Active Intelligence prompts for campaign analysis
These prompts are based on the Your Database Isn’t Too Big. Your Segmentation Strategy Is Too Simple. webinar, published on the AI Lab by ActiveCampaign.
Get the prompt template
How to use these prompts
These prompts are pulled from the workflow Don Purdy demonstrated live—the analysis prompt he runs before every information-session campaign, plus the five questions he uses to keep the loop running after a send.
Use with: ActiveCampaign Active Intelligence—applies to every prompt below unless noted. Copy, paste, and swap in details where you see [BRACKETS]. Every AI model behaves a little differently, so review the first output and refine from there. Don’s demo prompts ran in seconds against his live campaign data—the answers will only be as sharp as the campaigns Active Intelligence has to read.
Prompt 1: Analyze a past campaign before you build the next one
Best for: Recurring campaigns where the same audience and offer come around again—monthly events, quarterly newsletters, repeat nurture pushes—and you want the next send to be built on what the last one taught you.
Why this prompt: Don ran this exact analysis live in the webinar before drafting his next information-session campaign. The reason he reaches for it: the work used to be a half-day spreadsheet exercise—exporting data, building pivot tables, hunting for the story buried in the analysis. Active Intelligence cross-references his entire campaign history in seconds and surfaces what to keep and what to change before he writes a single line of new copy.
Primary prompt
Analyze my last [CAMPAIGN TYPE] campaign across these dimensions: open rate, click-through rate, subject line, preheader, and call to action.
Show me what worked, where we underperformed, and where there’s room to grow. Benchmark the numbers against [INDUSTRY] averages where you can.
Then draft the next [CAMPAIGN TYPE] campaign incorporating the changes you’d recommend—enhanced subject line, optimized copy, clearer CTA, visual improvements—using my loaded brand components.
Variables to fill in:
- [CAMPAIGN TYPE]—the recurring send you’re analyzing (e.g., “in-person information session,” “monthly newsletter,” “re-engagement nurture”)
- [INDUSTRY]—your industry for the benchmark (e.g., “higher education,” “B2B SaaS,” “ecommerce apparel”)
What to expect: A read on the past campaign’s strengths and gaps, a benchmark comparison against your industry, and a drafted next campaign on-brand and ready to edit. In Don’s run, the analysis and the campaign draft both returned in under a minute.
Follow-up prompt:
For the drafted campaign, give me three subject line variants to A/B test against the recommended one. For each, tell me what’s different about its angle (urgency, specificity, personalization) so I can read the test result.
Prompt 2: The five questions to ask after every send
Best for: Closing the loop after a campaign sends so the next one starts from a sharper baseline. Don’s framing: stop pulling reports and start having conversations with your data.
Why this prompt: Don’s third action item from the webinar is “ask Active Intelligence five questions” instead of running another report. The five questions cover engagement, subject line patterns, conversion gaps, and the next move—the same set he uses every cycle so the analysis stays consistent across campaigns.
Primary prompt
Looking at my recent [CAMPAIGN TYPE] sends to [SEGMENT], answer these five questions:
1. Which segment engaged the most, and what does that segment have in common?
2. What subject line patterns are working—are questions outperforming statements, is specificity beating urgency, is personalization moving the needle?
3. Who engaged but didn’t convert? What’s the gap between open and action?
4. What time of day or day of week is producing the strongest response from this segment?
5. Based on all of the above, what should I test in my next campaign?
Variables to fill in:
- [CAMPAIGN TYPE]—the cohort of campaigns you’re reading across (e.g., “event invitation,” “monthly nurture”)
- [SEGMENT]—the segment you’re focused on (e.g., “MBA prospects,” “registered alumni,” “dormant contacts”)
What to expect: Five short answers, each with the underlying data Active Intelligence pulled to back the conclusion. Treat the fifth answer as your hypothesis for the next campaign, not a guarantee—the system surfaces patterns, you decide what to test.
Follow-up prompt
Of the five answers above, which one would change my next campaign brief the most if I acted on it? Walk me through what the next campaign would look like if I prioritized that finding.
Prompt 3: Pattern-match across your top campaigns
Best for: When you’ve sent enough campaigns that individual analysis stops being useful and you want to see what’s true across all of them—the cross-campaign signal that no single send can show.
Why this prompt: Don described this as the question he asks Active Intelligence after every campaign closes: “What patterns are showing up across my top campaigns?” The answers shape the next brief—specificity, urgency, personalization, time of send—and they get sharper every time the loop runs.
Primary prompt
Look across my top 10 performing campaigns from the last [TIME WINDOW]. What patterns are showing up in the ones that worked?
Specifically, compare:
- Subject line structure (length, question vs. statement, specificity, urgency cues)
- Send time and day
- Segment size and profile
- Call-to-action wording and placement
- Email length and visual density
Tell me which two or three patterns are most strongly associated with the wins, and which I should test in the campaigns that aren’t performing yet.
Variables to fill in:
- [TIME WINDOW]—the period you want to read (e.g., “six months,” “this academic year,” “since January”)
What to expect: A short list of cross-campaign patterns with the data behind each, plus the two or three you should apply to your underperforming campaigns. Save the patterns that hold up across windows—they’re worth more than any single campaign’s read.
Follow-up prompt:
Of the patterns you just surfaced, which one is the most actionable in the next two weeks—meaning I could change a campaign in flight or queue up a test? Give me the specific change to make.
Tips for better results
- Run the analysis before you draft, not after. Don’s whole point: ask Active Intelligence what the last campaign taught you before you open a new campaign editor. The draft you’d have written from a blank page is rarely as good as the one shaped by the last data point.
- Load your brand components first. When you ask Active Intelligence to draft a campaign, the output pulls from the brand components you’ve loaded. If the draft looks off-brand, the brand components need an update—not the prompt.
- Be specific about industry and segment. The benchmark and segment fields do most of the work. “Higher education” and “MBA prospects” tell Active Intelligence which slice of data to read against. “Marketing” and “everyone” don’t.
- Treat the AI as a strategist, not a dashboard. Don’s framing: “It’s not just telling you, send this email on Tuesday. It’ll give you insights like send this email to people who have engaged with career content in the last 30 days, and here’s a subject line that performs best for them.” Ask the question that has a strategy answer, not one that has a number answer.
- Save what works to a running prompt library. ActiveCampaign also publishes a prompt library backed by real user prompts. When one of your own prompts produces a clean result, save it next to those—your next campaign starts three turns ahead.
See these prompts in action
Want to see Don run the analysis prompt live? The webinar shows the moment Active Intelligence pulls the past campaign’s metrics, surfaces the gaps, and drafts the next campaign with brand components applied—in seconds, while Don and Erin keep talking. The full recording also includes Don’s three-step homework for getting started this week.
Related
More data from the AI Lab.