI output
Brief on the 40-page facility agreement between Harapan Capital and Bank Utara: Purpose - RM5M term loan for warehouse expansion (cl.2). Key terms - 5-year tenor, 4.8% p.a., quarterly repayment (cl.3-4). Obligations - maintain debt-service coverage above 1.25x, submit quarterly accounts within 30 days of quarter-end (cl.8.2-8.3). Risks - personal guarantee required from two directors (cl.11); cross-default clause ties this loan to any default under other banking facilities (cl.14.1). Decision required: confirm directors' willingness to guarantee, check cl.14.1 exposure.
Legal Process Step Outline
Mapping the typical steps of a legal process, e.g. terminating a commercial lease
no file - specify the process
- List the typical steps in the named process, in sequence
- Note where local law governs versus where the contract governs
- Flag each point needing counsel confirmation - present as general information, not advice
Numbered step outline with clearly marked verify-with-counsel points.
See example AI output
Typical steps to terminate a commercial lease early (general info for Bumi Cafe): 1. Review the early-termination clause for notice period and penalty [VERIFY: exact wording]. 2. Serve written notice per the lease's method. 3. Settle outstanding rent, utilities and reinstatement costs. 4. Return premises in the agreed condition [VERIFY: contractual condition]. 5. Recover the security deposit, subject to deductions. Malaysian contract law generally lets lease terms govern early termination unless silent, then common law may apply [VERIFY with counsel].
Legal Memo to Talking Points
Converting a lawyer's memo into bullet points a business stakeholder can act on
no file - paste the legal memo
- Extract the points a business stakeholder needs to act on
- Convert them into short, actionable bullet points
- Preserve every caveat the memo states - do not overstate certainty where it hedges
Bullet-point talking points list.
See example AI output
Talking points from counsel's memo on the Sabah distribution deal for Cendana Beverages: We can proceed, but exclusivity should be capped at 18 months, not the requested 3 years. Memo flags uncertainty on whether Sabah licensing rules require separate registration - confirm with a local agent before signing. Termination for convenience should include a 90-day notice period. 60-day payment terms are acceptable but longer than our standard 30-day policy - flag to Finance. Caveat: analysis assumes the distributor isn't government-linked.
External Counsel Instruction Brief
Briefing external counsel tightly before instructing them on a matter, to control cost
no file - specify matter, attach documents
- Summarize the background of the matter concisely
- List specific questions for counsel to answer, not open-ended asks
- State the desired outcome and list documents provided - keep scope tight to control cost
Concise instruction brief: Background, Questions, Desired outcome, Documents attached.
See example AI output
Instruction brief for Messrs Tan & Rahman re: Selera Nusantara franchise dispute. Background: our Ipoh franchisee stopped paying royalties since March 2026, alleging we breached territory exclusivity by opening a new outlet 4km away. Questions: (1) Does our exclusivity clause (cl.9) support their claim? (2) What are our options to recover unpaid royalties (approx. RM84,000)? (3) Is termination for non-payment defensible given the counter-claim? Desired outcome: recover royalties and preserve termination rights, ideally without litigation. Documents attached: franchise agreement, royalty statements, demand letter dated 2 June 2026.
Risk and due diligence
Transaction Due-Diligence Checklist
Building a DD checklist at the start of a transaction, e.g. a share acquisition
no file - specify transaction type
- Group the checklist by legal, financial, operational, and compliance workstreams
- List the document to obtain for each item and the risk it addresses
- Prioritize items to the specific transaction type, avoiding generic filler
DD checklist table: Item, Workstream, Document to obtain, Risk addressed.
See example AI output
DD checklist excerpt for acquiring 70% of Delima Foods: LEGAL - verify no pending litigation - Document: litigation search + director confirmation - Risk: undisclosed liabilities. FINANCIAL - confirm revenue recognition policy - Document: 3 years audited accounts + management accounts - Risk: inflated historical revenue. OPERATIONAL - check key supplier contracts for change-of-control clauses - Document: top 10 supplier agreements - Risk: supply disruption. COMPLIANCE - confirm halal certification is current and transferable - Document: JAKIM certificate + renewal correspondence - Risk: loss of certification disrupting retail sales.
Due-Diligence Risk Register Summary
Turning collected DD documents into a ranked risk register for the deal committee
collected due-diligence documents
- Review the attached documents for risks flagged, basing findings only on evidence found
- Rank each risk by severity and likelihood
- Note the mitigation or deal implication for each and escalate material items clearly
Risk register table: Risk, Severity, Likelihood, Mitigation/Deal implication.
See example AI output
Risk register from Delima Foods DD: Risk 1 - RM1.2M unpaid supplier invoice not on the balance sheet - Severity: High - Likelihood: Confirmed - Implication: adjust purchase price or require pre-completion settlement. Risk 2 - Shah Alam factory lease expires in 8 months, no renewal option - Severity: High - Likelihood: Confirmed - Implication: negotiate renewal pre-completion or discount valuation. Risk 3 - two key sales staff have change-of-control exit clauses - Severity: Medium - Likelihood: Likely - Implication: negotiate retention agreements. ESCALATE Risks 1 and 2 to the deal committee.
Single Risk-Register Entry Draft
Formalizing a single identified issue into a proper risk-register entry
no file - describe the specific issue
- Describe the issue and its root cause
- Assess likelihood and impact, and list current controls in place
- Propose a specific, measurable mitigation with a named owner - avoid vague language like 'monitor closely'
Single structured risk-register entry.
See example AI output
Risk-register entry - Vantage Logistics: Issue - over-reliance on a single haulage subcontractor (Trek Cargo) for 65% of delivery volume, since no secondary-vendor arrangement was set up. Likelihood - Medium (Trek Cargo shows cash-flow strain). Impact - High: a sudden failure would disrupt deliveries to 40+ retail clients within days. Current controls - none formal, only ad-hoc spot-hire. Proposed mitigation - qualify and onboard a second haulage vendor covering at least 25% of volume by Q3 2026; owner - Head of Operations, Ravi Kumar, target 30 September 2026.
Document Red-Flag Identifier
Scanning target company financials or documents for anomalies before a deal
document under review, e.g. target financials
- Scan the document for items warranting further investigation
- Explain why each item is a red flag, focusing on genuine anomalies not routine items
- Rank the flags by level of concern - flag for counsel, don't draw a final conclusion
Ranked red-flag list for counsel with a one-line rationale each.
See example AI output
Red flags in Orkid Pharma's target accounts: HIGH - revenue jumped 45% in Q4 2025 with no corresponding rise in headcount or inventory - possible channel-stuffing or premature recognition. HIGH - RM600,000 'consultancy fee' paid to a company sharing a director with the target - undisclosed related-party transaction. MEDIUM - three senior managers resigned within 4 months - possible undisclosed internal issue. LOW - auditor changed twice in 3 years, worth asking why. Recommend counsel and a forensic accountant review the HIGH items before signing.
Management Risk Options Memo
Presenting management with options and a recommendation on an urgent risk decision
no file - describe the issue requiring a decision
- Explain the issue clearly and present the options available to management
- State the pros, cons, and risk level of each option in a balanced way
- Give a recommendation and flag the memo for counsel sign-off before acting
Short decision memo: Issue, Options with pros/cons/risk, Recommendation.
See example AI output
Memo: response to Cempaka Retail's data breach affecting ~3,000 customer records. Options: (1) Notify customers and the PDP Commissioner immediately - limits reputational damage; short-term hit; Risk: Low-Medium. (2) Delay notification pending investigation - more complete facts, but risks a PDPA penalty and a leak before we control the message; Risk: High. (3) Notify only regulators - limits publicity but likely breaches PDPA; Risk: High. Recommendation: Option 1. Flag for counsel sign-off before any notification is sent.
Communication
Plain-Language Query Response Draft
Answering another department's legal question directly and practically
no file - paste the query from the department
- Answer the department's question directly
- State the risk involved
- Give a practical, actionable recommendation without legalese, keeping necessary caveats intact
Short response: Answer, Risk, Recommendation.
See example AI output
Response to Sales: 'Can we offer a customer a 5% discount for paying upfront in cash, off-invoice?' Answer: You can offer a prompt-payment discount, but it must be documented on the invoice or a formal addendum - an off-book cash arrangement isn't advisable. Risk: undocumented cash discounts can create SST reporting discrepancies and look like under-reported revenue under LHDN audit. Recommendation: apply the discount as a stated invoice line item and route it through normal banking channels.
Legal Explanation for Executives
Rewriting a legal explanation into a decision-first note for an executive
no file - paste the legal explanation to rewrite
- Lead with the decision the executive needs to make
- Follow with the reasoning, preserving factual accuracy from the original
- End with the risk, concisely and without jargon
Short rewritten note: Decision, Reasoning, Risk.
See example AI output
Note to CEO, Anggun Beauty: Decision - Don't proceed with the influencer contract as drafted; ask for the exclusivity clause to be narrowed before signing. Reasoning - the clause stops the influencer working with any beauty or personal-care brand for 12 months post-campaign, broader than needed and could make the deal unattractive or unenforceable if challenged. Risk - signing as-is either invites a costly dispute if breached, or gives us an unenforceable clause, weakening our credibility in future influencer deals.
Firm Position Letter Draft
Drafting a formal position letter to a counterparty before it goes to counsel
no file - specify the matter & position
- State our position on the matter clearly, with the basis for it
- State the specific ask and include a clear deadline
- Keep a professional tone and no threats we can't back up - flag for counsel before sending
Draft letter: Position, Basis, Ask, Deadline.
See example AI output
Draft letter to Pertama Builders re: delayed handover of the Cyberjaya office fit-out. Position: the project is 6 weeks behind the 15 June 2026 completion date with no agreed extension. Basis: clause 7.2 entitles us to liquidated damages of RM2,000/day for delay. Ask: written confirmation of a revised completion date within 7 days, or we begin applying liquidated damages from 16 June 2026. Deadline: response required by 30 July 2026. [FLAG: counsel to confirm the calculation and enforceability before sending.]
Legal Advice to Action List
Converting received legal advice into an owner-assigned action list with deadlines
no file - paste the legal advice received
- Extract each decision the advice implies, staying faithful to the original
- Convert it into an action list with a clear owner and due date per item
- Do not add actions the advice didn't recommend
Action list: Decision, Reasoning, Owner, Due date.
See example AI output
Action list from counsel's advice on the Nusa Interiors trademark dispute: 1. File a response to the cease-and-desist within the statutory period - failing to respond risks a default admission. Owner: Legal Manager Aina, due 5 August 2026. 2. Commission a trademark-clearance search for our proposed new logo, to avoid repeating the conflict. Owner: Brand Manager Farid, due 15 August 2026. 3. Pause use of the disputed logo on new marketing material pending resolution. Owner: Marketing Lead Priya, due immediately.
Staff Guidance for PDPA Requests
Giving staff a step-by-step guide for handling a specific compliance situation, e.g. a DSAR
no file - specify the situation
- List the steps staff should follow for the named situation, in order
- State the applicable timeline and clarify what to do and not do
- State clearly when to escalate to legal
Staff guidance: Steps, Timeline, Do/Don't, Escalation point.
See example AI output
Staff guidance: handling a PDPA data-subject access request at Selesa Homestay. Steps: (1) Acknowledge in writing within 2 working days. (2) Verify the requester's identity before releasing data. (3) Locate all personal data across booking, CRM and accounting systems. (4) Compile and send within 21 days of the request. DO: log every request in the DSAR register. DON'T: release data without identity verification, or discuss the request with the individual's employer or family without consent. Escalate to Legal immediately if it involves a dispute or a minor.
More prompt packs by function
View the full Prompt LibraryFrequently Asked Questions
AI is a powerful first-drafter and summariser, but it can be confidently wrong and must never replace a qualified lawyer for anything consequential. Never paste privileged or confidential documents into public consumer tools — use approved enterprise AI, and always have counsel review the output. Every prompt here instructs the AI to flag items for counsel.
No. AI can explain concepts, draft and summarise, but it does not provide legal advice and can misstate the law. Treat everything as a draft for a qualified professional to verify. These prompts are written with that boundary built in.
You can attach documents to review, but only via your organisation's approved, confidential enterprise AI — never a public consumer tool for privileged or sensitive material. Confirm your firm's data policy first.
For confidentiality, enterprise deployments (Copilot within your Microsoft tenant, or an approved Claude/ChatGPT enterprise plan) matter most. These prompts work across the major assistants.
Yes. AITraining2U's courses are HRD Corp SBL-KHAS claimable for eligible Malaysian employers.
If a document contains personal data - names, NRIC numbers, phone numbers, addresses, or customer records - pasting it into a public AI tool can count as a data disclosure under the Personal Data Protection Act 2010, especially if the tool retains or trains on your input. Before using AI on contracts, HR records or customer files, redact identifiable personal data where possible, and confirm your organisation's approved AI tool has appropriate data-processing terms. When in doubt, use an enterprise-tier tool with no-training guarantees rather than a free consumer version.
No. General-purpose AI models can confidently cite the wrong section number, misstate a repealed provision, or blend Malaysian law with rules from another jurisdiction, because they aren't guaranteed to be current or jurisdiction-accurate. Treat every AI-generated legal reference - Act names, section numbers, case citations - as a claim to verify against the actual statute or a legal database, not a fact to repeat as-is. This is exactly why every prompt in this set builds in a 'flag for counsel' step before anything is relied upon.
AI can help draft and structure disclosure documents, but material, non-public information used to prepare an announcement is highly sensitive - if pasted into a public AI tool, it could create an inadvertent disclosure or insider-trading risk before the announcement is made. Listed companies should restrict any AI use involving price-sensitive information to approved, access-controlled enterprise tools, keep a clear audit trail, and have Company Secretary or counsel sign off before anything drafted with AI assistance is filed or released under Bursa Malaysia's Listing Requirements.
It reduces the amount of billable time spent on first drafts and summaries, not the need for a lawyer on anything consequential. AI is well suited to producing a first-pass contract review, a plain-language explainer, or a due-diligence checklist that a lawyer would otherwise draft from scratch - which can lower legal costs. But signing contracts, filing disclosures, or handling disputes based on AI output alone, without a qualified lawyer's review, exposes an SME to real risk that a modest legal fee would have prevented.
Use your organisation's approved enterprise AI plan (an enterprise Claude, Copilot or ChatGPT deployment with a signed data-processing agreement), not a free consumer account, since enterprise tiers typically commit not to train on your input. Where possible, strip out counterparty names, values and identifying details before pasting, and only attach full documents through tools your IT or compliance team has cleared. Check your NDA obligations too - some confidentiality clauses restrict sharing documents with third-party services, which a public AI tool may count as.
Go beyond prompts — train your team
Prompts are the start. AITraining2U runs hands-on, HRD Corp SBL-KHAS claimable AI training for Malaysian teams — from everyday AI productivity to building agents that run legal & compliance workflows end-to-end.