The Role of AI in Search Fund Operations: How Are We Adapting?
May 26, 2024
by a searcher from Maastricht University - School of Business and Economics in Fribourg, Switzerland
I would like to have a discussioAs we continue to navigate the complexities of acquiring and operating small to mid-sized businesses, it’s impossible to ignore the rising influence of artificial intelligence (AI) in our daily operations. From enhancing decision-making processes to streamlining administrative tasks, AI is becoming an indispensable tool. However, this rapid integration brings with it a host of challenges and opportunities that are worth exploring.
I’d like to open up a discussion on how AI is currently being utilized in our community and what the future might hold for search funds in this context.
Decision Support Systems: How many of you are using AI-driven analytics to inform your investment decisions or to identify potential acquisition targets? What tools are you using, and how have they impacted your success rates?
Operational Efficiency: AI can significantly automate routine tasks, from accounting to customer service (via chatbots, for example). Are there particular areas in your operations where AI has made a substantial difference? Have these tools integrated seamlessly with your existing processes?
Due Diligence: The due diligence process can be both time-consuming and costly. Are there AI platforms you've found effective in speeding up or enhancing the thoroughness of your due diligence? How reliable have these tools been?
Personalized Marketing: AI’s capability to analyze big data can offer unprecedented personalized marketing strategies. Has anyone leveraged AI for more targeted marketing in their portfolio companies? What results have you observed?
Challenges and Ethical Considerations: With all its benefits, AI also brings challenges such as job displacement, privacy issues, and the need for new types of management skills. What are the ethical considerations we need to discuss? How are we preparing our workforce to adapt to these new technologies?
Future Prospects: Finally, where do you see the role of AI heading in the context of search funds? Are there particular technologies on the horizon that could revolutionize our industry?
I encourage everyone to share their experiences, concerns, and predictions. Whether you’re an AI skeptic or enthusiast, your insights are valuable as we try to understand the broader implications of this technology in our field.
Looking forward to a lively and informative discussion!n how you use AI tools for your activities. Does it have an impact and if so - how?
from INSEAD in Singapore
- When you have a sledgehammer, everything looks like a nut--but in many cases it isn't the appropriate tool to be using and you can cause lots of unintended damage. We see more benefit from businesses creating targeted, deterministic, operational automations (e.g., intelligent use of Zapier or similar orchestration/workflow automation tooling), buying off-the-shelf SaaS tools that have integrated narrow LLMs to deliver specific product features (e.g., summarisation of conversations in a CRM) etc.
- Our strong guidance for any searcher/operator looking to use inherently stochastic LLMs as part of their workflows is to be crystal clear about the bounds in which you let it operate, and the risks of its actions/guidance/output. The business can be held liable for the LLM's decisions (e.g. https://www.forbes.com/sites/marisagarcia/2024/02/19/what-air-canada-lost-in-remarkable-lying-ai-chatbot-case/?sh=45bc10b9696f) so it isn't a simple matter of set and forget.
- Data Science is nothing new, and historically the vast majority of projects have failed--that is 60-90%. Gen AI/LLMs are a data science technique that deliver very layperson friendly output, but don't fundamentally differ from the many other DS approaches used in the past in that they consume data, look for patterns, and generate an output. Searchers/Operators need to be vigilant for the drivers of traditional DS project failures, i.e., poor problem definition, low input data quality, integration into existing systems, organisational buy-in, process and change management, ongoing monitoring and upkeep if they are to deliver what is still fundamentally a DS project even if it wears different clothing.
Happy to chat through any of these points in more detail if they stimulated thoughts.
from Stanford University in Pleasanton, CA, USA
Stanford's latest report on AI is interesting and has a lot of great information https://aiindex.stanford.edu/wp-content/uploads/2024/05/HAI_AI-Index-Report-2024.pdf
For searchers & businesses, AI2024 is good enough to automate repetitive, knowledge-based tasks such as prospecting, lead gen etc. If done well, it can give that searcher or organization a huge advantage over their competitors. Searchers & businesses can also drive more traffic in by optimizing their websites for AI search engines.
Regarding personalized marketing, I think one can try using AI2024 tools to research and summarize information about targets & prospects. But they have to be very careful how that is used. It is because data may not necessarily be accurate. For example, when I searched myself on Google's Bard AI, it gave a rich, detailed history of my life. But everything on that - other than my place of birth - was inaccurate and completely made-up/hallucinated. 😀
AI2024 is good enough for investors to have it scan data rooms and have it highlight or flag items that they usually use to make investment decisions. In my opinion, AI2024 is not remotely close to automating QoE, DD and other deal related tasks.
This is a pretty large area and is also rapidly evolving. So, things can change pretty quickly and could invalidate some of the issues that I mentioned above.
Hope this helps. If you would like to discuss further, please go ahead and DM me.