The best AI email assistants

Email is a digital hydra. For each one you bat out of your inbox, two others take its place. You finally pass the info on to your co-worker, only to drown in a jet of newsletters coming from nowhere. Wouldn’t it be great if AI could be your personal email Hercules?

It can, but don’t get too excited just yet: there aren’t a lot of options on the market. Most AI email apps have integrated OpenAI’s models into the compose window, but that’s about it. Sure, it’s a time-saver, but it only handles the writing part. Email is more than writing messages—it’s sorting, prioritizing, and deleting what doesn’t matter, too.

I spent two weeks researching and testing the best AI email assistants, and I’ve put together a shortlist of interesting options for you to try out—along with apps you should keep an eye on. And in case you’d like to build your own AI-powered email workflows, I’ve included a quick breakdown of all the ways this technology can help you do just that.

The best AI email assistants

• Sanebox for keeping your inbox organized

• Mailbutler for gathering contact details and tasks

• EmailTree for creating AI-powered workflows

What can AI bring to email?

This year, AI showed that it can understand and generate human language pretty well. While it’s not ready to work unsupervised, it’s a great tool to improve your existing automation with the ability to understand language. It can make a few decisions on its own, and it’s up to you to accept or edit the results.

But what does this mean when it comes to AI email apps? Let’s walk through the main AI features that you can use for automating your inbox.

Generative AI

Right now, the most common use of AI in email apps is for writing messages with generative AI. When composing a new email, you can start from a prompt—pretty much like in ChatGPT—or start generating a reply based on the sender’s message text.

Other features here include spell-checking your messages, rewriting with AI, offering quick reply suggestions, or even creating a pre-generated reply that you can send with one click.

Processing email content

Since AI can understand human language, it can analyze the content of each email. It can, for example:

• Run sentiment analysis on the message, providing an overview of how the sender is feeling. This is useful to understand if you need to be more tactful in your reply or if the person on the other side is open to learning more.

• Understand the topic of the email, which you can use to sort incoming messages with tags, labels, or folders. After your emails are organized, you can set up rules to forward each one to the appropriate person or just to sort the list with the highest priority at the top.

• Summarize the email or the entire thread, making it easy to file the conversation to analyze later or to bring new collaborators into the topic.

Data extraction and enrichment

Closely related to content processing, AI can also detect data types in your emails, extract that data, and store it in a database app. There are a lot of possibilities here. You can:

• Extract all the contact data from an email and add it to your CRM.

• Extract an invoice number and update information in your accounting app.

• Extract all the attachments from an email, read the contents of each file, and store them with the appropriate project pages in your project management app.

Data extraction pulls out information from an email. Data enrichment is almost the opposite: it surfaces information from your email history, your connected database apps, and public information to give you context on the current conversation. Imagine you’re talking to a friend you’ve known for 10 years. With data enrichment, you could summarize your entire email history and get some clues on where you should go on vacation next. 

Train your own AI language model to reply for you

You can train a model to contain all the relevant information about you and your business, so you can leverage that training to answer emails—and even produce other kinds of written content, too.

There are two ways to do this right now. If you have money to spare, you can hire a developer to fine-tune an OpenAI model. This is a method of tuning a GPT model with your data, helping it speak like you and know more information about you or your company.

If you’re looking for something easier, you can use a tool like Personal AI to train a personal language model rooted in your knowledge and writing style. You can then use Zapier to have it process all your emails and generate new drafts as necessary.

Combine them all

All of these features can be used with one another. Here’s an example:

1. Set up your AI to read everything in your inbox, extract relevant data, run sentiment analysis, and check your tech stack for relevant information.

2. Generate a reply with AI for each message based on the sender’s message and your data, organize emails with the most urgent first, and set your inbox to wait for your confirmation before sending. 

3. Pop by your inbox at the end of the day, start with email #1, and check the contextual information and pre-generated reply. If it looks good, hit send, and move to the next one.

And this is barely scratching the surface. The exciting thing about AI and email is that, depending on which features you use and the order in which you use them, you can get a completely different experience and results. It’s highly adaptable to each circumstance, meaning it can solve more complex problems—but, at the same time, it may take longer to reach a viable workflow.

What makes the best AI email assistant software?

Based on all the possibilities that this technology brings to the table, generative AI isn’t enough. The best AI email assistant software needs to go a bit beyond that. Here’s what I looked for as I was testing all the apps that claimed to be AI email clients. 

• Easy email integration. The integration process should be easy and fast with the most popular email services.

• Inbox management. I prioritized individual inbox management apps. For now, I’m leaving out apps that are focused on sales and outreach or customer service—I’m showing you how AI can transform your personal inbox, whether at home or at work.

• AI features. I was looking for AI features beyond generative text—for example, things like sentiment analysis, intent detection, automatic labeling, or data extraction.

• Productivity and automation. AI is one tool out of many, so these apps should offer other features and settings to help you save time and get to inbox zero faster.

• Data privacy. Your emails are private, so adding an app like this to your collection shouldn’t represent a security risk. I read the privacy policy of each app to understand how secure your data will be.

For two weeks, I dug the internet for all the AI email assistants that I could find. I filtered out those that just integrated generative AI into the compose window and was left with a very small selection. I signed up for each of them, configured the platform until it was ready to use, and sent a few emails back and forth between my email accounts. I played with the settings, observed how well the AI did what I wanted it to, and from there, picked these apps.

The best AI email assistants at a glance

Best AI email assistant for keeping your inbox organized

SaneBox (add-on)

SaneBox pros:

SaneBox cons:

SaneBox is really lo-fi. It sits in the background doing its thing. Once you go through onboarding and finish tweaking the settings, you won’t have to constantly log in to it. You can just use your inbox normally (on any client or device), and it will keep working.

When you start the configuration process, you can list the labels you’d like to use or have SaneBox do that for you. It’ll go through your inbox to weed out newsletters, conversations with friends and family, system messages from the apps you use, and a lot more. Once you finish running it, a few things will happen:

• You’ll see all your emails split by category, and you can easily trash useless messages (for example, email notifications from your apps).

Then, after that deep clean is complete:

• The important emails will stay in your inbox, so you can focus your attention on them first.

• All lower-priority emails will be added to the SaneLater folder. You can visit it once per day to read everything. It’s best to visit it frequently during the first weeks of using SaneBox: it’s still training, and its accuracy will improve over time. You can help train the model by dragging unimportant emails from your inbox to the SaneLater folder.

• You can create folders for any kind of topic or group of people, drag emails into those folders, and SaneBox will pick up on your intentions. It’ll start funneling similar emails in the future.

If you’re getting a lot of annoying messages from a single sender, you can throw them into the SaneBackHole: it’ll get all follow-up emails into its gravitational pull, so you’ll never hear from that person or organization again.

All these AI-powered sorting features will make sure that your attention will always be on your critical emails, keeping your inbox clean and organized throughout the day.

SaneBox price: From $0.07/day (or $24/year) for 1 feature on the Appetizer plan. If you want more than one email account and more than two features, you’ll need at least the Lunch plan, which is $12/month (or $169/24 months).

Best AI email assistant for gathering contact details and tasks

Mailbutler (add-on)

Mailbutler pros:

• Data extraction from email signatures

• Task extraction, plus integration with task management apps

Mailbuter cons:

Mailbutler went deep in integrating generative AI into its compose window. But it goes a little further too: it helps keep your contact information complete and can pick up on tasks mentioned in your emails, two features that set it apart from the competition.

It works as either a Chrome extension, Apple Mail extension, or Outlook add-in. It adds a sidebar to your inbox, letting you access the extra features there. You can access your contacts, notes, and tasks, as well as the contextual actions you can take when viewing an email.

Let’s start with data extraction. Mailbutler looks at the signatures of all emails you receive to grab phone numbers, addresses, and positions within a company. You can add extra fields if you’d like to, helping you keep a complete record of everyone you’re talking to.

In addition to signatures, it also processes the text of the emails looking for tasks, adding them on the sidebar so you won’t forget them. And since the app integrates with task management apps, you can send them straight over without having to retype anything.

The generative AI features are quite complete, too:

• Smart Compose writes an email for you from a prompt.

• Smart Respond lets you generate a positive or negative reply.

• Smart Summarize turns a long-winded email into a snappy snack.

• And Smart Improve keeps your grammar and spelling in check.

And all of this on top of a collection of useful features such as email open tracking, rich signatures, scheduled sends, team collaboration, and email snooze. 

Mailbutler price: From $4.95/month, but AI features are only available starting at the Smart plan for $14.95/month.

Best AI email assistant for AI-powered workflows

EmailTree (Web)

EmailTree pros:

EmailTree cons:

To start saving time, AI isn’t enough. You also need to have software that lets you configure triggers and the actions happening when those conditions are met. EmailTree brings both to the table: an in-house AI model and a tool to set up workflows and automation.

The AI model can do a nice set of things:

• AI email classification, helping to sort emails based on their content. You can set which categories you want to create, and EmailTree will start separating messages.

• Reply suggestions, which aren’t your traditional AI-generated emails. You have to write your own email templates and assign them as replies to emails. The AI model will train on these choices and automatically select which template is best for each message. This prevents an accidental AI hallucination making its way into your conversations.

• Sentiment analysis, which you can see on the left side of the screen when you open a message.

• Data extraction to squeeze out the important details of emails, ready to store in a database or to trigger automation.

• The possibility to generate a knowledge base with AI, saving the time it takes for you and your team to find answers when they need them.

After you’ve used EmailTree for a while, you’ll notice some incoming messages have a percentage next to the subject line. This is how much the AI trusts that it’s found a good reply for this message. You can click this percentage, see the reply suggestion, and if it’s good, send it out right away.

But the bulk of the time-saving features live in the Automation menu. There, you can pick a trigger from a long list—it can be any signal detected by the AI model or anything more traditional, such as a message from a specific sender address. Once you lock in on it, choose what happens when that condition is true (or when it isn’t). You can add as many steps as you want: forward the message to a team member, automatically reply with a message, classify the email with a custom label—among many other possibilities.

The pricing is a bit out of the ordinary with a three-year commitment for a total of $590 (that’s about $16/month, based on a promotion that’s still live as I’m writing this). If you think about it, it makes sense: the more the AI model trains, the better it’ll serve you, so there’ll be a higher likelihood you’ll stick around after the commitment period.

EmailTree price: From $590 for a full 3-year period (about $16/month).

Other AI email assistants to watch

There are a few AI email apps that didn’t quite make it due to my strict criteria, but I’d still like to share them with you. As they grow, they may incorporate more AI features, so definitely keep an eye out for them.

• Superhuman is a powerful email assistant with a great user experience. It offers robust generative AI features—but that’s that for now.

• Mailytica leans customer service, using AI to handle repetitive emails at scale. You’ll have your own model trained with your company’s messages, access email classification, email reply suggestions based on blocks (not on generation), sentiment analysis, and plenty of advanced features to save time and improve consistency.

• Lindy is an interesting app that goes beyond email: it’s developed to be your executive assistant, taking care of meeting scheduling, quickly sending a new message from a text prompt, pre-drafted emails, and task management. I’m looking forward to seeing this one grow.

Put your email on autopilot

I love the inbox zero feeling, but I confess I’ve been having it less and less, so it’s time to side with the bots. Having AI in your inbox means you’ll be able to be more agile with important messages, handle the lower-priority ones gracefully, and throw the rest out the window.

While we’re still waiting for more competition when it comes to AI email assistants, the apps above are a good start. You can also create your own AI email workflows using Zapier. Learn more about how to use GPT to automatically write business emails, sales emails, or email summaries with Zapier and OpenAI. Or get started with one of these pre-made workflows.

Related reading: