AI Tools for Client Sourcing

Aaron Martinez: What are some of the tools you’re using with clients to help identify deal-sourcing opportunities?

Ryan Orton: When we think about AI, we should view it as a way to augment human knowledge rather than replace it. It’s not going to take over the work we do in sourcing, but it will enhance and support it.

One common tool is sentiment analysis. This can include market sentiment analysis or customer sentiment analysis. Sentiment analysis examines written or verbal content to determine the intent and mood of the person providing that content. For example, you can analyze platforms like X (formerly Twitter) or Meta to understand market trends in your product domain or even customer-specific feedback.

This is a powerful early indicator of trends and can help identify opportunities, whether for companies experiencing growth or those facing challenges. It essentially captures the “voice of the customer.” You can also apply it to public information from, say, a CFO. By analyzing quarterly financial statements, you might notice shifts in tone or mood when discussing specific topics. These mood changes could serve as early indicators of underlying trends.

The goal is to gather sentiment from various sources and use it to identify marketplace trends — putting you one step ahead of where you’d traditionally be. Once you’ve sourced this information, you can rank and prioritize your opportunities, helping you focus on the most promising ones while discarding less relevant ones.

Language Model Development

AM: Are language models something companies need to develop themselves?

RO: Large language models can be developed by companies, but they can also be prebuilt. OpenAI is an example of a prebuilt large language model, along with Gemini or Cloud. There are many prebuilt models available.

Companies can also train their own language models. However, doing so requires significant computational resources and extensive datasets, so in practice, most companies opt to fine-tune existing models instead.

AI and Market Sentiment Study

AM: Traditionally, technology has been the domain of IT departments—bringing in experts, sourcing solutions, and handling related functions. But a lot of this is shifting.

Many of the tools we’re discussing today are being used directly by corporate development individuals. You no longer need to be a software engineer or an expert in large language models (LLMs) to leverage these tools effectively. What’s critical, however, is understanding how they work and how to use them to your advantage.

There’s another skillset that needs to be developed to understand how to effectively use these models. How do you use market sentiment in a practical way?

Chris Bernert: These tools are high-powered solutions. In my experience, they help chart the path for growth and ensure time is spent wisely. These tools enable decisions that might otherwise be influenced by bias or incomplete information.

When we talk about customer analysis, like the example with physician practices, these tools expand how you can engage with suppliers, vendors, and other stakeholders. Instead of relying solely on transactional relationships, you can gather insights from the entire stakeholder group to map out your deal-sourcing process more effectively.

This is a major shift from the traditional approach, which often relied on personal networks and informal conversations. For example, in family-founded businesses like mine, deal sourcing used to be, “I know this guy, and he knows that guy.” It was all about trust based on conversations. But now, tools like these allow us to modernize that process, making it more reliable and data-driven.

AI and Customer Cohort Analysis

AM: How can we use AI to dive deeper into customer cohort analysis?

RO: For some perspective, think about diligence 20 years ago. Back then, a company might have had 10 terabytes of data — roughly equivalent to 500 hours of HD streaming. Today, companies are dealing with 10 to 100 petabytes of data, which is comparable to the storage capacity required for every Netflix movie ever streamed, many times over.

To manage this immense volume of data, tools like machine learning and AI are essential. On top of that, tools like DatabaseUSA provide additional customer information, including demographic and psychographic data, which can be extremely valuable.

AI automates the analysis of this data, significantly reducing the time needed for tasks like cohort analysis. What used to take weeks or months can now be done in a single day. Additionally, these tools allow you to handle the sheer scale and complexity of data, enabling deeper insights into behavioral patterns. For example, you can identify which customers are likely to churn, which are likely to be retained, and calculate customer lifetime value.

The major shift here is moving from backward-looking analysis to forward-looking predictions. Instead of analyzing historical data alone, you’re using insights to predict the future performance of the company you’re buying. This approach helps you understand the true value of the company, not just its past performance.

Once you understand customer lifetime value, it feeds into the brand value and overall valuation of the company. This future-focused analysis is a game-changer in the diligence process.

AI and Preparing for a Sale

AM: We often talk about synergy analysis in transactions, and it’s usually focused on the cost structure of the business. For example, do we need two CFOs? These are the more tangible synergies that are easy to identify and quantify.

What we’re discussing here, however, is true revenue synergy, which has always been a more elusive concept. It’s difficult to predict how customers will respond to a merger or how to effectively combine two companies to unlock additional value. While we can talk endlessly about creative strategies for the future, having a predictive tool with AI could provide real insights and help quantify the potential impact.

How can we use AI to help identify the value of a synergy upfront?

CB: Years ago, someone told me, “M&A is the most expensive way to grow.” You need to identify opportunities and then integrate them, which can be costly and complex. But if you can improve your approach to diligence, you’re ahead of the game.

When I think about cohort analysis, we often talk about selling product A alongside product B and product C. If you can prove the value of that strategy with AI and data-driven horsepower, rather than relying on someone’s experience or intuition, you start to unlock exponential growth. You’re no longer limited to incremental gains; you’re aiming for results where “one plus one equals three” or even “one plus one equals ten.”

AI can help with this kind of analysis, allowing you to validate your assumptions and drive decisions with data. In the past, I’d present analyses to very smart people, and they’d respond positively because of my background in the public safety business and my ability to tell a compelling story. But the real challenge wasn’t about the past—it was about proving the future.

If you can leverage AI tools or smarter customer analysis to demonstrate future potential, you can confidently sell or buy something. This makes private equity groups happy and enables you to scale acquisitions more efficiently.

AI and Valuations

AM: How can you utilize AI to support valuations?

CB: AI can help identify the abstract shape of an asset. It helps with figuring out comps, ensuring they’re objective, directly comparable, and assessing how comparable they actually are.

That said, valuation still involves a degree of art, and there’s always a negotiated component to it. For now, I’m primarily using AI to identify blind spots, reduce upfront costs, and mitigate risk. These tools help facilitate better conversations with founders or investors by providing data to back up our positions.

The Cost of AI

AM: Do businesses need to make big capital expenditures to implement these kinds of AI tools? How accessible are they?

RO: The good news is that you don’t need to make huge capital expenditures. As Aaron mentioned earlier, this isn’t really the domain of the IT department anymore. AI now tends to fall under the CFO’s domain because most of these tools don’t require as much coding.

However, while many AI tools are low-code or no-code, more sophisticated/complex implementations often require some degree of customization and integration.

It’s a lot like putting together a puzzle. There are many tools available, and these tools can be used for customer lifetime value analysis, revenue forecasting, and everything else we’ve discussed today. They are designed for financial analysts and business analysts rather than IT professionals. If you’re looking to do forecasting or similar tasks, these are two powerful tools to consider.

You can download these tools and start using them right away. They typically have simple interfaces, starting with prompts like, “Do you want to input data?” Once you input the data, it guides you through steps in a process flow. For example, you might change column headers or numbers, and then decide what type of analysis you want to perform next. It’s all drag-and-drop functionality, making it straightforward to use while still being powerful. Although it can get complex, it’s typically quite user-friendly.

Another example involves large language models, which can either start blank or come pre-trained with specific knowledge. You can also train them to specialize in particular tasks. For instance, if you need a model specialized in financial data, you can download it to your laptop, and it functions as a “brain,” a large language model highly proficient in finance and financial terms. This makes it excellent for analyzing historical company data or asking financial questions.