Martin Michaux

AI engineer, builder, founder · Head of AI at Herald · Forbes 30 Under 30

AI email automation: an agent that works your inbox

For most small and mid-sized businesses, work arrives by email: requests, orders, questions, documents. I build AI agents that work inside the inbox, so every email is read, sorted, answered or turned into a task within minutes, and your team only handles what really needs them.

Background

Experience: email AI at Herald

Herald offers brokerages an AI agent that works inside their email to extract client data and send submissions, and I lead the AI behind it. Agents that live in the inbox are one of the fastest ways to save time, because that is where the work already is.

  • Works in Gmail and Outlook, including shared inboxes
  • Reads attachments, not just the email text
  • Asks for missing information instead of stalling
  • Hands anything unusual to a person with a short summary

What the inbox agent does

Sort and prioritise

Every email is classified by type and urgency and sent to the right person or queue.

Extract the data

Names, references, orders and documents are pulled from the email and attachments.

Draft the reply

Routine questions get a reply in your tone, ready to approve or sent automatically once trusted.

Create tasks and update systems

Requests become tickets, orders, CRM updates or calendar entries without copy-paste.

Chase missing information

When something is missing, the agent asks for it and follows up if nobody answers.

Escalate to a person

Complaints, exceptions and anything unclear go to your team with a summary and a suggested next step.

Where it fits best

Shared inboxes

info@, sales@ and support@ addresses that several people watch.

Orders and quote requests

Customers who order or ask for prices by email, in their own format.

Customer service

Questions about orders, appointments, invoices and policies.

Back office

Supplier emails, invoices, confirmations and documents that need to be filed.

How I usually work

  1. 1

    Understand the work

    Look at how the task is done today, how often, by whom, and what a good result looks like. Often a simple automation is enough and no AI is needed.

  2. 2

    Pick one workflow

    Choose the task with the most repetitive hours and a clear definition of done, and measure it before changing anything.

  3. 3

    Build and test on real cases

    Build the workflow on real data and tools. At first a person approves the AI’s work, so nothing goes out unchecked.

  4. 4

    Measure, then expand

    Compare before and after. Let it run where it has proven reliable, then move on to the next workflow.

Frequently asked questions

Can AI answer my emails automatically?

Yes, for routine requests. Most businesses start with the agent drafting replies for a person to approve, then let it send on its own for the types of email where it has proven reliable.

Does it work with Gmail and Outlook?

Yes. It connects to Google Workspace and Microsoft 365 mailboxes, including shared inboxes.

Could it send something wrong to a customer?

Not without your rules allowing it. You decide which types of email it may answer alone, and everything else waits for approval. Every message it sends is logged.

What happens to emails it does not understand?

They go to a person, with a short summary of what the email is about and a suggested next step.

Is my data safe? What about GDPR?

It should be designed with GDPR in mind from the start: data kept in the EU where possible, access limited to what the AI needs, and a person approving sensitive actions. The setup also needs documenting for your privacy records.

Related articles

What building AI agents in insurance taught me about AI workflows

Article · From the field · 4 min read

AI agent, automation or chatbot: which one does your business need?

Article · Fundamentals · 4 min read

Thinking about AI for your company, or working on something similar? I’m happy to talk it through. You can send me an email or book a call.

Book a call with Martin