Are you using ChatGPT to complete the same tasks over and over again? Do you find yourself copy-pasting the same prompts repeatedly? Custom GPTs might be the solution you’ve been looking for.

This is a bold statement – “Reclaim your time” – and I’m not here for empty promises. Custom GPTs won’t reclaim all your time, but they can certainly help streamline repetitive tasks you’re already using AI for.

Why You Should Consider Custom GPTs

If you’re using ChatGPT to generate social media posts, monthly reports, or any content you create on a regular basis, Custom GPTs can make this process so much easier. Here’s why they’re worth exploring:

• Save valuable time – No more retyping or copy-pasting the same prompts over and over
• Train AI to write in your unique voice – Avoid that cheesy AI language (what I call “AI banality”)
• Maintain consistent tone – Keep your content on-brand across all outputs
• Improve precision through ongoing training – The more you use it, the better it gets

You’ll still need to edit, fact-check, and proofread the output, of course, but it’s going to save you time compared to creating content from scratch or repeatedly training generic AI.

What Are Custom GPTs?

Custom GPTs are specialised versions of ChatGPT trained to perform specific tasks according to your instructions. Think of them as personal assistants trained to handle particular jobs for you.

For example, I’ve created several Custom GPTs that save me significant time:

• AI Publicist – I use this weekly for my segment on Ticker. I read news stories, paste them into my AI publicist, and it quickly pulls out specific talking points for me to prepare for interviews. Instead of manually extracting key points from lengthy articles, it gives me what I need in seconds.
• Key Message Generator – Based on my book “Strategic Social Media Management Theory and Practice,” this GPT creates key messages for social media strategies. I can either paste a website link for it to analyse, or it will prompt me for specific information to create structured messages.

How to Build Your Own Custom GPT: A 3-Step Process

Creating a Custom GPT might sound complicated, but it’s actually quite straightforward. Let me walk you through the three key steps:

Step 1: Plan and Prepare

Before diving in, you need to:

• Decide on a clear purpose – What specific task do you want your GPT to handle?
• Gather examples of your work – The Custom GPT will base its results on the examples you provide
• Define objectives – Structure, tone, format, and other specifications

Think of it like training a new staff member. You need to give clear, specific instructions and examples of what you want the output to look like.

Remember, AI is there to help with the task, not completely replace it. You’ll still need to:

• Prepare proper instructions and examples
• Review and refine the output
• Do the final fact-checking and editing to add your personal touch

Step 2: Build and Train

Now let’s get into the actual creation process:

1. Go to ChatGPT and navigate to the “GPTs” section
2. Click “Create” to start a new Custom GPT
3. Enter a name for your GPT (like “Dr. Karen’s LinkedIn Post Creator”)
4. Add an image (upload your own or use DALL-E to generate one)
5. Write a brief description of what your GPT does

The most important part is providing detailed instructions. Be as specific as possible about:

• Role and purpose – What is this GPT an expert in? What is it tasked with doing?
• Tone and style – How should it sound? Professional, friendly, academic?
• Structure – What components should be included in the output?
• Formatting details – Paragraph length, use of bullet points, etc.
• Training data – Upload examples of your best work (I recommend screenshots of your best-performing content)
• Example topics – Give it a few sample scenarios to work with
• Quality check instructions – Tell it what to look for when reviewing its own work

You can then add conversation starters – these are prompts that will appear when someone first uses your GPT.

Finally, decide on accessibility:

• Public (available in the GPT store for anyone to use)
• Link-sharing (only accessible to those with the link)
• Private (only for your personal use)

Step 3: Test and Tweak

Once your GPT is created:

1. Test it out with some sample prompts
2. Review the output and see how closely it matches your expectations
3. Go back and refine the instructions if needed
4. Keep updating the training data regularly to improve performance

It’s crucial to use your Custom GPT regularly and continue refining it. As you saw with my Key Message Generator example, if you don’t use it for a while, it might “forget” some of your specific instructions.

Creative Ways to Use Custom GPTs

There are countless applications for Custom GPTs depending on your needs:

• Agency work – Create a separate GPT for each client, trained on their specific brand voice
• Audience personas – Build GPTs that analyse or write from the perspective of different customer segments
• Platform-specific content creation – Develop GPTs specialised in creating content for particular social platforms
• Standardised reports – Like the one I built for a local council to help generate strategic plans for grant recipients

Final Thoughts

Custom GPTs aren’t as difficult to create as you might think. It’s really about providing clear instructions and good training data. While they require some maintenance and regular updates, they can save you significant time in the long run.

If you’re interested in learning more about using AI strategically in your communication, I’d love to connect:

• Connect with me on LinkedIn
• Book a free 30-minute chat to discuss any AI-related questions
• Join my free Facebook community focused on AI for strategic communication, where I share weekly news updates and offer discounts on monthly workshops
• Check out my new book: Artificial Intelligence for Strategic Communication

Remember, the key to success with Custom GPTs is choosing a repetitive task that would benefit from automation, providing clear instructions, and regularly updating your training data. Give it a try – you might be surprised at how much time it saves you.

What CustomGPTs have you built or plan to build in the future? Share them in the comments below.