Documentation
Use Insight Agent on ad:personam to generate guided campaign insights, analyse flexible date ranges, and ask follow-up questions for faster optimisation.
Insight Agent is ad:personam's guided insights service. It combines report-specific analytics logic with conversational follow-up, so you can start from a structured campaign review and then keep probing the result without starting over.
This makes Insight Agent especially useful for optimisation work. You can isolate a date range, generate a focused first analysis, and then ask follow-up questions to uncover what changed, why performance moved, and where to reallocate budget next.
What Insight Agent Does
Insight Agent is designed for users who want more guidance than a blank query box, but more flexibility than a one-shot report. Instead of asking you to frame the whole analysis from scratch, it starts with a predefined insight type and campaign context, then lets you deepen the analysis through chat.
Key advantages include:
- Guided starting point — choose a report type instead of composing a complex prompt from zero.
- Flexible date ranges — focus the initial analysis on the exact window you want to optimise.
- Multi-Turn Follow-Up — keep asking questions about the same result without rebuilding the report.
- Campaign-specific context — start from the selected campaign and optionally narrow to a Line Item.
- Actionable output — designed for optimisation decisions, not just metric recaps.
Available Guided Report Types
Insight Agent supports seven guided report types:
| Report Type | Best For |
|---|---|
| Campaign Analytics | Delivery performance, pacing, spend efficiency, and campaign-level optimisation priorities |
| Creative Analytics | Understanding which creatives drive stronger engagement and which assets need replacement |
| Domains Analytics | Finding the sites and publishers that deserve more or less budget |
| Audience Analytics | Evaluating which audience segments are contributing efficiently to performance |
| Supply Analytics | Reviewing supply-path quality, buying efficiency, and inventory source behaviour |
| Geo Analytics | Identifying strong and weak regions for budget concentration or exclusion |
| Device Analytics | Comparing devices, environments, and operating systems to refine delivery strategy |
How to Use Insight Agent
Navigate to Reporting > Insight Agent from the sidebar.
Step 1: Configure the First Insight
The initial configuration card includes:
| Field | Required | Details |
|---|---|---|
| Language | Yes | Choose the language for the generated analysis. |
| Campaign | Yes | Select the campaign you want to analyse. |
| Line Item | No | Narrow the insight to a specific Line Item when you want a more focused review. |
| Report Type | Yes | Choose one of the seven guided report types. |
| Date Range | No | Limit the first analysis to a specific period, such as last 7 days, last month, or a custom optimisation window. |
| Initial Instruction | No | Add extra direction such as "focus on spend efficiency" or "look for scaling opportunities". |
Step 2: Generate the Initial Insight
Click Generate Insight. ad:personam runs the selected guided analysis, streams the underlying data into the interface, and builds the written answer in real time.
Step 3: Use Multi-Turn Follow-Up
After the first answer appears, use the Follow-up Questions box to keep exploring the same context. For example, you can ask:
- Which regions should I deprioritise first?
- Which Line Items explain most of the wasted spend?
- What should I test next week based on this result?
This is the core difference between Insight Agent and a static report. The first analysis establishes context, and the follow-up thread helps you turn that context into sharper optimisation decisions.
Why Date Range Flexibility Matters
Campaign optimisation rarely happens on all-time data alone. You often need to compare a recent window against current pacing, investigate a short period with wasted spend, or focus only on the latest learning cycle after a creative or targeting change.
Insight Agent lets you do that directly in the first run. Use shorter windows to diagnose active problems, or broader windows to identify stable patterns before scaling. Follow-up questions then build on the stored context, so you can keep investigating the selected window without resetting the analysis.
Best Use Cases for Insight Agent
- Weekly optimisation reviews when you need recommendations grounded in a recent date range
- Campaign diagnostics when you want to understand why efficiency changed
- Budget reallocation decisions across regions, domains, audiences, devices, or creatives
- Stakeholder reviews where you need both narrative interpretation and supporting data in one workflow
- Insight discovery when you want a guided starting point and then a conversational drill-down
When to Use Another Reporting Tool
Use AI Reports when you want the original one-shot OpenAI report experience with a strong narrative voice.
Use Query Agent when you already know the question you want to ask and need a fast ad hoc answer in natural language.
Use DSP Reports Service when you need exportable CSV data rather than a conversational analysis.
Related Articles
- AI Reports — the original one-shot AI reporting workflow
- Query Agent — ask natural-language questions for quick ad hoc analysis
- DSP Reports Service — build detailed CSV reports with custom metrics and dimensions
- Reporting Overview — compare all reporting tools available on ad:personam