Most Marketing AI Is Still Just Chat

Hey everyone,

Welcome back for another bite to chew on.

Every tool in your stack shipped an AI feature this year. And almost every one of them is the same thing: a chat window bolted onto the sidebar that gives you advice you could have gotten for free. It writes a subject line. It suggests "best practices." It does not know your list, your revenue, or your flows.

Meanwhile the actual work of email marketing has not changed. You still open the dashboard, dig through reports, build the segment by hand, and stitch the campaign together screen by screen. The AI watches from the sidebar.

Omnisend just did something different. Instead of building a better chatbot inside their platform, they plugged the entire platform into Claude and ChatGPT. Your account data, your segments, your automations, your campaigns: all of it reachable from a single prompt in the chat you already use. Let's get into it.

On the Menu:

If your AI assistant cannot see your revenue, your list, or your automations, it is not an assistant. It is a search engine with better manners.

That is the gap between AI that talks about your marketing and AI that can actually touch it. Omnisend closed it. Their MCP connection plugs your full account into Claude and ChatGPT, and the latest update is the biggest upgrade the connection has had yet.

Switching from a different ESP? Migrate and in 5 days you could be paying up to 35% less.

→ Migrate to Omnisend in 5 days

The AI conversation in ecommerce has been stuck on the same question for two years: can it write good copy? Wrong question. The right one is whether it can do the work.

Open any tool in your stack and there is a sparkle icon somewhere. Click it and you get a general-purpose assistant that knows nothing about your business. It can explain what an abandoned cart flow is. It cannot tell you that yours has been quietly underperforming since March.

That is the ceiling of in-app AI. It is trained on the world, not on your account. So the advice is always generic, and generic advice is not the bottleneck for any operator we know. Execution is.

A real assistant pulls the report before the meeting. It notices the automation that stopped converting. It drafts the campaign, plans the send, and flags the deliverability problem before it costs you a sending domain.

That is what the Omnisend connection turns Claude and ChatGPT into. Because the AI is connected to your actual account through MCP, it stops guessing and starts reading. You ask a question about your marketing, and it answers with your numbers.

Connecting it is not a technical project either. Claude connects within a few clicks. The official app is coming very soon, and until then the custom connection takes minutes to set up.

Ask a generic chatbot "how do I improve my email revenue" and you get a listicle. Ask an AI that can see your Omnisend account and you get something closer to a consultant's audit: which campaigns actually drove revenue, which automations are pulling their weight, where subscribers are leaking out of your funnel.

Same model. Same chat window. Completely different output, because the input is your real data instead of the internet's average opinion.

The fastest way to understand this is to see what operators are actually asking. These are real, repeatable prompts pulled from the ones driving the most business value.

"Run a full health check on my Omnisend account. Score me out of 100 across three areas: campaign cadence, automation coverage, and product catalog completeness. Then give me the top 3 things I should fix first."

One prompt, and you get a scored audit of your entire email program plus a prioritized fix list. This is the kind of review most brands pay an agency to run quarterly.

"Look for underperforming areas in my lifecycle marketing and suggest the 3 highest-impact fixes for this week."

This is the prompt for Monday morning. Instead of scanning dashboards hunting for problems, you ask the AI to find where money is leaking and hand you the shortlist.

"Which of my automations made the most money in the last 30 days? For my top 3, break down which individual message inside each one is doing the heavy lifting, and which ones aren't pulling their weight."

Most operators know their welcome flow "works." Far fewer know that email 2 inside it is doing 70% of the revenue while emails 4 and 5 do nothing. This prompt finds that.

"Give me a monthly summary I can share with my team or client. What worked, what didn't, how this month compared to last month across campaigns, automations, revenue, and subscribers, and what we should focus on next month."

If you run email for clients, this one alone pays for the setup time. The month-end report writes itself from live account data.

Two notes for the cautious. First, the new granular permissions mean you can grant read-only access, so the AI can analyze everything without being able to change anything. Second, AI-created emails are now editable, so nothing goes out without your hands on it. Test your creation prompts before you send, like you would with any new workflow.

Here is the bigger question underneath the feature news: is your ESP going where email is going? Every ESP decision is a platform bet, and this update shows you what Omnisend is betting on.

When you pick an email platform, you are not buying today's feature list. You are buying the next three years of the roadmap. The connection to Claude and ChatGPT gets smarter, more flexible, and more powerful every month: granular permissions landed, email creation got sharper, and the pace is not slowing.

If your current platform's AI story is still a copywriting widget in the sidebar, that tells you something about where their roadmap is pointed.

The honest reason most brands stay on an ESP they have outgrown is that switching sounds painful. Rebuilding flows, re-uploading lists, recreating templates: nobody has a spare month for that.

Omnisend removes that reason entirely. Their team does the migration for you. Give them 5 days and show up to everything set: contacts, automations, segments, templates, all rebuilt and ready to send.

The move is not just about the AI. The numbers stand on their own: you could be paying up to 35% less than you pay today, on a platform where brands earned $79 for every $1 spent in 2025. Add 250+ ecommerce integrations, a 4.8/5 average Shopify rating from the 150,000+ brands running on it, and real human support that answers in 4 minutes on average.

Cheaper, better connected, and the ESP that lives inside your AI assistant. That is a hard combination to argue with.

Switching from a different ESP? Migrate and in 5 days you could be paying up to 35% less.

→ Migrate to Omnisend in 5 days