“For many in management, AI is just one more thing on an already overflowing plate — not a priority until forced to confront it.” That’s a candid take from one of the many General Partners, Operating Partners, and PE-backed CFOs gathered at a series of recent Accordion-hosted roundtables addressing the use of AI within portfolio companies.
Another take? “PE partners need to force our portfolio teams to confront AI…now. Because if we’re looking for an exit, we can’t not do anything.” To put an even finer point on it: “If we want that exit to be successful, we should be pushing them to think beyond efficiency plays – to the bigger bets they can make with AI.”
For portfolio companies, AI isn’t a nice-to-have; it’s a need-to-have. The question is no longer if portfolio teams use AI, but how they re-think the way their business scales with AI. And how, exactly, should portfolio companies use it for scaling? Here are 5 key takeaways from the roundtables.
1. Don’t just use AI; structure for it.
According to GPs, lean ops, smart infrastructure and basic AI fluency are the launchpad for real advantage.
“We had AI pilots everywhere but no consistent data strategy or operational model. It felt like building castles in the air.” The real breakthrough came only after the company simplified its operating processes and invested heavily in data infrastructure.
AI governance isn’t something to solve down the line. The moment you start doing AI work is the moment you should be shaping org-wide AI policies, capabilities, governance, strategy, and process. If you wait, you will be too late.
For example, a portfolio company in logistics slashed AI deployment time by 40% by standardizing data pipelines and training frontline managers on AI basics. This lean ops foundation enabled AI tools to generate actionable insights without overwhelming teams.
“If your team can’t even speak the basics of AI, it’s a non-starter.” Basic AI fluency must be the common language to unlock collaboration and faster decision-making.
2. Build AI based around your own data (differentiation won’t come from generic models)
You should be looking to transform your data into a decision-making engine. Ask yourself: Can I use my data to answer my most impactful business questions and make the kinds of decisions that drive outsized value creation?
The right engine will accelerate ongoing value creation through the hold period and keep you AI-ready for the next generation of technology.
A PE-backed insurance company moved beyond generic models by integrating proprietary claims data to tailor risk assessments uniquely to their book: “Off-the-shelf AI might give you a 5% lift. Our customized models delivered 15% better loss prediction accuracy—that’s a game changer.”
Another example came from a retail portfolio company that trained AI on its customer purchase history to design hyper-targeted marketing campaigns: “Our competitors used the same vendor models, but our data gave us a completely different lens on customer behavior.”
As for how they are developing proprietary data-based AI, clients are turning to a variety of sources. One manufacturing firm, for example, has piloted a proprietary data partnership with the University of Texas. At Accordion, we’re working with clients to leverage their data to deploy industry-specific Agentic AI and machine learning tools across areas like market/customer research revenue, predictive forecasting, inventory management, quoting and scheduling, and sales prediction.
3. Think beyond AI for efficiency; it should be adopted for product strategy
Here we are talking about organization-wide AI use cases across operations, pricing, customer success, revops, procurement, and more.
One PE-backed insurance company leveraged AI to enable the launch of a dynamic pricing product that adjusts coverage in real time based on behavioral data.
“We didn’t just automate underwriting—we created a product that adapts and learns with the customer.”
Similarly, a healthcare portfolio company used AI to personalize treatment recommendations, moving from a cost-saving tool to a competitive product differentiator: “AI helped us shift from ‘how do we save money?’ to ‘how do we deliver better outcomes?’”
Finally, a manufacturing firm with a product offering requiring bespoke drawings per order has used AI to help automate production, thereby accelerating the product design and client delivery process and “transforming [the company’s] ability to meaningfully scale.”
4. AI bets should match the investment horizon
“If you’re in for the long haul, invest in deep AI transformation—data lakes, new workflows, AI-native products. But if the exit is in sight, focus on optimizing what’s there. Don’t waste money chasing moonshots.”
For example, a software portfolio company with a 5+ year horizon invested heavily in embedding AI capabilities into their core platform. Conversely, a consumer brand preparing for exit focused on AI-driven marketing optimizations that improved margins immediately without disrupting operations.
In other words: sponsors want their portfolio companies’ AI strategy to align with the investment hold period.
Long runway? Get your data in order. Clean, reliable, accessible data is the key to building your own AI solutions and applications.
Near-term exit? There are three paths you could be working on today:
- Maximizing your existing technology (think CPM, EPM, ERP, Salesforce) – Use AI features within your existing tech stack.
- Developing your own AI solutions and applications using existing pockets of clean data.
- Implementing new tech to improve accounting and finance processes with AI (keep reading below).
5. Finance teams need to do more with AI…now
“Finance teams are the last to get AI attention but really need to be actively driving adoption.”
What are the ways finance should be using AI? Sponsors say they include AI-enabled models that forecast cash flow more accurately, leveraging AI-driven scenario analysis to better manage working capital and investment decisions, and deploying AI-powered bots to automate invoice processing, freeing finance staff for strategic analysis.
We’ve developed Accordion Intelligence, a dedicated service to help CFOs leverage AI to scale their business. As a first step in that journey, we’ve identified 5 discrete ways that PE-backed CFOs can use AI to improve workflows right now. They include:
- Close automation: Solutions to expedite, streamline, and improve visibility into close processes through AI-powered close checklists, automated reconciliations, and flux analyses.
- Cash flow forecasting: 13-week cash flow models leveraging machine learning tools to analyze historical baseline, recommend forecasting methodologies, populate standard outputs, and analyze variances and cash management opportunities.
- Invoice to cash: The introduction of automation and GenAI to expedite full invoice-to-cash processes, including credit checks, invoice matching, deductions, and collections correspondence with customers.
- Contract intelligence: GenAI is leveraged to automate extraction and analysis of supplier contractual terms. This enables the comparison of pricing, invoicing, and payment terms to actual operations, and the identification of cost and working capital improvements.
- Sell-side readiness: These are solutions that extract, load, and transform data from disparate sources and rapidly synthesize them into a sell-side data cube, with industry specific KPIs and benchmarks.
But while these are great (perhaps even critical) ways to start – they are only the beginning of the use of AI in finance.
At the end of the day, the goal for the CFO should not just be efficiency but scaling, revenue growth, and value creation: “AI can allow us to move from Finance as a cost center to Finance as a growth enabler.”
Another participant went further in terms of articulating the impact of AI on the role of the CFO: “Embracing enterprise-impacting AI for go-to-market uses cases, value creation, and product differentiation can elevate the CFO from a financial operator to a true strategic partner to the CEO.”
PE firms have reached a decisive consensus: AI is no longer optional for portfolio companies—it’s essential.