- The AI Exchange
- Posts
- Don't pivot to AI, solve problems instead
Don't pivot to AI, solve problems instead
Recommended steps for getting started with AI, deep dive on a consulting and agency use case
Welcome to the very first edition of what we’re determined to make the best damn newsletter in AI. Here we’ll break down AI topics that matter, open your mind to use cases, and keep you ahead of the curve.
Our #1 goal is to be useful. So please shoot us an email 📩 if you have questions or feedback, and especially if you implement something we share!
Here's what we're covering today:
A mental model for how to start using AI + free "ramp up" guide
Interesting reads on prompt engineering
Step by step on how consultants and agencies can use GPT to build Scopes of Work
... and if someone forwarded this email to you, thank them 😉, and subscribe here!
Let’s get to it! 👇
TODAY'S PERSPECTIVE
How the heck to start using AI in 2023 🚀
AI is moving fast, but don’t feel like that means you should pivot your entire business or career overnight. In fact, the biggest opportunities for you likely lie in what you are already doing.
Use AI to solve an internal or personal problem first. You’ll get more familiar with its capabilities to tackle bigger and more ambitious ideas. Are you thinking of launching a feature that writes emails for your customers? Build a version that writes emails for yourself first.
Take an inventory of how you or your business spends your time. Pick the biggest opportunities. Break up the steps or problem and begin applying AI to make those tasks more efficient.
If you’re looking to start a new business around AI, focus on customers and their problems - not the technology. Look for problems previously too expensive or too difficult to solve. Then see if current or emerging AI tech could unlock a solution.
And in case you didn’t get it yet, grab our FREE 10 page “Ramp up to AI” guide we shared on TikTok last week.
USE CASE DEEP DIVE
Customized Scopes of Work for Consultants, Agencies and Freelancers 📝
Statements and scopes of work are common documents in consulting and agency businesses. They can be tedious to create, and often need to be personalized to each client or project.
Creating Scopes of Work with ChatGPT
Many people have figured out that you can ask ChatGPT to write almost any document for you. The trick to getting a good result is to be specific. Here’s a template we tested out to get you started:
Can you create a Scope of Work for this new client? We are [designing a strategy] to [integrate AI into the product strategy for a product] called [Foobar]. This will include [step 1], [step 2], [step 3], and [step 4] [+ additional details like revisions]. Write in a professional, concise tone.
Customizing Scope of Work creation with GPT Playground
To take it one step further, you can use shots to train the AI to build a Scope of Work matching your style. (No idea what we’re talking about, here’s a brief walk through).
First, give GPT an example input. It's useful to think of this as a prompt template that you will swap out details for in step 2.
Then, show GPT an example output. We like to use the input => output => notation for readability, but it's not required.
beta.openai.com/playground
Then, add the input you want to generate for. And click submit. This sends all of the text you've shared (the prompt) to the API to generate the output.
Make sure you've increased the maximum length of the prompt so that you have a large enough window to send and receive the text. If this gets too long, you can consider breaking up the prompts into sections.
beta.openai.com/playground
The result will come back in green!
You can even go expert mode
If you’re creating SOWs often, or want to run a business that does this for agencies and consultants, you can build a no-code app or a set of automations that does this.
Reply if you’d be interested in a tutorial ✨
LINKS
For your reading list 📚
Notable releases
Open AI launches Point-E that generates 3D models, although even its researchers say it has a ways to go before it's in its prime time for use cases
Quora launched Poe (beta with waitlist) which is a ChatGPT-like interface, although it's not clear what data it's been trained on
Prompt engineering
A data scientist demonstrates how easy it is to reverse engineer the prompts used by major AI products, starting with Notion AI
Related to the last link, Riley Goodside, a prominent data scientist known for sharing GPT-3 prompts on Twitter, discusses how prompts are not a competitive edge
And if you're really nerdy ...
Jon Stokes on the differences between Stable Diffusion 2.0 & 2.1, as part of his “AI Wars” series
Open AI’s walk through of the ChatGPT architecture
That's all!
We'll see you again on Thursday. Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected].
... and if someone forwarded this email to you, thank them 😉, and subscribe here!
Cheers,
🪄 The AI Exchange Team