AI in M&A – how technology is changing company sales and advisors’ work

Only a few years ago, running an M&A process required a heavy load of analytical work. Advisory teams spent weeks scanning the market, searching for potential investors, benchmarking deals, preparing materials, reviewing documents and building successive versions of analysis.

Generative AI is rapidly changing that model. According to a 2025 Deloitte survey of 1,000 corporate and private equity representatives in the US, 86% of the organisations surveyed already used generative AI in at least part of their M&A processes. About 40% reported using it in more than half of their deals.

McKinsey also points to the first measurable effects of this shift. Organisations using generative AI in M&A reported an average cost reduction of roughly 20%, and 40% of respondents said the technology had shortened the deal cycle by 30–50%.

This does not mean AI will run the transaction for the owner or replace a good advisor. It changes something more fundamental: where the advisor creates value.

Source: Deloitte 2025 M&A Generative AI Study.

AI in M&A is no longer an experiment

The first wave of generative AI use covered mainly simple applications: summarising documents, research, drafting first versions of materials, and analysing large sets of information. Today, its use already spans practically the whole deal cycle.

In the Deloitte survey, GenAI use was concentrated in, among other areas, M&A strategy and market analysis, screening potential acquisition targets, due diligence, valuation, deal execution and post-deal integration. At the same time, Deloitte notes that companies rolling out AI more widely still place heavy emphasis on data quality, security, compliance and oversight of model outputs.

McKinsey, based on a survey of 200 market participants, in turn finds that the most commonly cited benefits are better information for decision-making, simpler processes, shorter deal timelines, lower costs and improved accuracy in parts of the analysis.

AI is becoming part of the deal infrastructure, much as databases, electronic data rooms and CRM systems did before it.

Which parts of the process are changing fastest?

The work most exposed to change is the kind that has, until now, mainly required time.

Research and market analysis

Until recently, mapping a market meant manually trawling databases, reports, websites and company registers. AI can now, in a short time:

• analyse hundreds of companies,
• group them by defined criteria,
• identify potential investors or acquisition targets,
• compare business models,
• surface sector linkages,
• analyse information found across many sources at once.

This does not automatically produce a good buyer list. But it sharply lowers the cost of putting one together in the first place.

Financial data analysis

AI can help with:

• analysing financial statements,
• flagging unusual line items,
• comparing historical results,
• drafting first versions of EBITDA normalisation,
• working-capital analysis,
• spotting anomalies,
• building scenarios and summaries.

The same applies to analysing large volumes of operational data that previously required manual processing.

Deal materials

A teaser, an information memorandum or an investor presentation still needs a good idea for how to tell the company’s story, but the technical work of preparing them is now much easier. AI can speed up:

• gathering information,
• structuring the document,
• drafting first versions of the narrative,
• analysing source data,
• comparing successive versions of the materials,
• preparing charts and exhibits.

Producing a polished-looking presentation is, on its own, an ever smaller edge.

Due diligence

This is probably one of the areas where AI’s impact will be especially large. Models can support the review of thousands of documents, flag contractual clauses, surface discrepancies, cluster risks and prepare first-pass summaries for deal teams.

In its 2026 M&A outlook, PwC notes that AI is already speeding up due diligence, valuation and the preparation of materials for investment committees, while stressing that human judgement remains critical.

Easier to produce an analysis. Not always easier to produce a good one

This is one of the most important consequences of the current shift. Modern models can produce, in a few minutes, an analysis that:

• is logically structured,
• looks professional,
• uses the right terminology,
• contains numbers and arguments,
• sounds convincing.

That does not mean it is correct. AI can draw on the wrong data, misread context, conflate concepts, adopt unfit assumptions, or state conclusions with more confidence than the underlying information warrants.

PwC points out that building real AI competence in deal processes takes more than buying access to a tool. What is needed includes the ability to trace where results come from, well-defined processes, systematic validation, and drawing on knowledge accumulated from past deals.

Access to AI is not the same as competence. The easier it becomes to produce a professional-looking analysis, the more it matters to know how to judge whether the data, assumptions and conclusions behind it actually hold up.

So what remains the advisor’s role?

If research, data analysis and producing materials are getting faster, that does not mean the need for advisory disappears. Its centre of gravity shifts instead.

Understanding what is actually being sold

Two companies with the same revenue and EBITDA can be worth very different amounts to specific investors. An advisor should understand:

• which parts of the business are genuinely valuable,
• what risks a buyer will see,
• where the synergy potential lies,
• which metrics will matter most,
• what could justify a higher valuation.

AI can help with the analysis, but it will not choose the right positioning strategy for the company on its own.

Judging what matters

A transaction generates a huge amount of data. Not all of it carries the same weight. An experienced advisor should be able to tell apart:

• a material issue from a technical one,
• a genuine investor from a party merely gathering information,
• an attractive offer from one that only looks good on the first page,
• a concession worth accepting from one that materially changes the deal economics.

Genuinely reaching investors

AI can efficiently draft a list of potential investors, but a list of contacts rarely translates into real interest on its own. Reaching the right people and getting them to spend time on the process takes something no model can generate: relationships, an understanding of how specific funds and strategic investors think, and the ability to run a conversation tailored to that particular counterpart. A good advisor knows who to call personally, how to sequence outreach, when an unlikely network connection is worth using, and when it is better to hold the first contact for the right moment in the process. It is this network and this creative approach to outreach – not simply appearing on a list – that determines whether a process attracts real, engaged buyers.

Creating competition among investors

One of the biggest sources of value in a good sale process arises when the seller is not negotiating with a single buyer. The difference between one offer and several credible offers is not only about price. Competition can also affect:

• the size and terms of an earn-out,
• how debt and working capital are settled,
• the scope of the seller’s liability,
• the terms of the owner staying on in the business,
• capital the seller retains in the company,
• non-compete provisions,
• conditions precedent,
• timing and certainty of closing.

No tool creates this competition on its own.

Negotiation and decisions

At some point, every deal stops being an analytical problem. It becomes a decision problem. Accept the offer? Grant exclusivity? Risk losing an investor by pushing for a higher price? Accept part of the price as an earn-out? Sell 100% of the company, or keep part of it?

Decisions like these need not just data, but also experience, risk judgement, and an understanding of both sides’ interests.

AI may reshape the traditional structure of advisory firms

For years, the typical investment-banking and M&A advisory model was built like a pyramid. A partner or senior banker owned the client relationship and the key decisions. Junior analysts did much of the analytical work. Successive layers of the organisation reviewed materials before they reached the most senior people.

That model made economic sense: senior time was scarce, and much of the work was labour-intensive. AI reduces the weight of that constraint. An experienced advisor can now personally carry out a much larger share of the analysis, using technology to process data quickly and generate successive variants, while continuously judging the quality of the output.

In our view, this may favour smaller, more senior-led teams, where less time and cost is lost handing work between layers of the organisation.

This does not mean large organisations will not use the same technologies. They will – and already do. The difference may lie in how quickly the technology translates into team structure, working practices and the economics of client service.

Will AI lower the cost of M&A advisory?

It is still too early to talk about a new pricing standard for the M&A market. Available research does show, though, that technology can meaningfully cut the time and cost of parts of deal work. McKinsey points to an average cost reduction of roughly 20% among organisations using GenAI in M&A, and some respondents report very significant reductions in process time.

In our view, this should over time be reflected in how advisory services are billed. That does not mean a fixed fee for running the process – often called a retainer – stops making sense.

The retainer covers more than the cost of research and preparing materials. It also rewards the team’s genuine commitment, senior availability, running a months-long process, and the advisor’s risk that the deal is interrupted or the owner changes their plans. These elements do not disappear with the arrival of AI.

What does shrink is the cost of the work a traditional retainer used to fund: manual research, processing information, screening investors, drafting first-pass analysis, or producing materials.

How we think about a fair fee model

At PROMETEIA, we believe the benefit of using technology should not sit solely with the advisor. If part of the work can now be done faster and with fewer resources, the client should share in that benefit too.

This is not about competing on the lowest price. It means the fixed part of the fee should reflect the actual work involved, senior involvement and process risk – not the cost of an organisational structure designed for the technology available ten or twenty years ago.

That is why we favour a model in which:

• the retainer covers the actual work and commitment needed to run the process,
• its level reflects the efficiency modern tools provide,
• a meaningful part of the advisor’s economics remains tied to the deal’s success,
• where appropriate, additional pay can be linked to a result that exceeds a previously agreed benchmark.

Technology should not exist solely to widen the advisor’s margin. It should also improve the economics of the process for the client. The point is not for the advisor to be “cheaper thanks to AI” – it is for the client to pay less for producing the process, and more for experience, sound judgement and the outcome.

Can AI replace an advisor?

In our view, that is the wrong question. AI will likely replace, or significantly reduce, a set of tasks people were still doing recently. It already affects research, document review, screening, modelling, preparing materials and other parts of the process.

But one could just as well say that spreadsheets replaced much of the manual work financial analysts used to do, and that electronic document repositories replaced physical rooms full of paper. That did not eliminate the need to make decisions. The tools changed. AI will be similar.

The best advisors will not compete with AI. They will use it to do more of the work faster and more accurately, while focusing their own time where their experience creates the most value.

PwC rightly notes that simply holding a licence for a tool does not build a lasting edge. An edge can only emerge by combining technology with a firm’s own processes, accumulated knowledge, validation of outputs, and experience from past deals.

AI is changing not just how deals are run, but what is being bought

The shift is not confined to advisors’ work. AI is increasingly shaping acquisition strategy itself. According to a PwC analysis, roughly one-third of the 100 largest corporate deals announced in 2025 cited AI as part of the stated strategic rationale for the transaction. This was especially pronounced in technology, but also visible in manufacturing and in energy and utilities.

Source: PwC, Global M&A Industry Trends 2026.

Buyers are therefore analysing not just a company’s current results, market, customers, margins and cash flows, but increasingly also:

• how resilient its business model is to AI,
• whether AI can improve its productivity,
• whether technology can lower barriers to entry in its sector,
• whether part of its products or services risk being displaced,
• whether the company holds data, capabilities or assets that become more valuable in an AI-driven environment.

PwC’s 2026 outlook goes as far as saying that assessing the impact of AI on a target’s business model is becoming part of modern due diligence and investment risk assessment.

For owners preparing a company for sale, this means the question “how does AI affect my business?” will increasingly become part of the conversation about company value.

What does this shift mean for a business owner?

Above all, it is worth thinking differently about what you are paying an M&A advisor for. Until recently, team size could serve as a proxy for an advisor’s analytical capacity. Today, headcount alone matters less and less. The more important questions become:

• Who will actually run my process?
• Who reviews the analysis that is produced?
• Do senior people stay involved between meetings too?
• Does the advisor know how to use technology properly?
• How is the buyer list built?
• How does the advisor plan to create competition among buyers?
• How will they support me in negotiations?
• Is the way they are paid aligned with my interests?

A professional-looking document is easier to produce today than ever before. What matters far more is whether sound reasoning stands behind it.

Technology increases the scale of analysis, but experience decides what to make of it.

At PROMETEIA we treat AI as a tool, not a substitute for expertise. We use technology where it lets us analyse data, the market and possible scenarios faster. As a result, experienced advisors can stay directly involved in more of the process, instead of relying on a multi-layered team structure.