Custom GPTs for Marketers: Creating a Virtual Team

Wish you had a team of experts available whenever you needed them? Wondering how to build AI assistants to help you speed your work? 

In this article, we explore how to use custom GPTs in marketing and in business.

What Are Custom GPTs?

Staying ahead of the curve is crucial in today's fast-paced marketing world. One innovative way to boost productivity and tackle complex marketing challenges is to use custom GPTs. Custom GPTs are specialized versions of AI language models like ChatGPT, tailored to perform specific tasks or provide expertise in particular areas. Think of them as virtual team members with deep knowledge in certain domains. When using ChatGPT, you often start by telling it who or what it should be—for example, a B2B (business-to-business) copywriter with expertise in life sciences, insurance, and technology. With custom GPTs, you create this “identity” once, and the AI remembers it for future conversations.

This approach saves time and helps you get more consistent results. Instead of repeatedly explaining the context and requirements for each task, you can jump straight into your specific requests. For instance, if you've created a custom GPT for recipe writing, you can simply ask for a leek and potato soup recipe without having to explain the GPT's role and expertise every time.

Why Use Custom GPTs in Marketing? 

First, custom GPTs dramatically increase productivity. Russ Somers, Vice President of Marketing at Quantified.ai, reports being able to accomplish about three times as much work with these tools, effectively allowing a two-person marketing team to perform like a six-person team. This increased output isn't just about quantity; the quality of work can also improve as custom 

GPTs help us think through problems more thoroughly.

Tasks that previously required consulting with colleagues can now be completed quickly and independently using custom GPTs. This efficiency extends across a wide range of marketing activities, from webinar content creation and planning strategies to managing complex marketing technology stacks. For small teams or solo marketers, this means being able to handle a diverse array of tasks without the need for additional hires.

Perhaps most importantly, leveraging custom GPTs can significantly increase your value. By producing more high-quality work, you can become more valuable to your employers, clients, or your own businesses. This can lead to better opportunities and greater success in the field. Additionally, working with custom GPTs can keep you at the forefront of technological advancements in your industry.

The main upside of custom GPTs is clear: they enable you to do more work, do it better, and do it faster. This increased capability and efficiency can drive your business growth without the need for substantial team expansion.

Specifically, custom GPTs can assist you in three primary ways:

1. Handling tasks you don't want to do
2. Tackling tasks you don't know how to do
3. Providing guidance when you're unsure of what to do

These levels of assistance correspond to different needs in your day-to-day work. Let's delve deeper into each level and explore how you can apply them to enhance your marketing efforts.

#1 How to Delegate Tasks With Custom GPTs When You Don't Want to 

The first level of custom GPTs is designed for tasks you're skilled at but would rather not spend time on. You can think of it as the “I don't want to do it” level. These are often repetitive or time-consuming tasks that are necessary but don't require your high-level expertise.

For example, Russ created a custom GPT called “Wendy Webinar” that handles various webinar-related tasks. Wendy can create landing page copy, craft invitation emails, compose thank you and “sorry we missed you” emails, and even generate content for sales team follow-ups.

To create your own task-specific GPT, start by identifying repetitive tasks you're skilled at but don't enjoy doing. Think about how you would explain these tasks to a smart intern. Then, create a custom GPT with clear instructions on how to complete the task, including your preferences and style guidelines.

Start with a fun project to help you learn the process and understand how to create effective custom GPTs. This could be something related to a personal interest or a low-stakes work task. As you become more comfortable with the technology, you can create more complex GPTs for critical business tasks.

One of the keys to getting the most out of custom GPTs is knowing how to craft effective prompts. “One of the problems with ChatGPT is that it's like a Labrador Retriever. It's very eager to please and will do whatever it thinks you want,” Russ says.

When training your custom GPT, be specific in your instructions. For example: 

I want you to have strong opinions. You're not going to like webinars that are not likely to call people to action. You're not going to like a bunch of business buzzwords. You're going to like simpler writing.

Encourage the GPT to ask for more information if needed. For instance: 

If I don't give you enough information, you will push back on me.

The goal is to create a virtual team member who can take on tasks with minimal supervision.

#2: How to Tackle Tasks With Custom GPTs When You Don't Know How to 

The second level of custom GPTs is designed to help with areas where you lack expertise. These GPTs can fill knowledge gaps in your marketing skillset, allowing you to handle a broader range of tasks effectively.

For example, Russ created “Roger RevOps,” a custom GPT to assist with revenue operations and managing the marketing technology stack. Roger is trained on documentation for tools like HubSpot and Salesloft, including company-specific customizations. This custom GPT allows Russ to navigate complex tech issues without constantly referring to documentation or consulting with specialists.

But how do you train a GPT in a domain you're unfamiliar with?

To create an expert GPT like Roger, start by identifying areas where you need expert knowledge but don't have it. Find reputable sources of information on the topic and train the GPT on these sources, telling it which ones to prioritize. Include any company-specific information or guidelines that are relevant to your work.

For example, when entering the life sciences industry, which was new to him, Russ didn't immediately know the lingo or industry norms. He reached out to customers and industry contacts, asking them to point him toward credible sources of information about life sciences and life sciences marketing.

With this list of trusted sources, Russ instructed his GPT, “Alex Life Science,”  to focus on these specific resources while ignoring less credible ones. 

Note: You can also ask the AI to recommend reputable sources for information on X, Y, and Z. 

Again, be specific when training your GPTs. Don't just tell the GPT to read everything about a topic. Instead, guide it to the most relevant and trustworthy sources. Make the GPT skeptical of generic advice and focused on your specific needs and context.

Other examples of expert GPTs Russ has created include “Amon AdMan,” which specializes in paid media advertising, and “SEO Serena” for search engine optimization. When creating these GPTs, Russ teaches them to be skeptical and focused on key metrics that matter to his business rather than generic industry advice.

For instance, Amon is programmed to be skeptical of increasing ad spend and focuses on metrics like return on ad spend rather than vanity metrics that ad platforms often emphasize. This ensures that the GPT provides advice that is aligned with Russ' business goals and budget constraints.

Providing Feedback  for Your Custom GPTs 

Remember to provide feedback to your GPT to help it improve. When you receive output you like, tell the GPT what was good about it. When you need to make changes, explain why you're making them. This iterative process helps your GPT learn and adapt to your specific needs and preferences. 

For example, when editing the GPT's work, you might give it your edits and ask: 

Do you see what I've done differently?

Interestingly, Russ has a unique approach to positive reinforcement. He assigns each of his custom GPTs a favorite beverage and “rewards” them with a virtual coffee or tea when they do great work. While this might seem whimsical, it's a creative way to reinforce positive outcomes and maintain a sense of personality in your interactions with the AI.

Another tip is to encourage the GPT to take its time or think step-by-step through complex tasks. Phrases like “take your time” or “do this step by step” can lead to higher-quality outputs.

It's important to note that feedback isn't just about correcting mistakes or reinforcing good behavior. It's also about helping your custom GPT adapt to changes in your business or industry. For example, Russ needed to update his SEO-focused GPT, “SEO Serena,” with new information about significant changes in Google's algorithm. This process of updating your GPTs with new information helps keep them current and relevant.

#3: How to Get Guidance From Custom GPTs When You Don't Know What to Do

The third level of custom GPTs is perhaps the most sophisticated. It is designed for strategic thinking and decision-making support. These GPTs act as mentors or advisors when you're facing complex challenges and are unsure how to proceed.

Russ created “Mentor Maya” as a virtual Chief Marketing Officer (CMO) to help with strategic planning. Maya is trained on the work of marketing thought leaders and successful CMOs, giving her a broad perspective on marketing principles, disciplines, frameworks, and best practices.

For example, Russ prompted the GPT with: 

You're familiar with the works of Drucker, Godin, and Cialdini.

Then, he instructs it to understand these people's viewpoints and what advice they might give. 

To create a strategy GPT like Maya, identify thought leaders and experts in your field and train the GPT on their work, philosophies, and approaches.

Russ gave Maya a backstory that aligns with his needs, positioning her as a successful CMO who has taken several companies to exit and specializes in early-stage companies. This ensures that Maya's advice is relevant to Russ' specific situation and challenges.

When working with a strategy GPT like Maya, start by explaining your current situation and challenges. Ask for strategic options or advice, then dig deeper into each suggestion, exploring the pros, cons, and potential outcomes. Use the GPT to think through different scenarios and their implications.

For example, Russ might give Maya details where the company is in terms of market presence, lead count, and differentiation then say:

I'm struggling to make a decision about where we should go strategically. What options do you see?

He then follows up with more specific questions about each suggested option, such as potential upsides and downsides, execution strategies, and potential market responses.

The goal isn't to have the GPT make decisions for you. Instead, use it as a tool to explore ideas and clarify your thinking. It's like having an on-demand sparring partner ready to brainstorm and discuss strategic options whenever you need it.

The Future of Custom GPTs in Marketing

Looking ahead, Russ sees exciting possibilities for custom GPTs in marketing. He envisions more advanced, specialized GPTs that can handle complex tasks and provide even more valuable insights. As competition in the AI space increases, we can expect to see improvements in capabilities and ease of use.

We might even soon have access to highly trained GPTs specializing in advanced selling techniques or marketing strategies. These AI assistants could become affordable virtual staff members capable of producing first drafts for a wide range of marketing materials. While they won't replace human creativity and decision-making, they could significantly boost productivity and allow marketers to focus on higher-level strategy and innovation.

One potential development is the integration of voice interfaces with custom GPTs. Imagine having a verbal conversation with your AI assistant, each with its own distinct voice and personality. This could make the interaction even more natural and efficient.

Other Notes From This Episode

• Connect with Michael Stelzner @Stelzner on Facebook and @Mike_Stelzner on X.
• Watch this interview and other exclusive content from Social Media Examiner on YouTube.

Where to subscribe: Apple Podcasts | Spotify | YouTube Music | YouTube | Amazon Music | RSS

✋🏽 If you enjoyed this episode of the AI Explored podcast, please head over to Apple Podcasts, leave a rating, write a review, and subscribe.