Adam Honig is the CEO and cofounder of Spiro.AI and is passionate about helping companies grow using artificial intelligence.
getty
With the boom of OpenAI's ChatGPT, the world has seen a huge uptick in demand for other types of generative AI: Stable Diffusion, DALL-E, Midjourney and others. From advanced chatbots and writing to image creation and 3D modeling, using generative AI models as a shortcut has become incredibly tempting.
But here's the thing: Generative AI is tricky. As we saw with the lawyer who ended up citing fake cases when using generative AI, you have good reason to be wary of the tech.
First of all, AI has to be "trained." Basically, you can think of it as teaching a child. Depending on what information you give that child, it'll think and learn based on that data. And since ChatGPT was only trained on information prior to September 2021, it's a mistake to rely on it for current or updated information.
Second, it needs a lot of information to provide meaningful (and accurate) content. ChatGPT and other models have been trained on petabytes of information. (You can think of a petabyte as the same thing as 20 million tall filing cabinets or 500 billion pages of printed text.) But it still has trouble with thinking “outside the box.” Just look at some of the image creation AIs handle hands.
But generative AI is really good at some things. All that data it was trained on, regardless of its topic, means it learned a lot about how we communicate. So if you want to write a thank-you note for a business meeting, or summarize a call and action items, it’s really good at saving time. It can also look at a lot of communications to customers across multiple people and summarize what is happening with each customer.
Simply put, AI excels at taking large amounts of information, finding patterns within it and then giving what it finds to a human for action.
These reasons are why you shouldn’t blindly let AI write for you, though. Yes, many products (including my company's) can draft an email based on your prompts. Tell AI you want an email that talks about your great conversation with your customer. Explain that you want to follow up in two days. AI can draft something for you to look at.
And that's the key: You should never finalize with AI or use it to send anything on your behalf. You should never, for example, give complete control over to AI to manage the human relationship with your prospects and customers. But using generative AI for what it's good at—drafting responses, creating summaries and finding patterns—is that can be incredibly beneficial if you take advantage of it with the correct human oversight.
After all, who wants to draft the same emails day in and day out? Your sales team certainly doesn’t, nor do they have the time to summarize every customer conversation for internal and external use. That’s where the power of generative AI can make your sales team more productive—provided that a human reviews and edits before hitting send.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
