Marketing analysis with answers you can verify
Generated by Klaviyo AI
Klaviyo’s Composer answers plain-language marketing questions using account data and shows the definitions, queries, and sources behind its findings. The guide explains scheduled reporting, industry comparisons, and follow-up audits and drafts, while emphasizing that AI-generated analysis needs verification and changes require human approval.
- Checkable answers: Composer uses Klaviyo’s definitions and displays queries and sources so users can verify its analysis.
- Broader analysis: Questions can span customers, products, revenue, campaigns, flows, segments, and forms, with industry benchmarks providing context.
- Repeatable reporting: Reports and audits can run on a schedule, while custom Skills preserve a standard review method.
- Consistent definitions: AI tools connected through the Klaviyo MCP server use the same Klaviyo definitions as Composer.
"What percent of your customers never make a second purchase?" Ask Composer, Klaviyo's AI marketing agent, and you get the number back with the definition it used and the query it ran, so you can check the math before it goes in a deck.
That's how Klaviyo built AI analytics and reporting: you ask questions about your marketing in plain language, and an AI agent queries your data, explains what it found, and shows its sources. Composer runs that analysis across customers, products, revenue, campaigns, flows, segments, and forms. It can also run the same analysis on a schedule, so Monday's performance review is waiting for you instead of the other way around.
Marketing reporting usually breaks in two places. Pulling the numbers means exports and spreadsheets, and two people pulling the same number get different answers because they defined revenue differently.
Composer works from the definitions Klaviyo already uses, so the same question gets the same answer whether you ask in Composer or in another AI tool connected to Klaviyo.
What does AI analytics and reporting do?
AI analytics and reporting answers questions about your marketing, explains why the numbers moved, and keeps the reporting going after you close the chat. In Composer, that looks like this:
- It answers questions in plain language. Ask "Why are we pacing behind goal?" and Composer analyzes up to 365 days of account data, in your account's currency, to explain what happened.
- It goes past campaign metrics. Questions about customers, products, and revenue get answered alongside campaign and flow performance, so you can ask about repeat purchase rates in the same conversation where you check last week's sends.
- It shows its work. Composer turns your question into SQL using Klaviyo's own definitions of terms like revenue, then shows how it reached the answer, with citations to the sources behind its findings.
- It puts your numbers in context. Composer brings industry benchmarks into its analysis, so you can tell whether a dip in click rate is about your program or the whole industry.
- It recommends what to do next. Ask "What should I focus on this week?" and Composer names the highest-impact opportunity, explains why it matters, and suggests next steps.
For a closer look at reading campaign results, see the guide to campaign performance insights.
Key features of AI analytics and reporting in Composer
These features keep the analysis going after the first question:
- Reports run on a schedule. Schedule reports, performance reviews, and audits, like a review of your campaigns and flows each Monday, and Composer runs them without anyone remembering to ask. Scheduled analysis is separate from scheduling messages, so nothing goes to customers on its own.
- Your reporting method becomes a Skill. Save your standard weekly review, including the metrics you check and the order you check them in, as a custom Skill, and Composer runs it the same way each time.
- Reports can become shareable pages. Anything Composer makes can become an artifact, a live web page with its own link, so a performance summary reaches your team without a screenshot.
- The same answers hold outside Composer. Ask the same question in Claude, ChatGPT, or another AI tool connected through the Klaviyo MCP server, and you get the same answer, because the question runs against Klaviyo's definitions.
- Analysis leads straight to audits and drafts. When a number points to a problem, Composer can audit the flow or campaign behind it and draft the fix for your review.
Learn more about what Composer can analyze and create.
Benefits of AI analytics and reporting
The payoff shows up in the hours between a question and a decision:
- Monday starts with the review already done. A scheduled performance review lands with findings and what deserves your attention, so the first hour of the week goes to acting on them.
- Fewer debates about whose number is right. Answers use Klaviyo's definitions and show the query, so an analyst can check the math and two teammates asking the same question get the same answer.
- The people with questions can answer them. A retention marketer can ask about repeat purchase rates or segment performance without filing a request or building a report.
- Insight moves to action in one conversation. When the analysis finds a weak flow, you can audit it, draft the change, and approve it in the same thread. Nothing goes live without human approval.
Start with one question about your own data. Ask Composer what deserves your attention this week.
Frequently asked questions about AI analytics and reporting
What should I look for in AI analytics and reporting for email and SMS?
Look for three things: answers drawn from your own account data, a way to check how each answer was calculated, and reporting that repeats without a new prompt each time. Composer covers all three with plain-language questions, answers that show the query and sources, and scheduled reports and reviews. It works across campaigns, flows, segments, forms, customers, and revenue.
Can AI compare my email performance to industry benchmarks?
Yes. Composer brings industry benchmarks into its analysis, so you can see whether a change in open or click rate is specific to your program or common across the industry. For a reference you can browse, see Klaviyo's email marketing benchmarks and SMS marketing benchmarks.
Why do different AI tools give different answers about my revenue?
Usually because each tool guesses what "revenue" means. Klaviyo turns a plain-language question into SQL using its own definitions, so you get the same answer whether you ask Composer or an AI tool connected through the Klaviyo MCP server. Composer also shows the definition and the query it used.
Can AI monitor my marketing performance and flag problems?
Composer can run analysis and audits on a recurring schedule, such as a review of your campaigns and flows each Monday, and surface the findings that deserve attention. You decide what to act on. Scheduled analysis is separate from sending, and Composer doesn't send marketing messages on its own.
How do I decide what to do after a campaign ends?
Ask Composer "What should I focus on this week?" It looks across your campaign, flow, and performance data, names the highest-impact opportunity, explains why it matters, and recommends next steps. From there, you can ask it to audit the flow it flagged or draft the next campaign.
How accurate is AI analysis of marketing data?
Treat AI analysis as a strong first read that you verify. Composer shows the query and definitions behind quantitative answers and cites the sources behind its findings, which makes checking faster. Because Composer uses large language models, its output can be wrong, so review the numbers before you act on them or share them.