The Dotdigital Marketing MCP server lets you connect AI tools such as Claude, ChatGPT, or Microsoft Copilot directly to your Marketing account, so you can work with your contacts, campaigns, and data using natural language.
If you're new to the MCP server, read these articles first. They explain how to connect, which permissions you need, and what each tool does:
What you can and can't do
You can ask about email campaign performance and content, industry benchmarks, contacts in a list or segment, individual contacts, and form responses. You can also create and update contacts and lists.
You can also create, send, and schedule email campaigns, create SMS campaigns, send single SMS messages, work with Insight data, and enrol contacts in programs.
You can't:
Search across all your contacts without choosing a list or segment.
Get contacts by campaign engagement or Insight data with Get contacts.
Delete contacts.
See which contacts opened or clicked a campaign.
Learn more in Dotdigital Marketing MCP server tools.
Use example prompts
The example prompts in this guide show common ways to work with the MCP server. You can use them exactly as written, or adapt them to your account, your naming conventions, and your goals.
Example prompts are grouped into categories:
Email campaign insights
Contacts and lists
Form insights
Account
Be specific about your criteria and dates
The MCP server works best when you're clear about who or what you're looking for, such as a location, subscription status, data field value, or list, and give a date range where relevant. If a prompt is vague, your AI tool may ask a follow-up question or return more than you expected.
Chain prompts together
You don't need to run every task as a separate request. You can chain prompts together, using the result of one prompt as the input for the next. See Form insights and list management for examples.
Ask your AI tool for thorough data analysis
Ask your AI tool to analyse data thoroughly, either in the conversation or in its general instructions. This helps avoid made-up numbers, wrong calculations, or analysis based on a sample rather than the complete dataset.
You can copy the following instructions, or write your own:
When analysing a dataset (file, query result, API response, log, spreadsheet),
process every record unless the user explicitly asks for a sample.
Rules:
1. Compute, don't eyeball. Never derive counts, totals, distributions, min/max,
or "most common" values by reading rows and estimating. Write and run code
over the full dataset and report what it returns.
2. Never extrapolate from a preview. The first page of an API response, or the
rows that fit in context, are for understanding structure only. Confirm
findings against the whole set before stating them.
3. State the denominator. Every quantitative claim must name how many records
it covers: "1,284 of 1,284 rows", not "most rows". If you did not process all
of them, say so and give the exact number you did.
4. Paginate to exhaustion. Keep fetching while a next-page token, cursor, or
hasMore flag is present. If a hard limit stops you, report where you stopped
and what remains.
5. Never invent values. If a field is missing, unreadable, or ambiguous, report
it as missing.
6. Account for every row. Report nulls, blanks, duplicates, malformed records,
and rows excluded by a filter as explicit counts. Excluded plus included must
equal the total.
7. Verify before concluding. Re-run the calculation, or check it a second way,
before presenting results.
8. Too big to process? Say so, don't guess. Say what the limit is and propose a
chunked approach, or agree a sampling method with the user and label every
output as sampled.
Avoid these words unless backed by a full computation: "approximately",
"roughly", "it looks like", "most", "typically", "should be around".
For anything involving percentages, rates, or comparisons between periods, add Double check any calculations to your prompt.
Ask for a chart
Once you have performance or benchmark data in the conversation, you can ask your AI tool to visualise it, for example, Now plot that on a line chart. You can keep adding to the same chart, such as another benchmark for comparison. Learn more in Visualise performance.
Name the connector if you have more than one
If you've connected other tools alongside Dotdigital, or more than one Dotdigital account, say which one you mean. This tells your AI tool which connection should act on the request.
Check before you change anything
Read what your AI tool is about to change before you approve it.
Before you update a contact, check they exist. If the email address doesn't match a contact, a new contact is created.
Results can vary between AI tools.
1. Email campaign insights
Email campaign performance
Get performance data and content for your email campaigns, over a date range or across their entire runtime.
Prompt |
How did my 'March newsletter' campaign perform last month? |
How did my 'Welcome' email campaign perform in February 2026 compared to February 2025? |
Give me the open and click rates for my last five campaigns |
What was the all-time performance of my 'Black Friday 2025' campaign? |
Compare click-through rates for my 'Spring sale' and 'Summer sale' campaigns |
Show me unsubscribe and complaint rates for campaigns sent last month |
Pull the content and subject line of my 'Product launch' campaign |
What's the bounce rate trend for my 'Abandoned cart' campaign over the last quarter? |
Which of my last ten campaigns had the highest click-to-open rate? |
Show me performance for all campaigns sent between 1 June and 30 June 2026 |
Give me a month-by-month breakdown of open rates for my 'Re-engagement' campaign this year |
Show me the subject line and preheader text used in my last three campaigns |
Which campaign in the last 90 days had the lowest open rate, and what was its subject line? |
Get all-time activity stats for my 'Annual report' campaign |
How does the performance of my campaigns with the tag 'Newsletter' compare month on month? |
How did my 'Re-engagement' campaign perform last month? |
Benchmarking
Compare your email engagement with marketing benchmarks, broken down by industry and region.
Prompt |
How does my email open rate compare to industry benchmarks? |
What's the average click-through rate for not-for-profit organisations in EMEA? |
Compare my unsubscribe rate to benchmarks for the travel industry |
What's a good bounce rate benchmark for financial services in North America? |
How does my 'Spring sale' campaign's click-to-open rate compare with my industry? Double check any calculations |
Is my open rate above or below average for my region and industry? |
Compare my campaign performance to industry benchmarks |
What's considered a strong click rate for hospitality brands? |
What's the average open rate benchmark for retail brands in APAC? |
Find patterns across campaigns
Ask what your best or worst performing campaigns have in common. Your AI tool reads and compares content across several campaigns to answer.
Prompt |
What do my best performing email campaigns this year have in common? |
What do my highest open-rate campaigns have in common in terms of subject line style or content type? |
Which subject line patterns tend to get the best open rates across my campaigns? |
Do my campaigns with a single clear call to action outperform ones with multiple links? |
Is there a common theme across the campaigns with the highest unsubscribe rates? |
Compare my top five and bottom five campaigns this year by click rate. What's different about them? |
Do shorter subject lines outperform longer ones across my recent campaigns in terms of open rates? |
What do my worst performing campaigns this quarter have in common? |
Create email reports
Turn campaign data into a summary you can share with your team or manager.
Prompt |
Create a Q1 email performance report that I can share with my team |
Put together a monthly performance summary for my newsletter, covering opens, clicks, and unsubscribes |
Create a year-to-date performance report for my email channel |
Build a report comparing this quarter's campaigns to last quarter's |
Create a report summarising my top five campaigns this year, with their open and click rates |
Draft a performance recap of last month's campaigns that I can send to my manager |
Visualise performance
Ask your AI tool to plot email campaign performance in charts. This works well in Claude, where charts appear in the conversation. ChatGPT and Microsoft Copilot don't currently show charts in the conversation.
Prompt |
Plot the open rate for my 'March newsletter' campaign on a line chart, broken down by month for this year so far |
Add the apparel and fashion open rate benchmark to that chart |
Now show that as a bar chart instead |
Chart my last six campaigns' open and click rates side by side |
Plot open rate against click rate for my last ten campaigns as a scatter chart |
Show my unsubscribe rate trend over the last six months as a chart |
Turn this month-by-month open rate breakdown into a chart I can use in a presentation |
2. Contacts and lists
Audience insights
Find contacts in a list or segment that match your criteria, or look up a specific contact.
Name a list or segment
Your AI tool can only search for contacts within a list or segment. It can't search across all your contacts. For large lists, ask how many contacts match rather than for every contact's details.
Prompt |
Find all contacts who live in France and have opted in to SMS |
Show me contacts added to my account in the last 30 days |
Get me the first name, last name, and email address for every contact in my 'Black Friday 2026' list |
Which contacts have opted in to SMS? |
Pull all data fields for contacts who signed up in the last week |
Find the contact record for jane@example.com |
Look up the contact with the mobile number +447700900123 |
Show me every list this contact belongs to |
What are this contact's current channel subscription statuses across email, SMS, and WhatsApp? |
Has this contact given consent for marketing communications? |
Show me contacts modified since the start of this year |
Find contacts with a loyalty tier of 'Gold' |
Find contacts created before March 2026 |
Find contacts with a UK mobile number so I can check SMS eligibility |
Show me the WhatsApp subscription status for contacts who joined in the last 90 days |
Show me consent records for contacts who signed up in the last week |
Find contacts whose email status is 'Suppressed' |
List management
Create lists, add contacts to them, and check whether contacts were added.
Prompt |
Add these contacts to my 'SMS' list |
Create a new list called 'Solo travellers' |
Create a list called 'Q4 webinar attendees' so I can add people to it |
Add these contacts from my CRM to my 'VIP' list in Dotdigital |
Check the status of the last contact import. How many were added, updated, or failed? |
Why did some contacts fail to import? Show me the reasons |
Add these details to my 'Customers' list, but don't overwrite any fields that already have a value |
Find contacts who live in France with a loyalty tier of 'Gold', then create a new list called 'France Gold members' and add them to it |
3. Form insights
Form performance
Get activity and response numbers for your forms, pages, and surveys.
Prompt |
Show me activity across all my forms this quarter |
How many responses has my 'Customer satisfaction' survey received this month? |
Which form in my account has the highest number of responses? |
How many people have submitted my 'Event registration' form so far? |
Compare response numbers across all my active forms this month |
What percentage of respondents to my 'Preferences' form selected email as their preferred channel? |
Qualitative insights
Ask your AI tool to read open-text responses and summarise themes, rather than just counting answers.
Prompt |
What are the top themes from our feedback survey? |
Summarise the common complaints in the open-text responses to our 'Post-purchase' survey |
What do respondents say they like most about our product, based on our 'Product feedback' form? |
Are there any recurring suggestions in the comments left on our 'Onboarding' survey? |
What's the overall sentiment of the free-text responses to our NPS follow-up question? |
Group the open-ended responses to our 'Customer satisfaction' form into common themes |
Create form reports
Turn form and survey responses into a summary you can share.
Prompt |
Create a report for my Q1 NPS feedback form that I can share with my manager |
Put together a summary report of responses to my 'Onboarding' survey |
Create a report showing response numbers and top themes from my 'Product feedback' form this quarter |
Build a one-page summary of my 'Customer satisfaction' survey results that I can send to the team |
Form insights and list management
Chain a form response query into a list action. Find the contacts who meet a condition, then act on them, all in one conversation.
Check your programs first
Before you add contacts to an existing list, check whether any programs use that list. Adding contacts may enrol them in the program.
Prompt |
Find all contacts who were highly satisfied based on our last NPS survey and add them to a new list called 'Promoters' |
Find respondents to our 'Product feedback' form who left a low rating, then add them to a list called 'At risk follow-up' |
Find contacts who said they're interested in solo travel on our 'Travel preferences' form, then add them to a new list called 'Solo travellers' |
Find contacts who said they're available for volunteering and add them to the 'Volunteering opportunities' list |
4. Account
Check which account you're connected to
Confirm which account and region your AI tool is connected to before you run other requests.
Prompt |
Confirm which account I'm connected to |
Give me the email address of the Dotdigital account I'm connected to |
Give me a summary of my account details |
What's the name of the account this MCP client is currently linked to? |
Work across more than one account
If you've connected more than one Dotdigital account, for example, one for each brand or region, your AI tool can work across all of them in the same conversation. Each account needs its own connection. Learn more in Connect to the Dotdigital Marketing MCP server.
Prompt |
What are my best performing email campaigns across my three connected Dotdigital accounts? |
Which of my connected accounts has the highest average open rate? |
Pull benchmarks for my industry, then check how each of my connected accounts compares |
Find contacts matching these criteria across all my connected Dotdigital accounts |