Mergers and acquisitions (M&A) fail at an alarming rate, missing the mark on projected deal value an estimated 75% of the time. This raises a critical question: What drives such widespread underperformance?

A key factor may have to do with how company executives allocate their time across the deal cycle. Executives tend to focus on the early stages — target selection and valuation — while relegating integration, the final phase, to unit and functional leaders. This dynamic may also explain why AI tools have been disproportionately deployed in the pre-deal phases, sidelining their transformational potential in integration — a missed opportunity with significant implications.

In a recent interview, NYU Stern Professor Baruch Lev offered a useful heuristic for understanding M&A success informed by an analysis of 40,000 deals: Target selection drives 45% of outcomes, pricing contributes 20%, and integration accounts for 35%. In his words, “Integration is the key.” This breakdown highlights the importance of using AI during integration as aggressively as in pre-deal phases.

Organizations that extend AI’s reach beyond the deal’s opening acts stand to significantly improve their odds of success, given a third or more of overall deal value hinges on successful integration.

Why AI stands to transform M&A integration

Integration teams face a dual burden: They must resolve a constant stream of strategic decisions while simultaneously managing an overwhelming array of time-sensitive, often manual tasks. This balancing act is often complicated by the combined company’s nascent and untested leadership structure, reporting hierarchies, and decision-making processes.

The team’s ability to efficiently resolve high-stakes decisions — such as product portfolio optimization, pricing strategies, organizational restructuring, and systems integration — becomes a key dependency for the timely completion of the integration and synergy capture.

AI presents a transformational opportunity in this context. By automating and streamlining many of the labor-intensive integration activities, AI can free up valuable human capital, allowing integration teams to focus on higher-order tasks. This way, AI has the potential to significantly increase transaction efficiency by accelerating the overall integration process.

From harmonizing policies to creating tailored communications, the numerous AI use cases for integration are promising, offering a powerful lever to achieve higher returns.

Where to start?

To unlock the transformational potential of AI in integration, organizations must take bold yet measured steps toward experimentation. Companies that have already invested in robust data quality and process documentation are particularly well-positioned, as these assets provide the high-quality inputs necessary to generate useful AI outputs.

For companies that are still early in their data quality journey, a lack of clean data and standard documentation can be a barrier to progress. The good news is that even small steps — such as starting an inventory of critical data elements, implementing basic master data governance procedures, or documenting key processes — can lay the groundwork for future AI success. By investing in data quality, governance, and process documentation now, organizations can steadily improve their AI readiness and unlock greater value from integration efficiencies down the road.

Further, organizations with mature M&A functions have the advantage of well-established integration frameworks, taxonomies, and playbooks, which help put all integration activities in plain sight for assessment against AI capabilities. By using these assets to systematically explore ways to apply AI to predefined activities, companies can fast-track AI-driven value capture in the integration phase.

All that said, even organizations starting from less mature foundations can reap substantial benefits from the efficiency gains AI offers.

Regardless of the starting point, a practical approach is to adopt a “think big, start small” mindset. Rather than waiting for a comprehensive AI strategy or an off-the-shelf vendor solution, companies should begin experimenting with targeted, high-impact use cases designed to address specific pain points or inefficiencies.

Over time, successful use cases can be further enhanced and woven into the integration playbook, ultimately creating a more digitally-enabled integration capability.

Identifying AI use cases for M&A integration

Some integration teams will eagerly dive into AI experimentation, while others may hesitate or struggle to find a starting point. Regardless of a team’s enthusiasm, a structured framework helps approach AI adoption systematically, prioritizing the highest-opportunity activities for maximum impact. To help integration teams identify prime opportunities for AI applications, consider the following five-question framework:

1. Data comparison and harmonization

Does the activity involve comparing multiple sets of data or information? AI can help analyze and harmonize disparate datasets, such as customer and vendor lists, company policies, and organizational ranks. AI can be used to compare artifacts from different companies or units, analyze differences, and suggest harmonized versions.

2. Text synthesis and summarization

Does the activity require synthesizing or summarizing text? AI can efficiently distill key information from multimodal inputs, including text, audio, and images. This capability can help create executive summaries, action item lists, and decision logs from lengthy meetings or workshops.

3. Integration artifact generation

Does the activity generate an output for which past examples are available? By leveraging documentation from prior integrations or similar projects, AI can create templates and draft documents such as service level agreements (SLAs), workstream charters, job descriptions, and status reports. This not only saves time but also ensures consistent deliverable quality across integration activities. 

4. Stakeholder communications

Does the activity involve communication? AI can craft tailored communications for various stakeholders — think press releases, earnings call commentary, executive memos, or employee updates — by stitching together project information from multiple sources into coherent, audience-appropriate messaging.

Additionally, retrieval-augmented generation (RAG) solutions, which combine generative AI with dynamic access to relevant organizational data, can help internal stakeholders access key information (e.g., specifics of updated policies, or how to use new systems) through intuitive chat interfaces reducing the burden on teams tasked with executing change management for the integration.

5. Stakeholder simulations

Would the activity benefit from a stakeholder simulation? AI can create sophisticated persona simulations based on available data, such as employee psychometric profiles or an inventory of analyst earnings call questions. These simulations allow teams to test stakeholder reactions to new policies or announcements before implementation, enabling more informed decision-making.

At this point, it’s important to note that all the known limitations of current AI models — imperfect pretraining data, hallucinations, inaccuracies, and other errors — still apply. A qualified human should always be “in the loop” for each activity, providing oversight and carefully reviewing and refining all AI outputs.

Masterstroke: Converting individual successes to a digital integration capability

To drive AI enablement and greater digitization of the integration process, business leaders must establish infrastructure that makes AI both safe and easy to use. Experimenting with AI on targeted use cases across the integration lifecycle should be an explicit expectation of the integration team.

Furthermore, experimentation should include formal touchpoints where teams share use cases where AI helped or failed, capturing learnings to apply going forward.

The true masterstroke lies in establishing protocols for workstream leads and key team members to share successful AI use cases with the Integration Management Office (IMO) or a similar governance body. This way, successful use cases can be disseminated across teams, cataloged for future use, and made available for peers to improve.

Upon integration completion, team members and leaders should hold workshops to formalize updates to the integration playbook by incorporating AI impacts. This ensures that individual AI use cases can evolve into a digitally transformed integration capability. To gauge this evolution, leaders from the IMO should define and track metrics — such as time saved, cost reductions, or synergy realization rates — while senior executives regularly review these results during and after integration completion.

Stop leaving deal value on the table

Although executives may invest more time in pre-deal phases, they must address the dearth of AI usage during integration — a critical phase of the deal cycle that contributes an estimated 35% of overall deal value. Following the approach outlined above — mandating experimentation, systematically identifying opportunities, and formally incorporating successes into the playbook — business leaders can ignite the process of digitally transforming their integration capabilities, unlocking significant efficiencies, and increasing their ability to consistently deliver projected deal value.

About the Author:

Joe Sagrilla is an independent management consultant and business advisor, top business school faculty, Board member, writer, and speaker. His specialties include business strategy, technology, transformation, process improvement, and organizational performance. He currently lives in Austin, TX.