Meet Your New Advisors: How AI-Language Learning Models Shape Expert Personas

Everyone is talking about how to prompt AI for better content creation. At least that’s how it seems when visiting networking platforms like LinkedIn. You can’t scroll two clicks without landing on several how-to guides on the topic.

Prompt engineers impart wisdom about how to get AI language learning models to do your bidding. Most of the advice centers around AI content creation and how to use the technology to replace professional copywriters and content creators on your marketing team.

It’s possible to use AI like ChatGPT, Google Bard, and Claude to create content with precise directions from a human. However, that’s not necessarily the best way to leverage the power of AI technology.

Michael Goldrich offers a more effective way to use AI to achieve your marketing goals. The founder and chief advisor at Vivander Advisors LLC, Goldrich relies on a 5-step process to guide companies through an AI transformation.

His new book, Too Many Hats, Too Little Time explores Generative AI and its transformative impact on the business world. Innovators and leaders can learn how to integrate AI while preserving human judgment and creativity.

Thinking outside the AI content creation box

“I see a lot of stuff online where people are selling tools for specific business problems, but I think it’s backward thinking,” Goldrich said. “You need to look at the entire company and see how people work before offering them specific solutions.”

So, what exactly does that mean?

When you understand the business’s overall challenges, you can provide more robust solutions. Instead of prompting AI to create content for your brand, Goldrich suggests something more groundbreaking: AI personas.

An AI persona is a character or identity created for an artificial intelligence system to interact with users. Creators give the AI persona a specific personality, tone, and style of communication to mimic a human-like quality.

Think of the last time you interacted with a chatbot or virtual assistant on the website of your favorite brand. They were created with an AI persona tooled toward improving the customer experience.

What Goldrich proposes cranks AI personas up a notch to serve a more innovative purpose.

Last year, he spent considerable time experimenting with ChatGPT 4. He first learned the art of creating effective AI prompts. “One of the first principles of creating a good prompt is to create a role. For example, you are a digital marketer, or you are a social media manager. After the role is created, you write the prompt to provide enough context that is associated with the role to deliver the output you want. Writing good prompts is the key to getting the AI to do what you want it to do,” he said. “Once I had that down, I started getting specific with it and asking the role to persist through the chat.”

Asking ChatGPT to persist means it maintains continuity and context throughout the conversation. Requesting persistence creates a more coherent and meaningful interaction between a human and AI.

Goldrich then used what he’d learned to create two separate AI personas and directed them to interact through ChatGPT. Each persona had unique characteristics such as a profession and a tone of voice. After much experimentation, he succeeded in using the AI personas to solve a specific conflict by interacting with one another based on his prompts.

He continued to experiment with creating AI personas until he had an entire “board” of them, each with an area of expertise beneficial to his business marketing goals. The personas could converse with one another and with him in a similar way you would expect to see a board of directors operate, Goldrich said.

A computer-generated human head shows swirls of light abote it to indicate it's thinking. It represents an AI persona.

5-step plan for AI persona success

He was so stunned by the results he decided to share them.

In his book, Goldrich outlines the steps he took to tailor personas to align with company goals and have them interact in meaningful ways. He includes screen captures of real, unedited conversations he had with his team of AI personas to demonstrate how the process works. “A quarter of the book is these back-and-forth conversations,” he said.

Goldrich expresses in his chapter on strategies for engaging with AI personas that how you choose to engage with the digital entities significantly impacts the quality of insights they provide.

He provides basic and advanced techniques for prompting and AI persona creation so you can get the results you need. Goldrich likens AI language models like ChatGPT 4 to a remote employee on the first day of work. “You can get them to do what you want, but you have to be clear and concise on what you want,” he said.

One of the challenges of prompting with AI is drift, Goldrich warned. “You can’t use the same prompt over and over because you won’t get the same response each time.” You must give it adequate context. For instance, you can’t just say, “be a copywriter” and expect it to understand your input and produce quality output. 

“One of the reasons people get so frustrated when prompting AI is because Google has trained us to type a series of keywords into a search bar,” said Goldrich. “ChatGPT is the exact opposite. It’s not 3 or 4 words but three or four comprehensive sentences. The more thoughtful you are, the better the output is going to be.”

His step-by-step guide makes understanding the purpose of AI personas and creating ones to suit your needs easy. Whether you’re seeking specialized advice from a legal AI persona or want a panel of AI collaborators, following his instructions makes it possible.

Goldrich’s plan is ideal for startups and small to medium-sized companies that lack the resources to hire expert advisors. “You can have a sandbox environment of advisors to help you identify the ‘gotchas’ with ideas for growth or other business operations,” he said.

His book includes a cautionary note about using AI personas. “It’s important to remember these AI personas draw upon training data to respond in a way that’s consistent with an expert’s public persona,” said Goldrich. “It’s emulation, not impersonation. They can’t capture the full scope of that person’s knowledge, experience, or consciousness.”

AI’s take on AI personas

Perhaps the most entertaining part of Too Many Hats, Too Little Time is the conclusion. It features book reviews from each of the AI personas he created.

“Some of them liked it, some of them didn’t,” Goldrich said, laughing. “The lawyer (persona) was concerned about privacy, while some of the others were concerned about losing the human in the loop.”

You can find Goldrich’s book on Amazon.

Interested in learning more about how his company helps organizations transform with the power of AI? Feel free to connect with him on LinkedIn or send him an inquiry via email.

1 thought on “Meet Your New Advisors: How AI-Language Learning Models Shape Expert Personas”

  1. Great site with quality based content. You’ve done a remarkable job in discussing. I look forward to seeing more of your great posts.

    Reply

Leave a Comment