Whimsical is a full-featured visual collaboration workspace that enables users to create many types of diagrams, including flowcharts, wireframes, projects and mind maps. But its mind mapping tool deserves special attention, because it recently added OpenAI integration.
This new capability makes it easy to add AI input to your mind maps. And it has a capability I’ve never seen before in AI-enhanced mind maps.
Here’s how it works:
First, you build out a mind map as you normally would, with an array of topics and subtopics. Whimsical’s simple, intuitive interface makes this process easy and fun.
When you need input from OpenAI, you simply write a prompt in a map topic and click on a “generate ideas” button in the context-sensitive floating toolbar that appears above it.
Whimsical parses its reply into a series of map subtopics, one per idea.
Simple and quick – easier than visiting the ChatGPT website or using one of a growing number of paid AI tools to write prompts, get ideas and then copy and paste them into your mind map.
Test driving Whimsical’s AI mind mapping tool
I experimented with Whimsical’s mind map AI engine by giving it this prompt:
“How can I teach people the value of mind maps?”
It returned 5 answers, all fairly thoughtful. I wanted more ideas. So I made sure my prompt was still selected and I clicked the “generate ideas” button again. OpenAI immediately added another 5 ideas to my mind map, below the existing ones.
Follow-up prompts are supported
One of the most effective ways to use OpenAI and tools based on it is to give it follow-up prompts – adding specific parameters to it or asking it to give you better or more nuanced answers. This isn’t a capability I’ve seen in AI-enhanced mind maps before, so I thought I’d give it a try.
I first asked Whimsical’s AI tool how I can teach people the value of mind maps. So I created a new subtopic below the answers that it gave me and entered this follow-up prompt:
“Give me more answers. But assume that the people I want to teach are the leaders of non-profit organizations.”
I was intentionally vague in the way I worded it. I didn’t mention the concepts of teaching or mind mapping. In order to answer this prompt accurately, OpenAI would need to refer my first query, which I conducted 12 hours later.
To my amazement, it did an excellent job answering my new prompt, still focused on mind map training but now drilling down into the ways in which I could train non-profit leaders on their us4. Very cool!
One final experiment
With ChatGPT and other AI writing tools, you can formulate prompts that specify a specific number of answers – for example, “Give me 10 answers for X” or “Suggest 20 ideas to improve Y.” I wanted to see if Whimsical’s AI tool could do the same. So I gave it a prompt that was aligned with my earlier queries about mind map training for the leaders of non-profit organizations:
“What are 10 things that non-profit leaders need help with, where mind mapping could help them?”
Once again, it returned only 5 answers. Why? I suspect that it’s a built-in throttling of the tool to limit the number of answers it can request from OpenAI. It may also be designed this way so you don’t bury your mind map in large quantities of AI-generated results.
Enhancing AI-generated topics
Whimsical treats AI-generated topics the same as those that were “organically” generated. You can add subtopics to them, outlining your reaction to specific ideas or riffing off of them with your own ideas. You can also enhance them with:
- Topic colors
- Topic shapes
- Attached files
These embellishments enable you to visually classify the answers the AI engine added to your mind maps. For example, you could color the topics green to represent ideas you plan to take action on. Unfortunately, Whimsical’s icon collection doesn’t contain letters or numbers (so I can rate ideas 1/2/3 or A/B/C). But that’s a minor complaint.
I’m pleased with what I’ve seen so far. I predict this tool will only get better as Whimsical moves its AI integration to ChatGPT4, just launched last week.