Persado MCP adds Persado's predictions, based on what customers did in more than 120,000 campaigns, to the AI assistant you already use. Paste in an email, a subject line or a banner headline, ask for Persado by name, and you see how well it is likely to perform with customers and get better-performing options, without leaving the conversation.
If you write subject lines, push notifications or ads in an AI assistant, you have probably asked it which version is better. It will give you an answer. What it can't tell you is how each version is likely to perform with customers, because it has never seen the results. Persado has. Persado is a platform that predicts how well emails, banners and other customer messages will perform, learned from what customers actually did in those campaigns.
What is Persado MCP?
MCP stands for Model Context Protocol, an open standard that lets an AI assistant use a tool it doesn't have built in. Persado MCP is the Persado connector for that standard. You add it to your assistant once. After that, you ask for Persado by name whenever you want to check a piece of content, and the answer comes back in the same chat.
There is no separate portal to open and no file to export. The draft, the question and the answer stay in one thread.

Illustrative example. Persado connected inside an AI assistant.
What you get
A performance prediction, and the words moving it
Persado gives each draft a score: a number from 0 to 100 that predicts how well it will perform with customers. It comes with the specific words and emotions pushing it up or down, so you know what to change, not just that something is off.
Better versions, ranked
Ask for alternatives and Persado returns several, listed best first. "Ranked" simply means ordered from most to least likely to perform. Each one shows its predicted lift against your own draft, so you can see how much better the top version is likely to be, not just that it is different.
Answers from your own results
Ask what has been working across your campaigns and get a plain answer, such as which emotion led last quarter. No dashboard to open.
Brand and compliance, for Enterprise
Enterprise teams can also check content against their brand voice and regulatory rules in the same chat. For financial services, that means problems are caught before launch.
What it looks like in practice
This is an illustrative example, not a real result. You type: "Persado, score this subject line before I send it: 'Your statement is ready.'" Persado answers that the line scores 58 out of 100, and that a generic opener is holding it back because nothing in the line tells the reader why to open it today.

Illustrative example. The subject line scores 58 out of 100, with your own line as the reference.
You ask: "Give me better versions, ranked." The top version, "Your May statement: one thing to check," scores 81, a predicted lift of about 10% in conversions compared with your line. Your original sits at the bottom of the list as the reference point.

Illustrative example. Three versions ranked best first, with your original line as the reference.
A score on its own is only a number. Next to your original and the other versions, it tells you what to send.
Why not just ask the assistant?
General-purpose assistants learn from the open web. They write well, and they are good at producing five versions of a line in a few seconds. What they don't have is a record of which of those versions customers actually responded to.
Persado's predictions come from labeled campaign results, with a learning loop that keeps improving them. When your assistant calls Persado, it is using that record of what worked, not a guess about what reads well. Your assistant keeps doing what it is good at: drafting, editing and explaining. Persado adds the part it can't do alone, which is a prediction grounded in results.
Where it works
Persado MCP works in any LLM assistant. It is available in the Microsoft Copilot marketplace and the Claude directory. You can connect it manually to ChatGPT and Gemini Enterprise, or to your in-house model. Connecting takes about five minutes, once.
Who it is for
Individuals and teams
If you write emails, banners, push notifications or ads and want data behind the choice between versions, this is built for you. You can start on your own, with no sales conversation.
Enterprise teams
Persado MCP works alongside Persado Studio, so teams can check content where they already work. 8 of the 10 largest US banks use Persado. [LEGAL TO CONFIRM WORDING]
Try it free for 14 days
Open your AI assistant, add Persado and score one piece of content you are working on this week.
FAQ
What is a Persado score?
A number from 0 to 100 that predicts how well a piece of content will perform with customers. It is learned from real campaign results, not from how the text reads. It always comes with the words driving it and a reference point, such as your original draft.
Which assistants does it work with?
Any LLM assistant. It is in the Microsoft Copilot marketplace and the Claude directory today. ChatGPT and Gemini Enterprise connect manually, and so does an in-house model.
What does the free trial include?
15 calls over 14 days, with no card. One score or one set of better versions counts as one call. The trial covers predicted performance on your own content. The 14 days start from your first rewrite.
Does it check my brand rules and compliance?
Brand and compliance checks are part of Enterprise. The free trial scores performance only.
Do I need a sales call to try it?
No. You connect it, sign in and start.

![[PRODUCT UPDATE] Three new ways to use Persado](https://a.storyblok.com/f/291846000560732/ec5ff57a41/q3-rundown-cover-v2.png)

