Close



Data Sources



General

Data Sources | General

Fields & Formulas

Data Sources | Fields & Formulas

Connections

Data Sources | Connections

Reports



Creating & Sharing

Reports | Creating & Sharing

Report Editing

Reports | Editing

Report Config

Reports | Report Config

Chart Cards

Reports | Chart Cards

Content Cards

Reports | Content Cards

Data Chat



General

Data Chat | General

AI Visibility



General

AI Visibility | General



General

AI | General

MCP (Model Context Protocol)

AI | MCP (Model Context Protocol)

Clients



General

Clients | General

Integrations



General

Integrations | General

Webhook — Local & Reviews

Integrations | Webhook — Local & Reviews

Webhook — Automation Payloads

Integrations | Webhook — Automation Payloads

Webhook — Other

Integrations | Webhook — Other

Settings



Account & Billing

Settings | Account & Billing

Users

Settings | Users

Comments & Notes

Settings | Comments & Notes

Notifications

Settings | Notifications

FAQ

Settings | FAQ

Other

Settings | Other

KPI Forecasting



General

KPI Forecasting | General

Settings & Status

KPI Forecasting | Settings & Status

Seasonality

KPI Forecasting | Seasonality

Metrics by Source

KPI Forecasting | Metrics by Source

Budget Tracking



General

Budget Tracking | General

Anomaly Detection



General

Anomaly Detection | General

Automation Jobs



General

Automation Jobs | General

Change History

Automation Jobs | Change History

URL & Page Checkers

Automation Jobs | URL & Page Checkers

Local & Reviews

Automation Jobs | Local & Reviews

Search Console

Automation Jobs | Search Console

Insights



Managing Insights

Insights | Managing Insights

Change History

Insights | Change History

URL & Page

Insights | URL & Page

Local & Reviews

Insights | Local & Reviews

Other Insight Types

Insights | Other Insight Types

Local Listings



General

Local Listings | General

Profiles & Content

Local Listings | Profiles & Content

Performance & Activity

Local Listings | Performance & Activity

Media

Local Listings | Media

Categories & Services

Local Listings | Categories & Services

Reviews



General

Reviews | General

General

Social Posts | General

Creating Posts

Social Posts | Creating Posts

Managing Posts

Social Posts | Managing Posts

Data Warehouse



Pipelines

Data Warehouse | Pipelines

Access & Credits

Data Warehouse | Access & Credits

Source Fields

Data Warehouse | Source Fields

Data Studio

Data Warehouse | Data Studio

BigQuery

Data Warehouse | BigQuery

General

Clients | General

General

Integrations | General

MCP

Integrations | MCP

Webhook — Automation Payloads

Integrations | Webhook — Automation Payloads

Webhook — Local & Reviews

Integrations | Webhook — Local & Reviews

Webhook — Other

Integrations | Webhook — Other

Account & Billing

Settings | Account & Billing

Users

Settings | Users

Comments & Notes

Settings | Comments & Notes

Notifications

Settings | Notifications

FAQ

Settings | FAQ

Other

Settings | Other

Recent Updates



Account Settings

Settings | Account & Billing

Acknowledging and Resolving Insights

Insights | Managing Insights

Add comment from Slack

Integrations | General

Add comment from Zapier

Settings | Comments & Notes

Adding a User

Settings | Users

AI Analysis Generation

Reports | Creating & Sharing

AI Credits

AI | General

AI Models

AI | General

AI Visibility Insights

AI Visibility | General

AI Visibility Overview

AI Visibility | General

AI Visibility Settings

AI Visibility | General

AI Visibility Webhook Payload

Integrations | Webhook — Other

Analysis Card

Reports | Content Cards

Anomalies Bulk Action

Anomaly Detection | General

Anomaly Algorithm

Anomaly Detection | General

Anomaly Detection

Anomaly Detection | General

Anomaly Detection Insight

Insights | Other Insight Types

Anomaly Detection Settings

Anomaly Detection | General

Anomaly Detection Webhook Payload

Integrations | Webhook — Other

Anomaly Job Common Errors

Anomaly Detection | General

An overview of Local Listings

Local Listings | General

An overview of Reviews

Reviews | General

Answers

AI Visibility | General

Area Chart Card

Reports | Chart Cards

Attributes

Local Listings | Profiles & Content

Automatic Schema Updates

Data Warehouse | Pipelines

Automation: Facebook Ads Destination URL Checker

Integrations | Webhook — Automation Payloads

Automation: Google Ads Change History

Integrations | Webhook — Automation Payloads

Automation: Google Ads Destination URL Checker

Integrations | Webhook — Automation Payloads

Automation: Google Analytics Change History

Integrations | Webhook — Automation Payloads

Automation: Google Business Profile Answer

Integrations | Webhook — Local & Reviews

Automation: Google Business Profile Change History

Integrations | Webhook — Automation Payloads

Automation: Google Business Profile Pending Updates

Integrations | Webhook — Local & Reviews

Automation: Google Business Profile Question

Integrations | Webhook — Local & Reviews

Automation: Google Review

Integrations | Webhook — Local & Reviews

Automation: Google Review Change History

Integrations | Webhook — Automation Payloads

Automation Job Common Errors

Automation Jobs | General

Automation Jobs

Automation Jobs | General

Automation: Microsoft Ads Destination URL Checker

Integrations | Webhook — Automation Payloads

Automation: Page Status Checker

Integrations | Webhook — Automation Payloads

Automation: Robots.txt Change Detection

Integrations | Webhook — Automation Payloads

Automations Bulk Action

Automation Jobs | General

Automation: Search Console Activity

Integrations | Webhook — Automation Payloads

Automation: Search Console Keyword Movement

Integrations | Webhook — Automation Payloads

Auto Resolving

Insights | Managing Insights

Bar Chart Card

Reports | Chart Cards

Branding

Reports | Creating & Sharing

Budget Campaign Dynamic Filter

Budget Tracking | General

Budget Common Errors

Budget Tracking | General

Budget history

Budget Tracking | General

Budgets Bulk Actions

Budget Tracking | General

Budget Settings

Budget Tracking | General

Budget Status

Budget Tracking | General

Budget Tracker Webhook Payload

Integrations | Webhook — Other

Budget Tracking

Budget Tracking | General

Budget Webhook URL

Budget Tracking | General

Build Reports with Jepto MCP

AI | MCP (Model Context Protocol)

Bulk Action

Local Listings | General

Bulk Upload Images

Local Listings | Media

Calendar

Social Posts | General

Cancel Subscription

Settings | Account & Billing

Card Chat

Reports | Editing

Card Data Filters

Reports | Creating & Sharing

AI Analysis Generation

‍
‍

When you use AI to create an Analysis, Jepto hands the reporting period to a dedicated background analysis agent. It researches your connected data, checks KPIs, budgets, recent account activity, and anomalies, then writes a narrative from what it finds.

Every analysis is shown in an Analysis card. The card is a container that shows an analysis when the report's date range overlaps the analysis's date range. For example, a report set to October 15 to November 30 shows an October analysis, but a report set to December 1 to December 31 does not.

1The agent researchesYour dataKPI targetsBudget pacingAnomaliesRecent changesAlso usesReport audienceAdditional contextRules

2It saves a Draft periodDraftReporting periodOctober 2026HeadlineContent

3You review and publishPublishedOn the reportThe headline shows on the Analysis card. Open full analysis shows the content.

What viewers see on the report The Analysis card on a report, showing a Summary headline and an Open full analysis button

Each new analysis is a draft, so nothing reaches the live report until you publish it.

At a glance

This is a full agent with its own tools. It queries your data, reads KPI targets, budget pacing, anomalies, and recent account changes, then writes from what it actually finds. It never invents a number.
It runs fully autonomously once started and does not pause to ask you a clarifying question. If something is ambiguous (like the exact reporting period), it makes the most reasonable call itself and finishes the draft.
Generation can take a few minutes per period as it's reasoning over your data and will not return straight away.
The analysis is saved as a Draft so that nothing goes live before you have had a chance to review it. To make it visible, you need to set the status to 'Published'. For more, see Analysis Card
Use Additional context to steer it toward what actually matters this period.
Report audience tailors the commentary - Business owner, Marketing manager, Finance team, Agency team, Client contact, and Leadership each produce a differently structured narrative and tone.

‍

How the agent works

Analysis generation utilizes a fully configurable AI agent that acts as a research assistant: it plans a short checklist of what it needs, calls a sequence of tools to gather real numbers and context, and only writes the narrative once it has them. Jepto's instructions to the agent explicitly forbid reporting any number that didn't come back from a tool call, and require converting raw platform values (like spend in micros) into readable currency before writing anything down.

Tools available to the agent

Category	What it can check
Your data	Which sources are connected, the same metrics/dimensions catalog as the rest of Jepto (Data Model), and live query results for period totals, campaign breakdowns, and trends
This report	The charts and tables already on the report, so the narrative adds context instead of repeating what's already visible - plus where on the page to place the analysis card
Performance signals	KPI targets (actual vs. target), budget pacing (on/off track), flagged anomalies, and recent account actions (bid changes, campaign launches/pauses) to explain why a number moved
Web Search	Allows the agent to conduct research across the web. If used within the analysis the agent is requested to cite the sources within the footer.

The agent also loads a channel-specific playbook depending on what's connected for the client - paid search, social, SEO, local SEO, email, or a blended approach when several channels are relevant - so the structure and focus of the narrative adapts to the account rather than using one generic template.

‍

Inputs

Field	What it does
Reporting period	One or more full calendar months. The agent can't analyze a partial or still-in-progress month.
Report audience	Who the narrative is written for - Business owner (default), Marketing manager, Finance team, Agency team, Client contact, or Leadership. Genuinely changes what the narrative leads with and how it's structured, not just tone - see Report audience below.
Additional context	Freeform notes - strategy changes, what to focus on, or anything unusual this period. You can also paste the full page item JSON from your report (layout, card ids, titles, grid positions) - there is no size limit. This is the main way you steer the output; see example prompts below.
AI model	The same model picker used elsewhere in Jepto (e.g. Card Chat, Data Chat) - the available list is account/config-driven and changes over time.
Account Rules	Organisation-wide instructions from Rules - tone, terminology, and things to avoid. Applied automatically for every run; you do not paste these into Additional context.
Client Rules	Per-client instructions from Rules. Applied automatically alongside Account Rules for the selected client.
Client background context	Industry, services, positioning, and profile fields from Client settings. Applied automatically so the narrative uses the right business framing.

‍

Report audience

Tailoring the reporting commentary to the intended audience - each option maps to its own full set of writing rules that change which sections expand or collapse, what the narrative leads with, and how technical the language is allowed to get.

Audience	What it changes
Business owner (default)	Collapses Channel Performance into one summary table per channel; skips Campaign Breakdown and Trends unless something critical shows up; closes with exactly 3 prioritized actions, each with an estimated dollar or lead impact.
Marketing manager	Leads with channel efficiency (return on ad spend, cost per lead/acquisition); expands Campaign Breakdown grouped by objective; emphasizes funnel metrics over impressions; calls out creative or audience signals (e.g. fatigue, quality score drops).
Finance team	Leads with total investment vs. total return; expands budget pacing per channel (spend, % of budget used, days remaining, projected month-end variance); flags any channel more than 10% off pacing; frames recommendations as budget reallocation with an estimated impact.
Agency team	Full metric tables in every section, all campaigns (not just top spenders), a week-by-week Trends section, and a short narrative after every table explaining what drove the movement - the most exhaustive option.
Client contact	Plain-language monthly update; opens with a short "what changed and why it matters" narrative; keeps Channel Performance to the 2-3 channels that mattered most; owner-ready recommendations; collaborative, reassuring tone.
Leadership	Strategic readout grouped into themes (growth engines, efficiency gaps, budget risk) instead of a channel-by-channel walkthrough; closes with a 3-item decision agenda; prefers 3-month trend/trajectory framing over single-period detail.

All six options share the same underlying data and tool calls - only the framing, structure, and level of detail change. Pick the audience that matches who will actually read the card, and use Additional context for anything specific to this run (a campaign change, a goal shift) rather than trying to convey it through the audience choice.

‍

What it won't do

It won't ask you questions mid-run. It's a one-shot background job - once you start it, there's no way to answer a follow-up. If the result misses the mark, edit it directly or regenerate with more detail in Additional context.
It won't invent numbers. Every figure in the draft came from an actual query against your connected data.

‍

Example prompts for Additional context

"Focus on lead volume and cost per lead this month - deprioritize impressions and reach."
"We paused the Google Ads brand campaign on the 12th - factor that into any spend or lead volume changes you explain."
"Call out budget pacing for each channel and flag anything more than 10% off track."
"Keep this brief - a short executive summary is enough, skip the campaign-level breakdown."
"This client's main goal is revenue, not lead count - frame recommendations around revenue impact."

‍

Tips

Name specific events - a paused campaign, a new budget, a shifted goal - in Additional context. The agent already checks recent actions and anomalies automatically, but naming the one that matters sharpens the explanation.
Stick to full calendar months - the reporting period can't include the current, still-running month.
Treat the draft as a strong first pass, not a final publish-ready document - it saves as Draft specifically so you review before it goes live.
For multi-month requests, expect one saved period per month, each reviewable and editable on its own.

‍

Troubleshooting

Problem	Likely cause	What to try
Generation is taking a while	Normal - it's making several tool calls per period before writing anything	Wait a few minutes; multi-month batches take longer since each month is researched separately
Draft feels generic or misses what mattered this period	The agent only knows what your data plus Additional context tells it	Add specifics (a channel to focus on, a recent change, a goal) and regenerate
A number looks wrong	The agent only reports what a query returns - if the underlying data is off, the narrative will be too	Cross-check the same metric on a chart or table card on the report; if it's still off, the issue is upstream in the Data Model, not the narrative
I wanted to redirect it partway through	Not supported - the agent can't pause for input once running	Let it finish, then edit the result directly, or regenerate with more detail in Additional context

Was this article helpful?



Nice one!

Thanks a lot for your feedback! If you’d like a member of our support team to respond to you, please send a message here



Please try again

Oops! Something went wrong while submitting the form.



Nice one!

Thanks a lot for your feedback! If you’d like a member of our support team to respond to you, please send a message here



Please try again

Oops! Something went wrong while submitting the form.