Create audiences for programs

In Adobe Marketo Optimizer, people lists define the audience for person journeys — either as dynamic filter-based lists that update automatically, or static lists with fixed membership. From the chat interface, the Audience Creation skill builds, adapts, and edits people lists through a guided conversation.

  • Skills - audience-creation and people-list-comparison
  • Invocation - Describe audience criteria directly, upload a Marketo Engage Smart List, or name an existing list to edit
  • Reads from / writes to - Marketo Optimizer; reads Marketo Engage when adapting Smart Lists

Supported workflows workflows

Coworker supports three audience creation workflows and determines which one applies from your request. If your intent is ambiguous, it asks before proceeding.

Workflow
When to use it
Example prompt
Create from scratch
You want a new people list defined by criteria or membership.
“Create a dynamic list of VPs of Marketing at SaaS companies in North America.”
Adapt a Marketo Engage Smart List
You have an existing Marketo Engage Smart List or Smart Campaign and want an equivalent people list.
“Adapt this Marketo smart list into a people list.” (attach the asset)
Edit an existing list
You want to add to or replace the rules on a list you already have.
“Add a rule to my ‘Enterprise Trials’ list for lead score over 50.”

Create a people list from scratch create-from-scratch

Before generating anything, Coworker confirms all four of the following. It asks for any that are missing — in a single message.

  1. Rules / criteria — A plain-language description of who belongs in the list.
  2. Name — What to call the list.
  3. Location — Which program the list should live in. Provide a program name and Coworker finds it; if there are multiple matches it asks you to pick.
  4. Type — Dynamic (filter-based, auto-updating) or static (fixed membership). This is required — Coworker will not guess; if you don’t specify, it asks.

Dynamic lists dynamic-lists

For dynamic lists, Coworker proactively suggests including personalization attributes to make targeting richer. These attributes are included by default — you opt out, not in:

Attribute
Why it helps
Derived Persona
AI-inferred buyer persona for persona-based content targeting.
Derived Intent
Inferred purchase-intent signals that surface in-market accounts.
Engagement Tier
Calculated engagement level that prioritizes engaged contacts.

Tell Coworker if you want to remove any of these before it proceeds.

Static lists static-lists

  • Static, no criteria — The list is created empty, ready for you to add members manually.
  • Static from criteria (a snapshot) — Coworker builds the matching set and copies those people in. Population is asynchronous — Coworker confirms the list was created but notes that it may take a few minutes for people to appear. It will not claim the list is ready immediately.

Review card review-card

Nothing is created until you approve it. After you describe your criteria, Coworker presents an interactive People List Creation Review card (for adaptations from Marketo Engage lists, the card is titled People List Conversion Review).

Each row in the card represents one condition:

Column
Meaning
Requirements (or the Marketo Engage list name for adaptations)
Your original request or the source Marketo filter.
(rule)
The actual attribute-based rule generated for that condition.
Include
A checkbox to keep or drop that rule.

Confidence levels:

  • High confidence rows are matched cleanly and checked by default.
  • Low confidence rows (approximate mappings or anything flagged) are shown with a notice indicator and are unchecked by default.
  • Rows the system could not map show “No equivalent found” — these have no rule and remain unchecked.

A Conversion Summary tallies N High Confidence and N Low Confidence, with a hint: low-confidence rules are unchecked by default; check them to include, or describe the change you want in chat.

Card actions:

  • Proceed — Creates the list using only the checked rules.
  • Describe the changes you want in chat — Pre-fills the input with "I want to change: " so you can refine; Coworker regenerates and shows a fresh card, keeping rules you had already approved.

You can also type a follow-up at any point (for example, “also restrict to companies over 500 employees”) and Coworker regenerates the card.

Attribute mapping attribute-mapping

When you describe criteria, Coworker translates each condition into a real, known person-level attribute. Three outcomes can appear on the Review card:

  1. Matched (high confidence) — Your condition maps directly to an attribute (for example, “email is acme.com” maps to the email attribute). Checked by default.
  2. Approximate (low confidence) — The closest available attribute differs in name or data model (for example, a Marketo Amount filter approximated as Lead Score). Shown with a note explaining the difference; unchecked by default.
  3. Not found — The condition couldn’t be mapped to any known attribute. Shown as “No equivalent found”; no rule is generated.

This is why a list you describe might come back with fewer rules than conditions you specified — unmatched conditions are surfaced explicitly rather than silently dropped. If important criteria land as “not found,” rephrase them using the attribute’s real name and Coworker re-tries.

NOTE
If you’re mapping spreadsheet columns to fields (a Field Mapping card with Source Column, Target Field, a confidence percentage, and an Unmapped Columns list), that is the lead import flow, not audience creation. See the import-leads skill.

Edit rules for an existing list edit-rules

When you ask to change rules on a list you already have, Coworker establishes which list and which edit mode:

  • Add / append (default for “add rules”, “add more rules”) — new rules are merged with the existing ones.
  • Replace (default for “replace rules”, “change rules to”) — new rules replace all existing rules on the list.

Coworker summarizes what will be applied and clearly states whether it’s adding or replacing, then asks you to confirm before it commits. After applying, it reports the total rule count and how many were added or replaced.

NOTE
Edits use a merge-aware path so an “add” operation never silently overwrites your existing rules.

Audience overlap overlap

Ask Coworker to compare two people lists (for example, “Show me the overlap between ‘Q3 Webinar’ and ‘Enterprise Accounts’”) and it renders a People List Overlap card:

  • A header badge showing the count: “{N} in common.”
  • A stats row with each list’s total member count and the overlap as “X% of A · Y% of B.”
  • A member table of the people in both lists, with a Name column and a second column you can steer — Email (default), Company, or Job Title depending on what you asked.
  • Click any name to open that person in the workspace.
  • If there is no overlap, the card says so clearly: “No members in common between these two lists.”

Limitations:

Limit
Detail
Table size
Shows up to 200 members; beyond that it notes “Showing 200 of N — ask me to refine the query to narrow results.”
Overlap computation
Computed on email address; people without an email are excluded from the intersection.
List size
Reads up to approximately the first ~1,000 members of each list. For larger lists Coworker tells you results are partial.
Draft dynamic lists
Can’t be compared — a list that hasn’t been published has no live segment. Coworker asks you to publish it first or use a static list instead.

QA validation qa-validation

After creating or updating a list, Coworker offers: “Would you like me to verify the list is correctly configured?” If you accept, it re-fetches the list and reports the following checks:

Check
Result
List found under the correct Program/Folder
Pass / fail
Filter count matches what was applied
N filters / mismatch
Personalization attributes present (if included)
Present / missing
List name matches what you requested
Pass / fail
Estimated member count
count or N/A

Limitations limitations

Limitation
Detail
Static-list adaptation from Marketo Engage
You can’t adapt a Marketo Engage static list (or an email or other non-filter asset) into a people list. Static lists are explicit member IDs and can’t be expressed as filters; Coworker asks for a Smart List or Smart Campaign instead.
Activity- and membership-based filters
When adapting from Marketo Engage, filters like Email Was Opened, Visited Web Page, Form Was Filled Out, Member of List, and Member of Smart Campaign have no people-list equivalent and come back as “No equivalent found.”
Company-level conditions
Translated to the nearest person-level attribute where possible (people lists operate on person attributes) and flagged low-confidence when the fit is loose.
Deeply nested AND/OR logic
Complex nested logic may collapse to a top-level AND/OR; Coworker notes this when it happens.
Name collisions
Not auto-resolved — if the name is taken, Coworker asks you for a different one rather than silently appending a suffix.
Approval required
Coworker will not create or modify a list until you click Proceed, confirm, or give a clear go-ahead (“approved”, “looks good”, “build it”).
Static snapshot population
Membership in static lists created from criteria fills in over a few minutes — not instantly.
recommendation-more-help
marketo-optimizer-help-user-guide