3 Powerful Nebula AI Agent Use Cases You Can Implement Today

Give access to custom mailboxes to your Nebula AI agent using Agentmail!!!

Note — Stuck behind medium paywall? Click here to read this blog for free!

Do you know what’s becoming harder with the rapid growth of AI models and tools these days? Finding the RELEVANT and REAL use cases where you should actually use AI in your daily life. You get to see new updates from big research labs all the time. Everyone is talking about it, but only a few are actually using it. In the words of Pedro —

thanks to X

There are many reasons why people are not able to use AI to solve real problems. One of them is that setting up AI agents via most tools available online is still kinda complex. They require some understanding of technical concepts before you can actually create useful AI agents.

While a few can do that, many are still hesitant to use AI directly. If you have read my previous blog, you already know that I was experimenting with Nebula AI Agents to see how I can use AI automation in my daily life. And that’s exactly where this platform comes in to help.

I was setting up a few use cases recently and realized that people either build overly complex integrations or go too simple for the sake of it, even if it doesn’t solve any real problem where AI could actually be useful.

Today, I’ll show you 3 simple yet powerful use cases that involve giving custom mailboxes to your Nebula AI Agent and seeing the possibilities where we can adopt AI in daily life to solve a real problem. Note that these are not the template use cases that AI platforms generally provide, but rather real problem-solving agents.

How to give custom mailboxes to your Nebula AI Agent?

Have you noticed that most of our new interaction with the internet happens via email in some way or another? Email is still the default interface of the internet.

The best way to integrate AI with the internet, as we humans do, is by having a mailbox. For any task on the web, we generally sign up. To communicate, we use email. To reach out to people or customers, we use custom emails. To hire, we use it. To apply, we use it, and many such problems where email becomes our first way of communication. In fact, on Nebula itself, we get a custom email for our account, which sends daily updates and reminders to our email ID to give us alerts.

Clearly, if we give a custom mailbox to our AI agent where it can read emails, reply to people, see attachments, find and sort emails with filters and keywords, etc., I think this will solve a lot of problems that we might not have thought of yet.

On Nebula, you have an option to connect your AI Agent with Agentmail. Agentmail gives AI agents their custom mailboxes. In fact, it is built specifically for agents, not humans. So if you want to perform any action via your AI agent that involves some form of communication via email, you should give it a try. All you need to do is go to settings, get your API key from Agentmail, and use it to connect to Nebula. The rest, AI will handle itself.

Getting AI newsletter daily digest via Nebula AI and Agentmail!

I’m sure you must have figured out multiple use cases just by reading the last few paragraphs.

The first and very minimal use case that I built via Agentmail and Nebula was getting my daily digest of the latest AI newsletters I have subscribed to. It’s weird, but I subscribed to at least 20–30 AI and tech newsletters via my personal Gmail, and I hardly read more than 2 or 3 sometimes!

One way for you to solve this is by giving access to your personal Gmail and then asking AI to specifically find only the newsletters, which I’m not a big fan of, unfortunately.

Another way is to give AI access to a custom email that it uses to subscribe to all these newsletters on your behalf and then send you a single daily digest of all the updates. It sounds good and is best if you are planning to find new newsletters from a different domain, maybe startups, finance, or robotics. You can have a custom mail for each type of newsletter or maybe a single mail subscribing and then categorizing in the final digest. But it’s useful when you let AI subscribe on your behalf. What if you have already subscribed and want to clear the mess from your Gmail?

You can simply create a “forward” on your Gmail of all your Substack newsletters so that every time you get a newsletter from a certain domain, for example, every mail coming from substack.com, it automatically gets forwarded to your custom agent mail, so that it only contains the newsletters that you have already subscribed to. This makes it even faster and might save you a few bucks finding relevant newsletters xd. This is important so that you do not end up losing tokens on irrelevant context, which you might not need for this specific use case.

go to gmail settings

I did this, and the results were as amazing as I thought. Nebula connects with your agent mail successfully, and it has access to certain capabilities that give superpowers to Nebula AI agents, like reading messages in your inbox, replying to emails, and reading attachments. More often, if a mail does not contain the required context and has hyperlinks, Nebula can take that URL, go to the website, create a more refined version of that newsletter, and then send it back to you.

So now, instead of you getting 20–30 newsletters in your mail, you get one 1 daily digest every day, which contains a summary of all the newsletters that you have subscribed to. So you do not miss anything and still end up cleaning your mess. You can schedule this agent to run daily at a certain time, and Nebula AI Agent will read your inbox, check if it has a new newsletter, and then send you a combined version at your scheduled time every day.

I personally loved this use case because this was the real problem I had, and Nebula solved this quite easily. Earlier, I had built one solution myself, but that was pretty hard to manage. Also, maybe I can have multiple agents running on Nebula at the same cost required to manage a standalone product I built for newsletters.

How to use Nebula AI Agent with Agentmail as a Founder?

This use case is especially relevant for startup founders who are looking to hire top talent for their company. One single post and you will get 1000s of entries for a job post. Without AI, filtering top candidates is difficult. With an AI agent, you can probably do this at a fraction of the cost. Even more optimized if you go with open-source models on Nebula.

If you are an early-stage startup founder hiring your initial team, instead of relying on Google Forms or any other portal, you can simply share your agent mail with candidates and ask them to share details on this mail with all the required information. They can share their resume, email, contact information, and other relevant details in the description.

You can imagine all of these emails would have completely different formats, descriptions, and details. Once you have connected your Agentmail on Nebula, you can ask it to go through all the messages, including attachments and portfolio links if shared, and then ask it to find the top 10 candidates for the next round.

I tried building this on Nebula and tested it end-to-end by myself. It simulated responses and sent 10 different candidate emails to my Agentmail inbox.

sample text showing how it sent multiple mails on my agentmail inbox

It runs API requests in parallel and can send multiple emails at once. Agentmail uses APIs to create, send, and receive emails in inboxes.

using Nebula mail to send details on Agentmail

I asked the Nebula AI Agent to send 10 different candidates as a dummy setup to test the validity of this problem statement. It sent all 10 to the custom inbox I created for this purpose. Most of them had descriptions, and a few of them had PDFs attached.

simulating 10 dummy profiles for testing this whole logic end to end

it can create a PDF during run time, store it and use it

Nebula AI uses its own code agent to generate and execute code automatically when working on such tasks.

code agent to write and execute code required for sub tasks

As the next and main step, I created another AI agent on Nebula called Hiring Agent, which would be responsible for finding relevant candidates based on my requirements from my Agentmail inbox. You might not have to do any of this unless you also want to test it yourself end-to-end before sharing the mail with candidates.

Using Hiring Agent to read mails and filter results

FYI, this is how the custom mailbox looks on the AgentMail website —

Agentmail Console

The moment I asked it to evaluate, Nebula sends API requests to Agentmail using the API key you provided, reads all the responses, and finds all the relevant profiles. If any of the mail has a PDF attached, you can clearly define these capabilities while creating the agent so that it reads PDFs/docs attached and visits portfolio links if required and provided.

Nebula using Agentmail APIs

Now, based on my requirement, it finds top candidates who meet certain criteria. You can customize this behavior while creating the agent and change the prompt as required. I asked for a minimum of 5 years of experience, and it filtered out irrelevant profiles.

Hiring Agent on Nebula to find top candidates

In the activity log in Nebula, you can find more details about what the agent is reading and how it is making its decision.

activity log of ai agent

Once it found the top 5 candidates, it gave the list in sequence of relevance to my criteria. It highlights the skills shared by the candidate, their education details, whether it matches all the requirements, and the impact they have created. If a candidate has mentioned these details, you already know that they were serious while applying for the job.

final result showcase

And then finally, a comparison table of top candidates, along with recommendations of profiles you can interview immediately.

comparison table for top candidates

top 3 recommendations

While this might seem like a lot at first sight, you know that creating a new AI agent on Nebula takes minutes. You just have to define it well, and that’s it. It’ll self-improve as you chat. If you haven’t read the previous blog on how to create AI agents, maybe give it a look.

But this use case related to hiring would be too relevant for young startup founders who do not have the budget to go for ATS systems and other expensive tools. You can now just build your own agent for all these tasks.

Regarding the third relevant use case to implement AI in your daily life, I’ll keep it to myself and would love to hear from you. If you had to build something using Nebula and Agentmail, what would you build? Maybe a sales agent? Or a customer service agent to talk to customers via custom mail? Or maybe reaching out to investors using Nebula? Possibilities are endless, but the most important thing is knowing what to do with AI. Once you are clear on your problem, AI can help you solve that quickly.

I’ll be trying out more use cases related to AI agents and will talk more about it soon. Till then, I hope you have subscribed to get the blogs in your mail directly. See ya soon. Cheers!