Generative AI in M&A: How LLMs Help Deals

Generative AI Has Arrived in M&A

Generative AI in M&A is no longer a thought experiment. In 2026, large language models (LLMs) — the technology behind ChatGPT, Claude, Gemini, and their successors — are actively reshaping how deals get sourced, evaluated, documented, and executed across Asia Pacific and globally.

The adoption has been faster than most predicted. When ChatGPT launched in late 2022, M&A professionals treated it as a curiosity. According to McKinsey’s 2025 State of AI report, 78% of organisations now use AI in at least one business function — up from 55% just a year earlier — and 65% are regularly using generative AI, double the rate from the year before. By 2026, the landscape has matured: purpose-built generative AI applications designed specifically for M&A workflows are delivering genuine productivity gains and, in some cases, competitive advantages.

But the landscape is also noisy. McKinsey’s research found that only about 6% of organisations qualify as “AI high performers” — those reporting meaningful EBIT impact — while the majority remain in experiment or pilot mode. Vendors slap “GenAI-powered” on everything from basic search filters to template-based document generators. Understanding what generative AI actually does well in M&A — and what it doesn’t — is essential for deal professionals evaluating these tools.

What Generative AI Means for M&A

Generative AI differs from earlier AI applications in M&A in a crucial way: it handles unstructured tasks.

Earlier AI (pre-2023): Structured data analysis — matching buyers to sellers based on financial criteria, screening targets against quantitative filters, tracking pipeline metrics. These are powerful but limited to tasks that can be expressed as data operations.

Generative AI (2023+): Unstructured content creation and analysis — drafting documents, summarising complex texts, extracting insights from natural language, generating personalised communications, and reasoning across multiple data sources.

The combination of both is where the real impact lies. A deal sourcing platform that uses traditional AI for matching and generative AI for outreach personalisation is more powerful than either approach alone.

Generative AI Use Cases by Deal Stage

Pre-Deal: Sourcing and Strategy

Investment thesis development. LLMs synthesise market research, industry reports, and competitive data into investment thesis frameworks. A PE associate can describe a thesis in plain language and receive a structured analysis of the market opportunity, key risks, and comparable precedents in minutes.

Target screening. Generative AI processes unstructured company descriptions, news articles, and management commentary to identify targets that fit qualitative criteria — “companies where the founder is signalling succession readiness” or “businesses benefiting from regulatory tailwinds in ASEAN fintech.” This goes beyond traditional database filtering.

Market intelligence. LLMs continuously process news, regulatory changes, earnings calls, and industry publications, producing daily briefings tailored to specific sectors and markets. For APAC deal teams covering 10+ markets, this is more efficient than manual monitoring.

Deal Marketing: Documents and Outreach

Teaser and CIM drafting. The most widely adopted generative AI use case in M&A. LLMs produce first drafts of investment teasers and CIM sections from structured deal data. The output respects M&A conventions — blind descriptions, financial ranges, thesis-led narrative — when purpose-built for the industry.

For a detailed analysis, see our article on AI teaser generation.

Buyer-specific outreach. Rather than generic email templates, generative AI produces substantively different outreach for each buyer segment — emphasising synergies for strategic buyers, returns for financial buyers, and strategic fit for corporate development teams. The messaging reflects each buyer’s publicly stated strategy and recent transaction activity.

Pitch deck content. LLMs draft slide content for mandate pitches, management presentations, and market update decks. The strategic narrative still requires senior banker input, but the underlying content — market data, competitive positioning, financial summaries — can be drafted by AI.

Due Diligence and Evaluation

Document summarisation. Generative AI excels at summarising large document sets — contracts, regulatory filings, technical documentation — into structured summaries with key terms extracted and risks flagged. A 200-page contract can be summarised into a 2-page brief with key provisions highlighted.

For a comprehensive look at AI in due diligence, see our article on AI due diligence in M&A.

Q&A preparation. Before management meetings, LLMs generate question sets based on the available information — identifying gaps in the data room, inconsistencies in financial statements, and areas where management commentary would add value. This ensures deal teams arrive at meetings well-prepared.

Risk analysis. Generative AI identifies and categorises risks from multiple data sources — contractual risks from legal documents, financial risks from statements, market risks from industry analysis — producing an integrated risk matrix.

Negotiation and Execution

Negotiation preparation. LLMs analyse comparable transaction terms, precedent transactions, and market benchmarks to inform negotiation strategy. They draft position papers, anticipate counterarguments, and structure talking points for deal meetings.

Agreement drafting support. While legal drafting remains the domain of experienced lawyers, LLMs assist with first drafts of ancillary documents, red-lining mark-ups, and clause comparison across versions. This accelerates the documentation phase without replacing legal judgement.

Real-World Examples: Where GenAI Is Working

APAC Cross-Border Process

A mid-market sell-side process marketing a Vietnamese manufacturing company to buyers across Asia Pacific used generative AI for:

• Teaser variants in English, Japanese, and Korean, each emphasising different aspects of the opportunity for different buyer types
• Buyer research briefs for 80+ potential buyers, synthesising each buyer’s public strategy, recent acquisitions, and likely interest level
• Outreach personalisation that referenced each buyer’s specific strategic priorities and ASEAN presence

Result: 3x broader buyer coverage, completed in one-third the time of a traditional process, with response rates 40% higher than the firm’s historical average.

PE Sector Scan

A PE fund evaluating an investment thesis around “healthcare services consolidation in Southeast Asia” used generative AI to:

• Map the entire landscape of healthcare services companies across six ASEAN markets from public data and news sources
• Generate investment memos for the 25 most promising targets, each with financial profile, competitive position, and risk assessment
• Draft thesis validation documents for their investment committee

Result: a process that would have taken two associates six weeks was completed in 10 days with AI assistance, with broader coverage than the manual approach would have achieved.

What’s Overhyped vs. What’s Genuinely Useful

Overhyped

“AI will replace investment bankers.” It won’t. Generative AI handles production tasks — drafting, summarising, analysing. The core of investment banking — relationships, judgement, negotiation, strategic advice — requires human expertise that AI can’t replicate.

“AI-generated documents are ready to send.” They’re not. Every AI-generated document requires human review, editing, and strategic framing. Treating AI output as final product risks factual errors, confidentiality breaches, and strategic missteps.

“ChatGPT is good enough.” For ad hoc tasks, generic LLMs are useful. For professional M&A workflows, purpose-built tools that understand deal conventions, confidentiality requirements, and industry terminology produce significantly better output.

“AI due diligence means you need fewer professionals.” AI accelerates due diligence but doesn’t reduce the need for experienced professionals. It changes what they spend time on — from document review to judgement-intensive analysis. Reducing the DD team because “AI handles it” increases risk.

Genuinely Useful

AI as a first-draft engine. The biggest productivity gain from generative AI is eliminating the blank page. Starting from a 70% complete first draft rather than scratch saves hours on every document, every outreach email, and every analysis.

AI as a coverage multiplier. Generative AI enables deal teams to cover more ground — more buyers contacted, more targets screened, more documents reviewed — without proportionally increasing headcount. McKinsey estimates that generative AI could unlock US$2.6–4.4 trillion in additional value across the enterprise. For M&A teams, the coverage advantage is the sustainable competitive edge.

AI for cross-border complexity. In APAC, where deals span multiple languages, regulatory frameworks, and business cultures, generative AI’s ability to process multilingual content and synthesise cross-market intelligence is particularly valuable.

AI for institutional knowledge. LLMs integrated with a firm’s deal history and sector expertise create a form of institutional memory. New team members can access synthesised knowledge that would otherwise take years of deal experience to accumulate.

Risks and Guardrails

Confidentiality Risk

M&A data is highly sensitive. Using public LLMs (like ChatGPT) for deal-related work creates confidentiality risk — deal details may be used in model training or exposed through data breaches. Deal teams should use enterprise-grade or self-hosted AI tools with appropriate data handling guarantees.

Hallucination Risk

LLMs generate plausible-sounding content that may be factually incorrect. In M&A, a hallucinated financial figure, a fabricated regulatory requirement, or an invented precedent transaction can have material consequences. Human verification of all factual claims in AI-generated content is non-negotiable.

Over-Reliance Risk

Teams that become dependent on AI output without maintaining the underlying expertise risk losing the ability to do critical analysis independently. AI should augment expertise, not substitute for it. Junior team members still need to learn the fundamentals of deal analysis, even as AI handles routine production.

Regulatory Risk

Financial regulators across APAC are developing frameworks for AI use in investment services. Compliance requirements will vary by jurisdiction and may affect how AI tools can be used in client-facing activities. Staying ahead of regulatory developments is essential for firms embedding AI in their workflows.

Building a Generative AI Strategy for Your Firm

For M&A teams looking to adopt generative AI effectively:

Start with production workflows. Document generation, outreach personalisation, and research synthesis offer the clearest, lowest-risk starting points. Gains are immediate and measurable.

Choose purpose-built over generic. M&A-specific AI tools outperform generic LLMs for deal work. The difference in output quality — understanding of conventions, confidentiality handling, workflow integration — justifies the investment.

Maintain human oversight. Every AI output should be reviewed by someone with the expertise to identify errors and add strategic value. Build review workflows, not autopilot systems.

Invest in data security. Ensure your AI tools meet the confidentiality standards your deals require. Enterprise-grade deployment, data isolation, and clear data handling policies are minimum requirements.

Measure and iterate. Track the impact of AI adoption on team productivity, deal velocity, and output quality. The data will inform where to invest further and where to pull back.

This is how we work at Amafi — generative AI is built into our core advisory workflow (sourcing, matching, document generation, outreach) with the security and domain expertise that professional dealmaking requires. Our clients get the speed of AI with the judgement of experienced advisors.

---

This is what AI-native M&A advisory looks like. Amafi uses generative AI across every phase of the deal — from buyer identification to document preparation — delivering faster processes and better outcomes for sellers. No retainers, success fee only. Book a valuation meeting to see the difference.