To cut through the noise, we sat down with elite, experienced AI consultants in our Catalant Practice Community who spend their days working directly with deal teams and portfolio operations leaders and executing AI rollouts on the ground. These are professionals who are out in the field helping firms build their strategic AI roadmaps, wrangle data sets, and introduce advanced technology across mid-market and large-cap portfolios.
Together, we dug into what is actually working and how PE leaders can craft and execute on AI strategies to accelerate value creation without increasing risk. This report is the result: a direct transmission from those who are hands-on in AI strategy and integration in private equity.
Setting the scene
An Environment Reshaped by AI
The private equity market is experiencing a profound operational shift. Firms are feeling immense pressure to keep up with competitors, demonstrate technical modernization, and capture value at an accelerated pace. Market expectations for building new categories and driving growth have compressed from years down to months, weeks, or days.
But behind inspiring press releases about AI rollouts, the landscape is chaotic. At the portfolio company level, many AI integrations are seen as hotbeds of confusion and disorder. Firm leaders know they need to act, but the technology is moving so fast that traditional playbooks have been rendered almost entirely useless.
Old methods of transformation — massive, multi-year initiatives — are outdated. Because AI technology is evolving at a breakneck speed, rigid long-term planning has become a liability rather than an asset. Some consultants are advising firms to abandon AI roadmaps that stretch longer than six to twelve months because the pace of change will make those plans obsolete before they even finish. Instead of waiting to build the perfect, all-encompassing system, firms need to stop over-planning and start moving.
People, especially those with an aversion to AI, tend to judge it based on what it can do today and assume that’s what it will look like in the future. The better question is: what will this look like in 3, 5, or 10 years, and how should that shape what we do today? Build for where it’s going, not where it is.
Forward-thinking players are separating themselves from the rest of the pack by mastering three core philosophical themes:
Strategy must lead technology
Foundations must come first
Humans must anchor change
Here is how they are executing on each.
THEME 1
Strategy Must Lead Technology
One of the most frequent missteps PE firms and their portfolio companies make is letting the tools dictate the strategy. A firm will see a flashy demo for a new AI product, make the purchase, and then scramble to find a use case for it. The question, “Where can we plug in AI?” is fundamentally the wrong place to start.
Sumeet Bhatia, an award-winning data and AI leader, advises firms that “most AI strategies in private equity fail because they start top-down. Real value is created bottom-up within workflows across portfolio companies.” He adds, “embedding AI for the sake of it will not bring value; if you don’t define the problem well, AI will hold you back.”
The smartest firms are treating AI like pain medicine. They start by identifying the major pain points and opportunities within the business, and only then do they ask if AI can enable a faster or more holistic solution. Often, the answer is yes. The key is that the business problem is the focal point.
AI can also be an impactful thought partner for asking these questions, helping to identify where gaps and possibilities exist and what risks may stand in the way of progress. Take, for example, the process of planning for the first 100 days after an acquisition. There is real value in how AI can improve decision quality and execution speed in that process.
10.7x
YoY increase in mentions of portfolio company AI strategy as a value creation priority
Source: Catalant call data
“AI allows firms to expand the range of opportunities they consider and quickly narrow the funnel to what matters most,” shared strategy and innovation consultant Kevin Darbelnet. “AI can simulate growth scenarios or identify operational efficiency opportunities that strengthen a deal thesis, but the real advantage comes from turning those ideas into focused, executable initiatives.”
Post-deal, leaders must balance the reliable wins of the back office with the exponential value of the front office. Back-office administrative automations can be the ultimate low-hanging fruit. They provide quick wins and help stabilize broken internal processes without the complexities of going to market. In the current environment, many firms point AI at one core goal: EBITDA growth.
AI consultant Deva I Justhy specifies that “when a PE firm enters a business, there is a broad set of opportunities available to introduce AI in core areas where valuation can be gained — improvements in safety, efficiency, growth, and compliance goals, specifically. I call these SEGC pillars. When analyzing an enterprise for AI opportunities, these business goal categories and related business processes directly linked to EBITDA impact are the places to start.”
There is also a tremendous amount of value coming from the ability to aggregate portfolio-wide data. By leveraging AI to ingest and normalize scattered information from across portfolio companies, private equity leaders gain a unified view that allows them to negotiate favorable vendor contracts at scale, improve decision-making, and repeat successful AI initiatives across portfolio companies.
Value is often hidden within workflows across portfolio companies. The firms seeing real impact go deep into functions, identify value chain friction points, embed AI, and build repeatable playbooks that scale across portcos. Single use cases create demos. Value chain thinking creates EBITDA impact.
To illustrate where the smart money is actually going — and what your competitors are doing to get ahead — here is a look at some of the most effective applications consultants see being deployed right now:
Firm-level AI applications
Deal lifecycle acceleration:
Firms are leveraging AI across the entire deal lifecycle, utilizing LLM wrappers specifically designed for investment banking and PE, such as Rogo or Blueflame, to analyze market intelligence with limited configuration. At Catalant, we’ve seen a 14x increase year over year in PE clients and prospects referencing AI due diligence alone.
Agentic investment committees:
Firms are beginning to use tools like StackAI to pull in third-party data providers, analyzing new opportunities against historical investment performance and even bringing AI agents in as a functional part of the investment committee process.
Cross-portfolio data, analytics, and insights:
Firms are utilizing internal AI portfolio analytics groups to ingest messy data from diverse portfolio companies, normalizing it into a single, cohesive format. This allows the firm to accurately benchmark performance and recommend strategic initiatives across the board.
Portfolio company AI applications
Go-to-market (GTM) optimization:
We are seeing AI deployed for rapid lead qualification in e-commerce, as dedicated sales assistants in legal SaaS, and more. Instead of operating in a silo, these AI initiatives are meticulously aligned with existing growth strategies to reduce churn, drive efficiency, and accelerate the sales cycle.
Product expansion:
Companies are leveraging AI to steal market share by establishing new capabilities in competitive categories. For example, a company that is historically strong in one area but weak in another can use AI to supplement core offerings, allowing them to instantly spin up low-cost add-ons and increase wallet share.
Specialized knowledge retrieval:
While generic chatbots fail, highly specific ones thrive. Examples include chatbots designed to mine old CAD files for design firms, sift through decades of old service manuals for industrial companies, or allow leaders to interact directly with granular procurement and pricing data.
“Vibe coding” for tactical scraping:
Non-technical operators are increasingly using LLMs to “vibe code” — generating scripts to execute rapid, one-off tasks, like scraping websites for market research. This saves hours of manual labor and bypasses the need for a trained software engineer.
theme 2
Foundations Must Come First
One of the more widely known secrets among AI consultants is that private equity portfolio companies are littered with unusable, unscalable AI prototypes left behind by massive legacy consulting firms.
A common playbook for legacy consulting teams is to take a pre-existing Azure or AWS chatbot platform, do some minor tailoring using offshore teams, and drop the resulting tool into a portfolio company as a “custom” proof of concept. The result? The tool often isn’t properly tested, it fundamentally cannot scale to production, and the employees refuse to use it because it wasn’t designed for their specific daily workflows.
However, not every discarded prototype represents a failure. New tools allow operators to spin up working models rapidly, often using plain English requirements, simply to test ideas. Treating prototypes as disposable proofs of concept can help firms move quickly, prove a solution’s efficacy, and iterate before committing resources to a full-scale build.
One of the biggest hurdles to scaling AI — and a major contributor to many prototypes’ inability to scale — is data and technical architecture. AI is only as effective as the underlying data it feeds on, and many portfolio companies operate on legacy tech stacks with fragile integrations. Building an AI solution with a weak foundation is inherently unsustainable.
Jatin Kamat, a technical product and AI expert, underscores the importance of stabilizing the underlying platform before incorporating AI. “When you walk into a portfolio company with a legacy tech stack and broken processes, you won’t get the outcome you want from implementing AI.” He advises firms to first focus on “getting the platform stable: replatform, build a better tech stack, clean the data, and ensure connections are reliable. Only then can you start applying AI on top.”
The portfolio companies seeing the quickest value from AI are those that already have their data in order. For most companies, achieving this requires a massive cleanup effort and highly specialized data and AI expertise. Central knowledge repositories must be established, and messy, siloed data must be organized and normalized.
Setting this foundation is critical for getting better outcomes from AI. “If you don’t have the right knowledge base, AI doesn’t just fail to create value; it leads to poor decisions at scale,” said Bhatia. “AI amplifies whatever data and context you provide, so getting that foundation right is critical.”
12.5x
YoY increase in mentions of AI talent and workforce needs when discussing value creation priorities
Source: Catalant call data
Even in the most successful deployments, AI isn’t doing all the heavy lifting; rigorous engineering and traditional data structuring drive the core value. Justhy cautions that “today’s most powerful AI models can act like opaque decision engines — impressive, but not fully transparent — so disciplined leaders need to insist on understanding data and assumptions before they bet the business on an output.”
This data reality collides with the constraints of compliance and security. Data privacy is a significant headwind across the industry, with compliance teams often shutting down AI initiatives out of fear that proprietary data will leak or be used to train external models.
To bypass these confidentiality concerns, a trend has emerged: deploying local, off-grid large language models (LLMs) such as Llama or Alibaba’s Qwen. Companies that cannot, for regulatory or operational reasons, upload confidential information to the internet via API-based models are pivoting to these localized systems to ensure their data remains secure while still capturing the operational power of AI.
In all environments, it is not enough to believe that a data foundation has been set up properly or that AI tools are operating correctly and generating appropriate results. AI must be thoroughly tested before it is deployed, especially if it will be interacting with unpredictable human users.
Technical expert Samer Najia warns that testing is “very much an afterthought in a lot of organizations. Companies are focused on delivery, so testing itself can be either incomplete or dependent on seeing if the tool works: yes or no.” Najia advises companies to ask, “How badly can it fail? How does it break? What impact does that failure have?” and to “push the system to its limits to ensure that a simple error won’t compromise the business.”
Ultimately, building a pristine data foundation and effective models is necessary, but it’s only part of the battle. It’s also important to recognize that AI cannot actually think like a human.
There is an idea that AI is a panacea that’s going to solve all of your problems. It’s not. It cannot think for you. Machines don’t have intuition, and they cannot make a bet or have a hunch. You need a human in the loop to make course corrections on a smaller scale so you can make big gains in the long run.
While AI can accelerate the generation of data-driven insights, it cannot replace human judgment.
theme 3
Humans Must Anchor Change
No matter how appealing an AI tool seems, if your teams cannot or will not use it, it is a poor investment. As Darbelnet shares, “companies are investing heavily in AI to generate new value, but leaders often overlook the change management required to get people on board and actually realize that value.”
Time and again, our experts note that one of the primary bottlenecks in AI implementation is not a technological hurdle. The actual selection of tools and the orchestration of agents is the easy part; getting employees to overcome their natural resistance and fear can be incredibly difficult.
Amid all of the hysteria around AI, enablement and adoption consultant Kelle Snow advises organizations to start by “decoding the resistance. Resistance in the form of ‘I don’t want to’ needs trust building, team building, and buy-in. Resistance in the form of ‘I don’t think I can’ represents a lack of technical confidence and a need for education.” Addressing these pain points is one key component of mitigating the fear and uncertainty many teams experience.
To drive real adoption, leadership teams must also establish a culture of understanding with a crisp, unified message. If an executive team is asked why they are implementing AI, what the benefits are, and how it helps employees do their jobs better, they need to tell the same story across the C-suite. If the messaging fractures or is inconsistent, the entire system falls apart.
One way to mitigate the adoption crisis is to move away from top-down mandates and instead identify internal change champions and super-users to drive adoption natively. Leveraging the team members who are driving AI usage to gain buy-in and upskill the rest of the workforce makes the technology seem more approachable and less scary.
Many of the hurdles of adoption are also impacted by the underlying factor of organizational culture. “Is your company AI-first? Is it AI-ready?” Kamat identifies these as “foundational traits that determine whether AI will flourish organically. Yet, organizational readiness and company mindset are two of the main things that get overlooked because of compressed PE timelines.”
At this stage, AI is still so new that most companies haven’t yet had the opportunity to build an AI-ready culture, much less an AI-first one. Doing so requires dedication at every level.
Even more critically, companies have to be ready for change in general. Firms need to cultivate agile structures and mindsets in portfolio companies to support and maintain the pace of change that is necessary to create value. Snow counsels leaders to ask: “Can our organization actually support this? Do we have the right structure, the right change champions, and the agility to adapt? The companies that succeed with AI will be the ones that are built for change, because change is inevitable and AI isn’t going away. This is about building capacity for the many things that come next.”
looking ahead
Where PE Goes from Here
The firms that win in the upcoming years will be those that build dedicated data and AI teams, focus on solving real business pain points, and take a hands-on role in managing the human element of technological adoption.
At Catalant, we are seeing this reality reflected in our current client demand. Firms are no longer asking for abstract advice. They are actively seeking experts for the rigorous management of AI-driven transformation and value creation initiatives, framing AI as a critical EBITDA lever and acting with a high level of urgency.
Trending AI needs
AI support is one of the most frequent needs addressed in our conversations with PE firms of all sizes. In the past quarter alone, AI has been a key topic for more than ⅓ of the PE firms our team has talked to.
The most common AI work types, ranked by frequency of mention in PE client conversations:
AI implementation for portcos | 35%
AI strategy for portcos | 32%
AI due diligence | 29%
AI talent/workforce at portcos | 26%
Function-specific AI | 23%
Gen AI or LLM use cases | 21%
Process automation | 19%
Over time, the technology will only grow more complex. There is likely to be a heavy shift toward model context protocol (MCPs) and agent-to-agent communication, where AI tools across different enterprise platforms begin talking directly to each other seamlessly. But you cannot jump to the future if you haven’t mastered the present. Companies must master the fundamentals: build a pristine data architecture, solve real pain points, empower your super-users, and create the infrastructure needed to respond quickly when the future changes in six months.
The AI landscape in private equity is moving too fast to navigate alone, and the cost of deploying bad technology is simply too high. Whether you are looking to assess your portfolio’s vulnerabilities, stand up an AI Center of Excellence, or finally production-harden those unscalable prototypes, you need battle-tested experts in your corner.
Ready to build real value?
Connect with top-tier AI operators who can turn your strategy into execution.
Meet the Consultants
Learn more about the expert consultants from Catalant’s Practice Community who are featured in this article:
Deva I Justhy draws on 25+ years of data, AI, and digital transformation experience, partnering with PE firms and their portfolio companies to create enterprise value.
Jatin Kamat is a technical product leader focused on AI-enabled value creation who brings 25+ years of experience building and scaling platforms.
Kelle Snow specializes in AI enablement, learning strategy, change adoption, and operating design, helping organizations translate innovation into sustainable ways of working.
Kevin Darbelnet is a strategy and transformation leader with 20+ years of experience leading R&D, innovation portfolios, and engineering programs.
Samer Najia brings 20+ years of technical and strategic leadership experience, which he uses to deliver measurable outcomes in AI, security, and digital transformation.
