AI-decisioning health
AI-decisioning health checks the data that powers personalization in Adobe Marketo Optimizer. It reports on lead coverage, persona classification, and story richness across demographic, firmographic, technographic, and psychographic categories. Then it flags missing data to identify where to start.
Use AI-decisioning health to see what data flows in from Marketo Engage and where gaps exist. Closing those gaps improves how AI decisioning scores and routes each person.
Open AI-decisioning health open
Open the report from either the home page or the Coworker chat.
- On the Home page, select the AI-decisioning health card in the Quick access row. The card leads the row and shows story volume and persona-classification progress, such as 929 stories, 32% persona classified.
- In the Coworker chat box, ask about your personalization data directly, or type
/and select AI-decisioning health.
Both paths open the report inside the Coworker workspace.
Chat welcome and follow-up prompts chat-welcome
Opening AI-decisioning health from chat displays a welcome message, *Welcome to AI-decisioning health, a summary of report checks, and a card to open the full report.
Below the card, under What would you like to do next?, AI-decisioning health suggests follow-up prompts based on your own data’s specific gaps. For example, if 67.7% of your leads lack a persona classification, one suggested prompt reads Why are 67.7% of leads unclassified by persona? Select a suggested prompt, or ask your own question, to get a direct answer without leaving chat.
Report overview report-overview
The workspace report opens with a Highlights callout that lists the strongest and weakest areas of your data in plain language, such as Demographic data reaches 100% of leads with strong field depth or 67.7% of leads remain unclassified into any persona. A checkmark marks a healthy result, and a slashed circle marks a gap.
Next to the highlights, a radar chart plots overall Coverage across six dimensions: demographic, firmographic, technographic, psychographic, persona, and intent. A larger shaded area means broader coverage.
Persona classification persona-classification
The Persona classification section shows how many of your stories are classified into a persona, for example: 300 of 929 stories classified · 32.3% classified · 67.7% unclassified. A stacked bar breaks classified stories by persona, with a legend showing the story count and percentage for each.
Select a persona segment to open a detail card with example job titles for that persona. For example, the Other segment might show: 272 stories / 29.3%, with examples such as Industry Specialist, Independent Advisor, Freelance Consultant, and Subject Matter Expert.
Coverage coverage
The Coverage section lists five data categories: demographic, firmographic, technographic, psychographic, and intent and activity. Each category shows the percentage of stories with at least one available attribute in that category.
Select a category to expand it, then choose one of two tabs:
- Attributes - Attributes grouped by type, such as Personal Details or Location under demographic. Each attribute shows how many stories have a value for it, for example:
firstName (906 stories). - Flags - Gaps specific to that category, or No open flags on this category when coverage is healthy.
Use the search field above the category list to jump directly to a category or attribute by name.
Flags flags
The Flags section at the bottom of the report lists every gap found across all categories, ranked by severity:
- Critical - Gaps that block a capability outright, such as Technographic coverage is 0% across all leads.
- Watch - Gaps that reduce effectiveness but don’t block a capability, such as Psychographic coverage reaches only 7.2% of leads.
Filter the list by severity, then select a flag to expand it and read a one-sentence explanation of its business impact, for example: Unclassified leads cannot enter persona-specific journeys or receive role-tailored messaging, reducing campaign relevance and conversion rates.
Recently accessed recently-accessed
If you open AI-decisioning health and then navigate away, it reappears under Recently accessed in the empty workspace, so you can jump back into the report without returning to the home page.
Planned enhancements for AI-decisioning health include:
- A dedicated entry in the Coworker skills catalog.
- Guided “ask how” actions that walk you through fixing a flag.
- A dedicated next steps tab.