The No-Code AI Toolkit
Here are 2 tools and 5 use Cases to power your existing business workflows. Most AI enthusiasts fail to build useful AI automations because of 3 reasons:
- They aren’t using the fastest and simplest no-code solutions – many of which are free
- They’re overwhelmed by the noise (1,000 prompt ideas!)
- They’re not applying AI to the right types of challenges – because they’re still in the old-school “If A, then B” automation mindset
Tool 1: Google Sheets
Spreadsheets don’t exactly have a reputation for being cutting-edge and exciting – but they are incredibly useful and powerful. So we’re going to use them as the basis for a lot of the business workflow automations we demonstrate today.
For our examples, we’ll use Google Sheets, since everybody can get free access to it (and it’s very powerful and full-featured). But if you want to level up, you can connect these same concepts to tools like Airtable instead.
Resources:
Google Sheets Alternatives
Airtable alternatives
The spreadsheet is where you’ll keep your input and output. Start with a super-simple sheet, like this:
Now, you can hook this up to Zapier and start applying AI magic to your data.
Tool 2: Zapier
You can start with a free Zapier account, but you may eventually need to pay for more tasks and/or faster processing times. Many people also like Make.com as an alternative to Zapier.
Resources: Zapier alternatives
Create a new Zap (that’s what Zapier calls an automation) and hook it up to your Google spreadsheet. Use the “New or Updated Spreadsheet Row” trigger.
Zapier will now “watch” your Google Sheet for any changes. Then, when a row changes, you can send the contents of one or more columns to the next step (e.g. ChatGPT).
Use Case 1: Translation
Now, let’s look at a (relatively) simple use case: translating content from one language to another. This is a common request that I receive from clients as a web developer, and while you can go out and hire a professional translator, we’re going to let ChatGPT do the work here.
🤓 Pro Tip: GPT-4 is better at translation than Google Translate
For this use case, the first column in our sheet will be our English text. Then, we’ll have Zapier and GPT-4 translate it to both Spanish and French, instantly. Notice that, even for this simple use case, this is way faster than sitting there with a ChatGPT window open and copying and pasting prompts.
💰 You’ll need to pay for ChatGPT to get an API key to hook up to Zapier
We’ll add a new Action in Zapier to send ChatGPT a prompt whenever a user updates a row in the spreadsheet. In this case, we will send the value of Column A (English) as part of the following prompt. Zapier takes care of pulling the dynamic data out of Column A and sending it to ChatGPT.
Prompt: “Translate to Spanish: [Column A Value]”
Here’s what this looks like inside your Zap:
In this example, Zapier knows my column name is “English” because that’s the header (first row) of the column in the spreadsheet. And I typed “Submit” into my first row of data, which is why you see that in gray within the Zapier screen above. Behind the scenes, that’s “Column A.”
Now, let’s get the response from ChatGPT and insert it into Column B. To do this, we add a new action called “Update Spreadsheet Row.” We pick the same Google Sheet that we used for the Trigger (the first step in our Zap), and then we dynamically bring in values from our previous steps into this step.
There are two dynamic values here: The Row ID (which is the ID of the row that was updated in the Trigger that started this process) and the Reply from ChatGPT, which in this case is the translated text in Spanish.
🤓 Pro Tip: ChatGPT sends back a lot of superfluous data, like how many tokens your request used. 99% of the time you just need to find the “Reply” value.
Voilà! Your sheet is updated with the Spanish translation. Now, I’ll do the same thing for French (create a second Action just like the one we just made for Spanish) and then publish the Zap so that it listens in real time for spreadsheet changes.
Here’s what the result looks like in Zapier:
Now, let’s put it into action. I can hand this spreadsheet to a team member who’s creating content for my website, and they can drop in English text (words, sentences, paragraphs, etc.) and instantly get their translations. (By the way, this is something my web design firm has been paid thousands of dollars to do in the pre-AI era. AI workflow automation makes it no big deal.)
Here’s the result:
Use Case 2: Proposal Creation
OK, let’s get more complex. Another common repetitive task for my web design firm is creating proposals for client projects. They’re about 80% the same every time, but I also need to customize each one – for example, with a personalized cover letter or requirements list. We can use the same input-output approach (with AI in between) to instantly create these proposal sections based on simple information about each project.
↗️ Resources: I use Qwilr for my proposals, and you can even integrate that directly into Zapier so that it creates a new proposal and inserts your AI results as variables
Here’s our example sheet. Our goal will be to generate a project summary with ChatGPT based on the easy-to-input data you add to each column.
🤓 Pro Tip: When you have lots of columns, you may want to have Zapier trigger the workflow when a “Status” column changes. That way, you’re not running your workflow every time anything in the row changes. Instead, you could have it only run when “Status” changes to “Ready for Automation.”
We’ll then set up a new Zap (see details in the Translation use case if you skipped over that section) and send the values of each column to ChatGPT in a dynamic prompt.
Prompt: Write a three-paragraph cover letter addressed to [Contact Person]. Express your excitement for starting the new website project for [Client Company] and mention your firm’s experience in the [Company Type] sector. Then, summarize the project, which include design and development of a new website on the [CMS] content management system, with [Number of Templates] unique page templates and the following extra features: [Extra Features]
Here’s our starting data:
🤓 Pro Tip: I duplicated my “Content Translator” Zap and then made a few tweaks to turn it into a “Proposal Writer,” saving me the time of connecting all my triggers and actions again.
And here is one of the cover letters that our automation adds to the final column:
“Dear Carrie,
I hope this letter finds you well. I am writing to convey my delight in working with you on the exciting Website Redesign project for the esteemed organization, Western Land Conservancy. As a firm, we have a rewarding history of creating impactful designs with meaningful missions. We take great pride in our extensive experience within the Non-Profit sector, and I believe that our deep understanding of the sector's unique needs and demands will fully cater to the needs of Western Land Conservancy.
Our team is looking forward to designing and developing a fresh, innovative website in line with your vision. We plan to leverage the reliable and dynamic capabilities of the WordPress content management system to build ten distinct page templates. Each page will be tailored to best serve its purpose while maintaining a cohesive feel to ensure brand recognition and enhance user experience. Our past feats with the WordPress platform make us the perfect fit for turning your vision into reality.
At the heart of our project plan, we will integrate features to enrich the user interaction and increase the organization's fundraising potential. The platform will feature a strong Mailchimp integration, designed to streamline your communication efforts and foster a stronger relationship with your supporters. For your fundraising needs, we will include an intuitive e-commerce platform that not only simplifies donations but also gives you the ability to manage and track contributions efficiently. We are committed to delivering a platform that meets your needs and goes a step further to keep your organization ahead of the curve.
Thank you, and we can't wait to get started on this inspiring project.
Best regards,
[Your Name]"
Boom! Say goodbye to writer’s block, and get your proposals off your to-do list at 10x the speed.
Use Case 3: E-mail Automation
Your brain is probably buzzing with new automation ideas – so let’s take a really powerful next step. You can forward your e-mail to Zapier and have it run automations for you.
I use this at Innovating with AI to solve a specific problem – sometimes, students sign up with two different e-mails in two different places, but I want to know that they’re one person so I don’t send them duplicate or unnecessary messages. For example, if you sign up for my ConvertKit mailing list with your personal e-mail, but then buy a course on my Thinkific site with your business e-mail, I would like to treat your personal e-mail address as a “course owner.”
The problem is that the specific data I need to create this connection exists only in one email from my course platform, and that e-mail is kind of a mess.
This e-mail is an automation nightmare because:
- It’s always structured differently (sometimes there is no user photo, sometimes there is no last name, etc.)
- The parts of the URL before and after the “subscriber_id” that I really need change depending on the first page the user visited and will never be entirely predictable
- There is no way to access this data via the API or any other type of structured request, it only comes in via e-mail
If you were a coder, you’d have to do something like “find the first https://, then find the next =, then find the next &” … and pray that the format of the URL and e-mail never changes.
Instead, with ChatGPT, you can use the prompt below, and it will be smart enough to find what you need every time, even if the patterns and structures of the e-mail change:
Prompt: “Below is the content of an e-mail. Provide me with the value of the subscriber_id parameter and the e-mail address that you find in this message. Return only the two values, separated by a comma.”
The manual approach takes maybe 30-60 seconds per e-mail, but when you multiply that by thousands of new students, it gets overwhelming. And I also can’t always do it immediately – if a user signs up at 12 AM, I won’t have time to map their multiple e-mail addresses before the 9 AM e-mail goes out the next day.
Here’s the Zapier workflow I built to do it automatically:
- Gmail: If sender is Thinkific and subject line contains “a learner signed up,” forward to Zapier
- Zapier: When the email arrives, send the body text to ChatGPT
- ChatGPT Prompt (Zapier Action): “Below is the content of an e-mail. Provide me with the value of the subscriber_id parameter and the e-mail address that you find in this message. Return only the two values, separated by a comma.”
- Zapier Formatter: Split the ChatGPT reply on the comma, so we now have two values (the subscriber_id and email address)
- Google Sheets (Zapier Action): Create a new row in my “User Mapping” spreadsheet with the email address as Column 1 and the subscriber_id as Column 2.
- ConvertKit (Zapier Action): Get the e-mail address for the subscriber matching the subscriber_id.
- Google Sheets (Zapier Action): Add the ConvertKit e-mail address from the previous step to Column 3.
- Zapier Filter: If column 1 and column 3 are different, proceed. If not, stop the workflow here. (We only want to do the next step if the user has two different e-mails on Thinkific and ConvertKit.)
- ConvertKit (Zapier Action): Add the tag “course-owner” to the subscriber with the e-mail address from Column 3.
Thousands of minutes saved – and we never send our customers the wrong message just because they signed up with two different e-mails. 🚀
Once you get handy with funneling your e-mail through Zapier and ChatGPT, you can do pretty much everything – even have ChatGPT respond to e-mails for you! The only limit is your imagination.
Use Case 4: Slack Summaries and Client Support
One of the biggest challenges at my web design firm is how to properly handle client communications and support requests without having the whole team distracted all day – after all, software developers work best when nobody is bugging them.
So, how do we balance quick responses with focus? Let’s bring Zapier, ChatGPT and Slack (our favorite communication tool) into the mix.
Here’s how this automation works:
- When a client e-mails support@[my domain].com, forward it to Zapier (see E-mail Automation for details on how to do this)
- ChatGPT (Zapier Action): Prompt: “Summarize the request the client is making in this email: [email contents]”
- ChatGPT (Zapier Action): Prompt: “Write a brief e-mail response to the message below, letting the client know we are working on their request and suggesting anything the client might do to resolve the issue themselves.”
- Slack (Zapier Action): Send a message to a public channel (#support): “New Support Request from [email address, pulled from Step 1] – [Reply from Step 2]. Potential Quick Response – [Reply from Step 3]
This ensures that our team is always up to date on support requests, rarely has to spend time opening their email (or a support ticket tool) and deciphering strange requests (the summary step handles that), and has a reasonable draft response ready instantly. And if a team member wants to ignore support for a while, they can just mute the #support channel in Slack.
🤓 Pro Tip: If you use Slack a lot, you can create even more advanced automations and use it as your “user interface” to your AI tools. For example, “if I send a block of text to the channel #summary, read the message and respond in the same channel with a summary of the text.”
Use Case 5: Smart Calendar Management
As you think about where AI can fit into your business workflows, it’s important to distinguish between things that are just repetitive and things that are both repetitive and variable. If it’s just repetitive, you could probably have automated it before ChatGPT existed. But the real challenge is the stuff that’s repetitive and variable – and that’s where AI really shines.
Let’s look at my calendar for some examples. I could already say, “Hey Siri, add a haircut to my calendar at 2pm on Tuesday” and (if she understood me) it would work. I can already click on the “Add to Calendar” button when I make an OpenTable reservation, and the process of creating the event is reasonably smooth. Calendly already creates a Zoom room for a new meeting.
The next level of calendar management is setting up your AI assistant to do the things you might have needed a personal assistant for in the past.
For example, I would like my personal calendar events to be a different color than my business calendar events. However, I am lazy and do not like to do anything but type the name of the event – I am not going to consistently choose from dropdowns to change colors or put things in different categories.
With Zapier and ChatGPT, we can automate the categorization of Personal and Business events – something you could never do before, because the input is highly variable (i.e. there’s no way for a coder to plan for every possible event title).
Prompt: Based on the title of this calendar event, is it more likely to be a Personal or Business event? Respond with only "Personal" or "Business"
Here’s how I set that up in Zapier:
- Google Calendar (Zapier Trigger): New or Updated Event
- ChatGPT (Zapier Action): Use the prompt above and drop in the title and description from the Google Calendar event
- Zapier Filter: Only continue if the ChatGPT Reply contains “Personal” (if it’s Business, the default color will be used, so we only change the color for Personal events)
- Google Calendar (Zapier Action): Update Event and change the color (there’s a list of colors you can pick from in Zapier)
🤓 Pro Tip: Zapier Filters provide a simple “stop or continue” conditional. Zapier Paths allow more complexity, and would be great here if we wanted three or more different categories rather than two.
And here’s my calendar, beautifully color-coordinated! Yellow is personal, blue is business, and I never think about it again. 🙂
