The Problem

Your need for content continues to increase but the time to write it does not. People are talking a lot about this “GPT3” thing but it isn’t clear how a nonprofit might actually use this day-to-day. With purpose-built writing/image tools there are many practical uses:

  • Thoughtful email marketing messages are needed regularly.
  • Senior staff doesn’t have the time to write thought leadership articles.
  • Clever social media posts and replies are needed weekly.
  • Text and image ad iterations.
  • Grant application writing buries the staff.
  • Creative images are needed for storytelling to donors.

The Solution

Whole Whale uses your data and writing voice to build easy-to-use AI writing applications and then trains your team on how to use them. These tools will help your team create organizational content 5X faster with purpose-built AI applications that KNOW your voice.

Example: here is a generally trained nonprofit email marketing writer.

IE: “Write a fundraising email for Power Poetry, a safe, creative, free online platform for teens to create poetry. This year Power Poetry has supported over 1.1 million poets. This email is asking for donations around GivingTuesday.”

Pro-tip: You can copy and paste the result to generate additional parts of the email. More about this AI email writer.

Describe the org/email you want
GPT3 Response

This tool can be trained on your organization’s voice to create quality first drafts for your team. Once these AI applications are built the next step is training, where Whole Whale will work to show and fine-tune how the tool gets used in the organization.

Whole Whale cares deeply about how the social impact adopts this new powerful tech, which is why we created a simple, free AI course.


Whole Whale is ready to work with great organizations interested in seeing what is possible with AI-generated content and custom-build useful AI writing tools.


What about other open writers?

Tools like,,, and others built on are powerful but run into the convergence problem. Once more companies start using the same tool with the same underlying training data the outputs will start too all look very similar. IE if 10 environmental organizations ask the same thing of the same bot, the outcomes will look like copies. This issue was well described by Chamath Palihapitiya on the popular All-in Podcast.

Is this just stuff copied from the internet and duplicate content?

Nope. It is a neural net trained on the entire internet up to 2019, this category of Natural Language Processor – GPT3 (General Pre-Trained Processor) is creating each word based on probability settings. 

Example of a response showing these probabilities. 

Wait, does Google penalize AI writers?

They say it technically violates a dated policy, but they  have no way to detect this when used properly as a writing assistant. Our stance is that if the tool is used as a writing assistant to create quick drafts and refine the writing process it will be considered acceptable use in the same way that automated writing already works in Google autocomplete…

QA with Google Devs  April, 2022: Google Says AI Generated Content Is Against Guidelines