The power of Generative AI is unfolding at unnerving speed. Getting the most out of the generative AI revolution, no matter what field you’re in, requires a reframed way of thinking about technology that’s unlike anything that has come before. To help get CEOs and their teams tackling this transformative tech, Chief Executive pulled together
The post Special Report: Practical AI for Everyone appeared first on ChiefExecutive.net.
The power of Generative AI is unfolding at unnerving speed. Getting the most out of the generative AI revolution, no matter what field you’re in, requires a reframed way of thinking about technology that’s unlike anything that has come before.
To help get CEOs and their teams tackling this transformative tech, Chief Executive pulled together a group of all-star AI players to form the faculty of our AI360Academy—you’ll meet them on the pages that follow. These aren’t your usual academics, consultants and theoreticians. These are player-coaches, with day jobs and decades of experience in their areas of specialty, who got the gen AI bug early and found creative new ways to deploy it in core areas of any company’s operations: sales, marketing, HR and finance.
Over the course of eight virtual sessions last year (replays: chiefexecutive.net/ai360academy) they helped more than 1,000 executives accelerate their understanding of how to use AI to get better at what they do (we’ll have more in 2024).
During all of that, a few key takeaways emerged: The best way to win is to play—literally. ChatGPT and other platforms reward curiosity and experimentation over precision and sophistication. They also demand—like any other good employee—clear communication and continuous feedback. Don’t get hung up on nailing the “magic prompt.” Despite the hype, it doesn’t exist. You just need to keep iterating and trying—and being clear about what you want.
Also, when it comes to using generative AI, those with the most to gain are those with the most talent. ChatGPT—at least for now—doesn’t know excellent work from pure dreck, and its output is entirely reflective of the input. Got someone who can’t write great ad copy? ChatGPT can’t make them David Ogilvy. But if you’ve got a master wordsmith with the confidence to shake hands with the future, they’re going to love being set free from menial work—and you’re going to love not having to add another mouth to feed.
Discovering the AI Mindset
Conor Grennan, AI Advisor to Companies; Dean of Students, NYU Stern School of Business and Head of GenerativeAI@Stern
AI promises efficiency, innovation and unprecedented growth for companies across industries—if only they can get it right. Few, however, are taking the right approach to realize that potential, notes NYU Stern School of Business’s Conor Grennan, one of the most influential and pragmatic thinkers working with ChatGPT today. “People think about it as a tool that the company has to slot in somewhere, but it’s really a completely different way of working,” he says. “Only a tiny percentage of people are using AI effectively.”
Grennan offers these tips to help companies develop a framework for participating in the generative AI revolution:
AI is more, far more, than Google on steroids. “Our brains are wired to, when we see something new, put it in a category, so we say, ‘Oh, this is Google,’” explains Grennan. “But if you’re using ChatGPT the way you use Google, not only are you not getting enough out of it, you’re using it in the exact opposite way it should be used.” Instead, treat it like a person, a coworker with whom you’re collaborating, rather than a tool that you plug commands into. “If ChatGPT offers you five ideas, give it feedback, just as you would in a brainstorming session with colleagues,” says Grennan. “You’ll get to your answer faster if you tell it what you like and what you don’t.”
Know, but don’t obsess about, the risks. Some companies are so focused on data protection policies that they focus on developing policies about employee use of AI at the expense of understanding its capabilities, warns Grennan.“Regulation protecting proprietary data is important, but there’s a sense of getting on a plane and hearing the whole safety presentation and then having to get off before the plane leaves,” he says. “There are real risks, but your focus needs to be on understanding how much you can do in this space without getting yourself in trouble.”
Invest the time to learn. The rewards for companies able to approach AI with curiosity, wonder and respect will be profound, says Grennan, who urges business leaders to spend time engaging with the technology and to encourage their employees to do the same. The good news? You don’t need any expertise in technology to dive in, he says. “In my experience, it’s not an understanding of technology that helps people do well with AI right out of the gate. It’s two things: creativity and intellectual curiosity.”
Using AI In Sales
Thomas Cheriyan, Director of Revenue Enablement, Rattle; Adjunct Professor, City College of New York
However good a company’s product may be, improving its sales process—something AI is uniquely equipped to do—will improve its prospects, says Thomas Cheriyan, director of revenue enablement at Rattle. He cites several areas where ChatGPT can be a competitive differentiator in sales:
Personas. AI can help sales teams create foundational content around personas, including clear definitions of key personas and ideal buyer profiles, as well as information on situational needs, pain points, challenges, goals and objectives faced by each persona. “Just remember to provide background on your company, because even if it’s been around for 100 years, ChatGPT may have no idea what it is all about, and any information it does have will be dated,” cautions Cheriyan.
Refining role-playing. Cheriyan also suggests providing ChatGPT with background about your company and its sales products, assigning it an identity and asking it to role-play with a group of your sales team members. “I tee this up by saying, “ChatGPT, you are now a potential client named Susan, and I am the sales rep at Rattle. I am going to cold-call you with the goal of getting a meeting. On a difficulty level of one to 10 with 10 being difficult, I want you to be an 11 and not give me the meeting,” he says. “It’s a great drill where the sales team reps co-participate in knocking down each objection. They’re practicing the basics and getting better at it all day.”
Support and development materials. ChatGPT can also assist in developing playbooks, scripts, social media posts and email templates, as well as training materials that educate sales team members on marketplace trends and competitors’ offerings. “What I have found is AI is not going to replace humans in the sales process any time soon, but it can be a competitive differentiator that makes you and your team more valuable to yourselves, to your career and to your company,” says Cheriyan, who also urges sales executives to guard AI successes as proprietary corporate practices. “Creating effective prompts will be the next generation of intellectual property, so if you develop an AI practice that you plan to share with your team, make sure everyone is aware that this is now intellectual property not to be shared externally.”
Transforming HR With AI
Stephen Lytle, Principal Consultant, 813HR
AI can be a powerful ally in the HR function, enhancing efficiency and engagement with employees during recruitment and training on through to delivering ongoing development. In addition to assisting with the creation of job postings, interview guides and training materials, ChatGPT can also help HR professionals with annual reviews, performance feedback and goal-setting, says Stephen Lytle, a principal consultant with 813HR.
Lytle cites the example of a hypothetical healthcare company whose director of operations emails HR urgently looking to hire a manager for a new department. “I can give ChatGPT the information from that email along with the company’s profile and ask it to create a job description that keeps these insights in mind and aligns with the company’s values and goals,” explains Lytle. “It will instantly pop up a job description that provides a company overview and describes the responsibilities of the role.”
Next, an HR executive would review the output and follow up the initial request with feedback that will enable ChatGPT to fine-tune the posting. “Here’s where you go back and revise,” explains Lytle. “Maybe you need the posting to follow a company template or to focus more closely on specific KPIs, such as leading a team or delivering patient experience.”
Given the right level of detail and prompt, ChatGPT can do much more than write job descriptions and postings, including creating personalized onboarding materials for new hires, providing constructive feedback and goal-setting guidance and crafting internal communications and employee satisfaction surveys. “To get the most out of ChatGPT, you have to understand its capabilities,” Lytle cautions. “If you put bad inputs in, you’ll get bad outputs.”
These tips can help maximize ChatGPT’s HR potential:
Treat it like an intern. “The instructions you give an intern are going to be a little more detailed, a little more specific than you might give an experienced colleague,” he says. “And when you get that output back, you definitely need to check it. You wouldn’t trust what an intern gives you if it’s their first week on the job.”
Ask questions. AI users often make the mistake of instructing the technology on what to do and hoping for the best, says Lytle, who recommends a more interactive approach. “Ask ChatGPT, ‘What other information do you need for me to make a better response or the best response?’ And it will tell you the exact information you need.”
Don’t limit your options. “Whenever appropriate, ask for multiple options when making a request, rather than just one,” says Lytle. “Tell it to give you 10, then pick the three you like best and ask for 10 more like those. You have to prompt it and then continue working with it interactively to get what you want.”
Leverage social media. The speed at which AI is being embraced by companies of all sizes across industries is formidable—which makes staying current a challenge. “There are plenty of places, from Facebook to LinkedIn, where information about what’s being done and learned about these tools is being shared,” says Lytle. “Following people who post about this is a great way to stay informed as use of AI in HR evolves from early adoption to a mature integrated strategy.”
Finance Reimagined
Glenn Hopper, CFO, Eventus; Author, Deep Finance: Corporate Finance in the Information Age
Finance departments are already using AI to streamline processes and enhance efficiency, but that’s just the beginning of what it can offer, says Glenn Hopper, an author and CFO of Eventus Advisory Group, who sees vast potential for the technology to inform decision-making. “The way to overcome finance being called a cost center is to find ways to provide real value to the business,” he says, adding that large language models’ capabilities go beyond basic financial computations. “Unlike previous methods of machine learning and artificial intelligence, generative AI is actually generating new data—creating something new based on data it is trained on.”
Hopper recounts a recent experience checking those capabilities by feeding publicly available Microsoft Corporation financial reports and statements in PDF form to ChatPDF for a financial analysis. “I basically uploaded Microsoft’s 10k and five years of quarterly statements and then asked questions as if it were a person who had that information memorized,” he explains. “For example, you can ask, ‘What does the ND&A section of this document say?’ and it will virtually instantly provide a summary and also give you the page number so that you can easily find and verify it.” AI can also pull out highlights and create charts.
The real transformative power, however, comes when finance executives fine-tune a finance and accounting chatbot to enable it to predict future trends, market fluctuations and potential risks, helping finance teams mitigate risks and deliver value. Hopper was able to generate sophisticated financial reports analyzing the publicly available Microsoft reports. Python serves as the backbone of the process, fetching and processing data from various sources that can then be run through statistical models or machine learning algorithms that incorporate Excel formulas.
“One of the things generative AI is great at is analysis of data, so I asked it to calculate key financial ratios and data points, debt to equity, return on equity, profit margin over the years, in 30 minutes flat,” explains Hopper. “Another way I like to use this is looking for correlations that aren’t universally known or readily apparent—for example, something on the expense side that is positively correlated to future revenue or daily stock prices to annual revenue.”
While it’s critical for companies to start the journey toward understanding how to integrate AI in the finance function, Hopper cautions against sharing proprietary financial information with ChatGPT or other large-language models—at least for now.
“When you log into ChatGPT and upload information, it’s not as if you’re leaving information on a table in Times Square, but these models are not enterprise-secure,” he says. “Security features are being developed, but for now I advise everyone to stick with publicly available marketing and finance material.”
Harnessing the Power of AI in Marketing
Ashley Couto, AI Marketing Consultant
Ashley Couto, a digital marketing veteran, shared core principles and tactics that guide her approach to using ChatGPT, Midjourney, Headline Studio by CoSchedule and other high-impact AI tools to develop marketing campaigns.
Human-centric communication. Couto emphasizes the importance of approaching every interaction with AI as if communicating with a human. “I start by thinking, ‘If I was giving this task to a person on a marketing team, would they have enough information right now about how to do this successfully?’” she explains. “Or, ‘Does ChatGPT have enough information about my company, about the different audience personas that we target, about the specific project that we’re hiring them to work on and the tasks that lead from A to Z?’” People who expect an instant answer from AI aren’t realizing its full potential. Giving AI what it needs to truly deliver requires writing robust prompts and then participating in a back-and-forth conversation, explains Couto. “It’s not unusual for me to spend 45 to 60 minutes in ChatGPT working back and forth with it on something before it gives me an output that is 90 percent where I need it to be.”
AI reference points are inherently outdated. One of the challenges with using AI tools is that they are trained on available data accessible up to a certain point in time. Since marketing is a domain that evolves rapidly, with trends, platforms and algorithms changing frequently, AI models may lack information about recent trends and marketplace developments. This lack of awareness can lead to the AI providing information or recommendations based on outdated data. Marketers need to be conscious of each AI tool’s cutoff date and vigilant about making sure AI-provided information is not outdated. “Any marketer knows that things are constantly shifting, and you’re constantly having to be aware of what’s happening now, what is ongoing and what is happening next at a rate that is much faster than what these tools can learn,” Couto notes. “Marketing changes very quickly—a lot faster than some of the other domains and departments within a business.”
Without human help, AI will operate in generalities. While AI tools can enhance productivity, they are not a replacement for human knowledge, creativity and the ability to interpret nuanced situations. “AI operates to the lowest common denominator, so most of the time it will output standard, conventional advice,” says Couto. “It’s our job as marketers to make sure we’re taking the things we know, our skillsets and experiences, and directing the tool to operate with those things in mind.”
Be sure to provide context and clear direction. AI will need instructions in order to ensure that output is delivered in a voice, style and format appropriate for the brand, the audience and the message. “You’ll want to tell it about the company and about the consumer base or target audience,” says Couto, who adds that for those who struggle with getting prompts right, there’s a simple solution: Ask AI. “Let’s say I need to come up with a new Facebook ad campaign for a new property that my realty company is putting on the market. I can tell it, ‘Ask me as many questions as you need to write a comprehensive and high-converting Facebook ad.’”
Constant learning is a two-way street. Marketers who hope to use AI tools need to commit to evolving right alongside them. Couto and her team maintain an internal database logging what went well and what did not for every test they conduct with any AI tool. “We have weekly meetings where we talk about where tests flopped or succeeded, why those things happened and how we can move forward and try a different approach,” she says. “The key with tools like this is understanding that you will need to constantly play with them and test and trial their limits in order to get from point A to point B.”
The post Special Report: Practical AI for Everyone appeared first on ChiefExecutive.net.