|
|
|
|
Guest: Brian Mergelas, a coach at CEO Coaching International. Brian is an accomplished leader with over 25 years of experience in driving growth, innovation, and operational excellence for global technology companies and startups.
Quick Background: As AI usage becomes more and more mainstream, companies that didn’t have the foresight to include an AI road map in their last few annual planning sessions are scrambling onto the bandwagon. But if your company is in catch-up mode, you’ll never gain ground in this rapidly evolving space without a plan for analyzing your options, building up your infrastructure, and streamlining integration.
On today’s show, Brian Mergelas explains how leaders can use AI to unlock value in their businesses. Brian shares his insights on developing an AI strategy, navigating its challenges, and its potential to revolutionize industries, including yours. We also explore practical examples of leveraging AI for process improvements, innovation, and decision-making to Make BIG Happen.
4 Keys to an AI Upgrade from Brian Mergelas
1. Understand the tech.
No, generative AI can’t really paint an original “Picasso,” write a new “Shakespeare” play, or run your business plan past “Warren Buffett.” Essentially, ChatGPT, Google Gemini, and other consumer-grade chatbots are just really, really good at math and guessing.
“AI tools are not really creating new knowledge,” Brian Mergelas says. “What they’re doing is they’re recombining things. Really artificial intelligence is all about predicting. The large language models (LLMs) have the benefit of being trained on an enormous set of data — basically everything that’s on the Internet — but it’s math. And so what they’re doing is predicting the next word in a sentence or the next sentence in a paragraph or the next paragraph in a text by looking at what the probabilistic outcome is, what the most likely response should be to a question.”
At higher levels of processing power and complexity than LLM chatbots, AI’s “guesswork” can be extraordinary. For example, medical AIs are helping scientists discover shortcuts toward new drug development and early detection of diseases. Financial firms are using AI to optimize their client’s portfolios and stay ahead of the markets.
These more advanced integrations tend to be proprietary tools that companies develop in-house or with a third party. In addition to encouraging their teams to experiment with commercial AI, like chatbots, CEOs should also be exploring what kinds of problems custom AI could solve for their businesses.
“It’s all very customized,” Brian explains, “because the data that Company A is going to have, the format, the structure, what it’s looking at, the calculations that need to be done in the background, are going to be different than what you’re going to see for Company B. Some of CEO Coaching International’s strategic partners can help companies build those internal applications and to do the QA/QC to make sure that the results are going to drive the response they’re expecting.”
2. Refine your inputs.
As much as AI has advanced in recent years, “garbage in, garbage out” still applies in order to generate the most useful formation and avoid inaccurate “hallucinations.”
“ Think about the first interactions that we had with an AI,” Brian Mergelas says. “We probably all went through like a Google search. You type in something and you get something back and you go, ‘Well, that wasn’t very interesting.’ Interacting with AI is not meant to be a ‘one and done.’ It’s meant to be a conversation. You shouldn’t be looking for that one killer question that’s going to give you all of the answers. It’s having a conversation with that guy at the party. He might say something that sounds really interesting and you can dig into it. ‘Please explain a little bit more about this. Can you give me some references for where you got that?’ And by asking a series of questions, you can end up with a better set of results.”
Asking follow-up questions or breaking down multipart questions into individual chats force the AI to generate deeper responses about the more intricate details of your overall query. It’s also important to remember that much of what generative AI does is guessing what it thinks you want. So if you ask ChatGPT a leading question, it might pick up on some unintentional cues and hallucinate a bad answer.
“If you ask AI, ‘What is the best seafood restaurant in this town?’ it might come up with names like ‘King Prawn Seafood,'” Brian explains, “when the correct answer is, ‘There are no seafood restaurants in this town.’ The AI didn’t know that because you asked it a leading question. And so it’s more likely to the AI that a fictitious restaurant name is the best answer, as opposed to saying that there is no restaurant. If, instead, you ask, ‘Using information that you can verify, please give me a listing of restaurants that have seafood on the menu,’ it might come back and say ‘There are no restaurants that have seafood on the menu.’ Same series of questions, but completely different answers.
3. Establish guardrails.
Some LLMs and chatbots have also improved their outputs by “remembering” conversations and by allowing users to upload additional information for them to analyze and incorporate into answers — such as PDFs or spreadsheets. Extra inputs can help AI generate answers that are more unique and specific.
For companies, however, these options raise some serious privacy concerns. Even if your company operates behind a firewall and stores info on its own servers, it’s likely that anything your employees upload to an LLM is going to be stored on the LLM’s servers. That means, theoretically, a competitor could ask the LLM a question and receive your company’s data as part of the answer.
“If you’re uploading documentation, do not upload proprietary information,” Brian warns. “Some companies, like Open AI, do provide APIs, which allow you to use the same model, but then the data is proprietary. So you can, in essence, build a contained version of ChatGPT that’s locked around your company. We at CEO Coaching International can help you set that up. But you absolutely want to make sure you take privacy into consideration when you get into the fine-tuning.”
Perhaps the strongest safeguard CEOs have at their disposal, both for privacy and quality control, is their teams. AI can draft a marketing campaign, but you’ll still need a human marketer to embed the soul of what you’re trying to communicate and the values that drive your company. AI might plow through your data in an instant, but you’ll need your executive leaders to follow that data towards solutions that improve the end experience you deliver for customers and employees alike.
”AI can do many things that people can do,” Brian says. “If you’re trying to summarize a document, or you’re trying to compare and contrast information, there are lots of use cases where AI is going to be more efficient. But it doesn’t replace the critical thought processes that we have. You can use AI to help educate you on a topic, and you can easily get to a 6 in terms of understanding. But if you’re dealing with an expert in that topic, 6 or 7 is actually pretty poor. So you still need a few experts, maybe with AI helping augment some of the other people that don’t have the expertise.”
4. Upgrade today.
AI isn’t magic. It’s not going to fix a leaky supply chain. It’s not going to broaden your customer base or accelerate your cash flow cycle.
But, with proper implementation and analysis, it can help your company achieve these and other essential tasks quicker and more efficiently.
If your company has made progress down your AI roadmap, start talking to your tech team about identifying the next upgrade.
And if you still haven’t grabbed low-hanging fruit like letting AI draft emails or take notes during meetings, stop waiting and start doing.
“First and foremost, don’t ignore AI,” Brian Mergelas warns. “There’s a head-in-the-sand attitude that sometimes exists. And on the other side, CEOs shouldn’t jump on the bandwagon believing that it’s a panacea. There are some areas where it’s going to add tremendous value. If you understand your business processes and you map them out, then you should look at AI in the context of, ‘With the new tools that exist, why are we doing it the way we’ve always done it?’ CEOs should be looking at engagement, pilot programs, use cases. Don’t do them all, but get started. If something doesn’t work, fine, find another application. But to ignore AI will be the death of businesses. You have to have some idea of what’s coming.”
Top Takeaways From Brian Mergelas
1. Stay ahead of the curve by educating yourself on how AI is developing and test-driving new applications.
2. Protect your data with a combination of established best practices and human oversight.
3. Work with experts who can coach you on the best AI integrations for your company’s specific needs.
Links:
The AI Imperative: Why CEOs Must Act Now to Harness Artificial Intelligence – At the 2024 Make BIG Happen Summit, Peter Diamandis issued a stark warning: “There are going to be two companies at the end of this decade: those that are fully utilizing AI, and those that are out of business.”
How To Turn AI Hype Into BIG Profits by Executing a CEO AI Roadmap – CEO Coaching International’s Randy Koch gives an overview of the process CEOs can use to build their own CEO AI Roadmaps, leverage AI to its full potential, and Make BIG Happen.
About CEO Coaching International
CEO Coaching International works with CEOs and their leadership teams to achieve extraordinary results quarter after quarter, year after year. Known globally for its success in coaching growth-focused entrepreneurs to meaningful exits, the firm has coached more than 1,500+ CEOs and entrepreneurs across 100+ industries and 60 countries. Its coaches—former CEOs, presidents, and executives—have led businesses ranging from startups to over $10 billion, driving double-digit sales and profit growth, many culminating in eight, nine, or ten-figure exits.
Companies that have worked with CEO Coaching International for two years or more have achieved an average revenue CAGR of 31% (2.6X the U.S. average) and an average EBITDA CAGR of 52.3% (more than 5X the U.S. average).
Discover how coaching can transform your leadership journey at ceocoachinginternational.com.
Learn more about executive coaching | Meet our world-class coaches
