FlowRunner
PricingContact
Theme
Start Free

Adobe Workfront

Project Management

Manage Adobe Workfront work at scale with AI agents. Agents search and update projects, tasks, and issues, look up users and documents, and create new work items from signals in other systems.

10 actions API key available
A campaign request row lands in the marketing intake sheet
Search Projects checks whether a project for this campaign already exists in the portfolio
Search Users resolves the requesting stakeholder and the traffic manager's user IDs by email
Create Project stands up the project with name, description, priority, and planned dates
Create Task seeds the standard workback schedule, brief, creative, review, launch, with assignees and dates
Agent verifies every task carries an assignee and the dates respect the workback logic
The traffic manager reviews the generated plan before the assignments notify half the creative team
The requester receives the project link and the first milestone date

What This Integration Enables

Workfront instances live or die on operational discipline: an enterprise work-management system where the data lags reality becomes an expensive place to store fiction. The discipline problem is a throughput problem, thousands of small updates nobody has time to make, and throughput is what agents are for. FlowRunner agents search and update projects, tasks, and issues across the instance, resolve users for assignment, create work from intake signals, and locate project documents with download URLs. The instance stops depending on everyone's diligence and starts depending on flows that do not forget. - Convert intake forms, CRM signals, and spreadsheet rows into projects with seeded task plans - Sync task status and percent complete from the tools where work happens - Sweep projects, tasks, and issues for status reporting and SLA watch - Resolve assignees by name or email before any work is routed - Find and fetch project documents for review and archival flows

Without FlowRunner

Intake dies in a queue Work requests wait days for someone to translate them into properly structured projects
Percent complete is fiction Status fields update when someone remembers, so portfolio reports describe last week
Duplicate requests spawn duplicate work The same campaign gets requested twice and two teams quietly build it in parallel

With FlowRunner

Requests become structured work in minutes Projects arrive with tasks, assignees, and workback dates already in place
Status mirrors the source systems Agents sync percent complete and status from where work actually happens
Duplicates get caught at the door Every intake is searched against existing projects and issues before anything is created

Use Case Scenarios

The intake sheet that staffs itself

Marketing requests accumulate in [Google Sheets](/integrations/google-sheets). On each new row, the agent runs Search Projects for a duplicate, then Create Project with the campaign name, priority, and planned dates, then Create Task for each phase of the standard workback plan, assignees resolved through Search Users. The traffic manager approves the plan, assignments go live, and the request-to-staffed gap shrinks from a week of queue time to the same afternoon.

Percent complete that tells the truth

When creative work moves in the production stack, the agent calls Update Task to set status and percent complete on the linked Workfront task. A daily sweep with Search Tasks then pulls anything overdue or stalled and posts a digest to [Slack](/integrations/slack) with owners named. The Monday portfolio review runs on data that was true Sunday night, and the PMO stops spending Mondays interviewing project managers about what the system should already know.

Bug intake with duplicate detection

A defect report arrives via [Gmail](/integrations/gmail). The agent runs Search Issues scoped to the product's project to check whether the defect is already logged, and Search Users to resolve the reporter. Genuine new issues get filed with context attached; duplicates get a pointer to the existing issue instead of a second copy. Search Documents pulls the current spec with its download URL so the assignee opens the issue with the reference material already in hand.

Human-in-Loop Highlight

Update Project can move a project's status, and in Workfront a status change is not a label, it is a broadcast. Flipping a project out of Current reshapes portfolio rollups, stops timesheet expectations, and tells every assigned contributor their work is paused or dead; flipping one to Complete asserts to the business that deliverables shipped. An agent inferring "this looks done" from percent complete alone will eventually complete a project whose final deliverable is sitting unapproved in legal review. So FlowRunner flows let agents update freely inside a live project, dates, descriptions, task percent complete, but any project-level status transition stops for the project owner with the evidence: "All 34 tasks at Complete, last document uploaded 9 days ago, no open issues. Mark GTM-Q3-Launch as Complete?" The owner, who knows what the system does not, answers in one click, and the decision is recorded with the evidence that prompted it. Task-level truth is automated; project-level truth is signed.

Agent processes routinely
Detects exception requiring judgment
Clear match Continues automatically
Ambiguous Routes to human via preferred channel
Human decides
Agent resumes with decision

Agent Capabilities

10 actions

Projects

4
  • Search Projects Searches projects with partial-match name filtering, status, and portfolio scoping. Returns core fields including status, priority, percent complete, and planned dates. The duplicate check and portfolio sweep primitive.
  • Get Project Retrieves one project by ID with the full detail set: status, priority, percent complete, planned and actual dates, owner, portfolio, and program.
  • Create Project Creates a project with a name plus optional description, status, priority, and planned start and completion dates. The intake-to-work write.
  • Update Project Updates only the fields provided on an existing project. The write FlowRunner gates when it changes project status.

Tasks

3
  • Search Tasks Searches tasks, typically scoped to a project, with filters for name, status, and assignee. Returns status, priority, percent complete, planned dates, and assignment.
  • Create Task Creates a task in a project with a name, plus optional description, priority, assignee, and planned dates. The workback-plan seeder.
  • Update Task Updates a task's name, description, status, priority, percent complete, assignee, or dates, changing only supplied fields. The sync verb that keeps Workfront honest.

Issues

1
  • Search Issues Searches issues, bugs, requests, and change orders, optionally scoped to a project and filtered by name and status. The duplicate detector for defect intake.

Users

1
  • Search Users Searches active users by name or email, returning contact details, title, role, and home group. Supplies the user IDs behind every assignment.

Documents

1
  • Search Documents Searches documents, typically scoped to a project, returning file name, extension, size, owning project or task, and a download URL where available. The retrieval step for review and archival flows.

Frequently Asked Questions

What can FlowRunner do with Adobe Workfront?

FlowRunner agents can run Search Projects, Get Project, and Create Project in Adobe Workfront, plus 7 more actions.

Does connecting Adobe Workfront to FlowRunner require OAuth?

No. Adobe Workfront connects to FlowRunner with an API key, no OAuth flow required.

Can Adobe Workfront trigger a FlowRunner workflow automatically?

Adobe Workfront doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with Adobe Workfront

$100 in credits. No card required. Connect in minutes.