> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gradial.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Frequently Asked Questions

> Answers to the most common questions about Gradial — what it is, how it works, integrations, security, and getting your team up and running.

<AccordionGroup>
  <Accordion title="What is Gradial?">
    Gradial is the secure marketing harness that turns a fragmented lifecycle into governed execution — a layer that sits across your existing martech stack and executes the jobs between brief and live.

    Enterprise organizations already have the right tools: a CMS, a DAM, a ticketing system, design tools, email platforms. The problem is the workflow between them. Every handoff is manual, every tool creates a queue, and quality is inconsistent.

    Gradial fills that gap with a coordinated team of specialized AI agents that can author, update, tag, QA, schedule, and publish content end-to-end — across your existing systems, without replacing them, with brand, accessibility, and compliance checks applied to every output.

    [Learn more about what Gradial is →](/docs/getting-started/what-is-gradial)
  </Accordion>

  <Accordion title="What is an enterprise marketing harness?">
    An enterprise marketing harness is the scaffolding around a large language model that makes it actually work: the tools it can call, the integrations into your systems, the context it reasons over, the memory it carries between jobs, the workspace it operates in, and the guardrails that constrain it.

    The general concept is also called an *agentic harness* or *agent harness*. What makes one an **enterprise marketing** harness is that all of it is applied to the systems, skills, standards, and approval paths enterprise marketing runs on — which is why Gradial spans the whole lifecycle rather than covering one stage of it.

    [What is an enterprise marketing harness? →](/docs/getting-started/what-is-an-agentic-harness)
  </Accordion>

  <Accordion title="How is an enterprise marketing harness different from an AI point solution or a copilot?">
    A point solution gives you depth at one lifecycle stage, but nothing carries between stages — each tool is another contract, integration, and manual handoff. A copilot drafts and suggests inside one application, but a person still performs the work in the system of record.

    A harness executes across the whole lifecycle, inside your existing systems, under one set of governed controls. The practical test: can it change something in your CMS, against your standards, with an audit trail, without a person carrying the work across a seam?

    [Compare the approaches →](/docs/getting-started/what-is-an-agentic-harness#enterprise-marketing-harness-vs-other-approaches)
  </Accordion>

  <Accordion title="Is Gradial a chatbot or AI assistant?">
    No. Gradial is an execution platform, not a chat interface or copilot.

    While you can give Gradial tasks using natural language, the platform's agents actually complete the work end-to-end inside your connected systems — navigating your CMS, DAM, ticketing tools, and email platforms the same way a skilled team member would. The output is a published page, a tagged asset, a completed ticket, or a QA report — not a draft sitting in a text box for a human to copy and paste.
  </Accordion>

  <Accordion title="What kinds of work can Gradial handle?">
    Gradial's agents cover a broad range of marketing and content operations work:

    * **Content updates** — from a single copy change to bulk updates across hundreds of pages simultaneously
    * **New page creation** — from a brief, Figma design, Word doc, or existing page reference
    * **Email authoring** — building and publishing emails, variants and journeys in connected platforms including Marketo, Salesforce Marketing Cloud, HubSpot, Braze, Airship and MoEngage
    * **Asset management** — uploading, tagging, naming, applying metadata schemas, and retrieving assets across DAMs
    * **QA and accessibility** — brand accuracy, WCAG 2.2 AA compliance, SEO, and custom governance rules on every output
    * **Translation and localization** — multi-language and multi-region content execution at scale
    * **Campaign execution** — turning a campaign brief into a structured set of work across every channel it touches: web, email, social, paid media and commerce
    * **GEO analysis** — tracking how your brand appears in AI-generated answers and surfacing optimization opportunities

    [See the full scope of Gradial in Action →](/docs/gradial-prompt-library/customer-use-cases)
  </Accordion>

  <Accordion title="How is Gradial different from other AI content tools?">
    Four things set Gradial apart:

    1. **Execution, not just generation.** Gradial's agents don't hand you a draft — they complete work inside your actual systems.
    2. **Orchestration across your full stack.** A single task can touch your CMS, DAM, ticketing system, and email platform without manual handoffs between each.
    3. **Built-in governance.** QA, accessibility, brand, and compliance checks run automatically on every output — not sampled occasionally.
    4. **Multi-model by design.** Gradial selects the best AI model for each task and continuously evaluates output quality in production — so performance improves over time and you're never locked into a single vendor.

    [See what makes Gradial different →](/docs/getting-started/what-is-gradial#what-makes-gradial-different)
  </Accordion>
</AccordionGroup>

***

## Getting Started & Access

<AccordionGroup>
  <Accordion title="How do I get access to Gradial?">
    * **Existing customers:** Log in at [app.gradial.com](https://app.gradial.com). If you don't have an account yet, ask your Gradial org admin to invite you — see [Manage Team Members](/docs/integrations-administration-and-setup/managing-access/manage-team-members).
    * **Evaluating Gradial:** [Request a demo](https://www.gradial.com/request-demo). Demos are run against your actual CMS and content — not a generic sandbox — so you see agents working on your environment.
  </Accordion>

  <Accordion title="How long does it take to set up?">
    End-to-end onboarding typically takes **6–12 weeks**, depending on the complexity of your setup and the number of use cases in scope. The process follows four phases:

    1. **Environment Discovery** — structured sessions to map your tech stack, workflows, and use case priorities. You can start gathering inputs before any sessions are scheduled.
    2. **Credentials & Access** — connecting your CMS, ticketing system, SSO, DAM, and other integrations. The actual configuration is fast once credentials are in hand. This phase is almost always the longest because it depends on your IT team's internal security review and provisioning process, not on Gradial.
    3. **Platform Onboarding** — configuring golden pages, design patterns, QA rules, and running end-to-end validation of your use cases with your Gradial onboarding manager.
    4. **Training & Graduation** — Gradial trains your core team; your core team trains the wider user group. Onboarding concludes with a formal **Grady's Graduation** sign-off, confirming all use cases are validated, your team is trained, and named owners are in place.

    The single biggest lever on timeline is how quickly your organization can issue credentials. Submitting one consolidated request — CMS service account, ticketing API credentials, SSO configuration — rather than spreading asks across multiple tickets can take weeks off Phase 2.

    [Onboarding Checklist →](/docs/getting-started/onboarding-checklist)
  </Accordion>

  <Accordion title="Do I have to have an engineer help me set up Gradial?">
    You'll need one of your own engineers or MarTech architects to handle the integration setup — someone with access to your CMS, ticketing system, and other endpoints who can provision service accounts and configure permissions. This is an internal resource requirement, not a developer engagement.

    A Gradial engineer is **not normally required** for setup. The onboarding checklist and integration documentation are designed to be self-service, and your Gradial onboarding manager guides you through each phase. For complex environments or custom configurations, a Gradial engineer is available to support.

    Once integrations are live, day-to-day use requires no technical involvement. Content managers, editors, and communications leads run tasks directly — no engineering support needed.

    [Onboarding Checklist →](/docs/getting-started/onboarding-checklist) · [Integration Guides →](/docs/integrations-administration-and-setup/integrations/get-started-with-integrations)
  </Accordion>

  <Accordion title="Is there a test environment to experiment in?">
    Gradial does not provide a generic sandbox environment. Because every customer's CMS, DAM, and content structure is unique, environments are set up as part of your contracted onboarding — not as a shared demo instance.

    During Phase 3 of onboarding, your team sets up a dedicated sandbox or testing area within your own CMS so you can run agents and validate outputs safely before going live. Your Gradial onboarding manager guides you through this step.

    All tasks also go through a review step before anything is published, so nothing reaches your live site until a human approves it — providing a safety net regardless of environment.

    [Onboarding Checklist: Phase 3 →](/docs/getting-started/onboarding-checklist)
  </Accordion>

  <Accordion title="What do I need before starting setup?">
    Before technical setup begins, you'll want to gather:

    * A service account with appropriate CMS permissions (AEM, Sitecore, Contentful, or your connected platform)
    * API credentials for your ticketing system (Jira, Workfront, Wrike, or Asana)
    * SSO configuration details for your identity provider (Entra ID, Okta, or custom OIDC)
    * Example tickets that represent your team's most common task types
    * Any brand guidelines, QA checklists, or content standards documents
    * Candidate pages for golden page configuration (examples of ideal page structures)

    [Phase 1 of the onboarding checklist walks through this in detail →](/docs/getting-started/onboarding-checklist)
  </Accordion>

  <Accordion title="How do I add or remove team members?">
    Go to **Settings → Team Members** in the Gradial app. You can invite users by email, assign org-level and workspace-level roles, and deactivate users who no longer need access.

    [Manage Team Members →](/docs/integrations-administration-and-setup/managing-access/manage-team-members)
  </Accordion>

  <Accordion title="Does Gradial support Single Sign-On (SSO)?">
    Yes. Gradial supports:

    * Microsoft Entra ID (Azure AD)
    * Okta (OIDC)
    * Custom OIDC providers

    SSO is configured by your IT or platform admin during initial setup.

    [Configure SSO →](/docs/integrations-administration-and-setup/managing-access/sso)
  </Accordion>

  <Accordion title="Who can work in the platform once it's configured?">
    Anyone on your team can be given access — Gradial is designed for content managers, authors, editors, and communications leads, not just technical users.

    Permissions are structured at two levels:

    * **Organization roles** — control platform-wide access (Owner, Admin, Member, Viewer)
    * **Workspace roles** — control what a user can do inside a specific workspace (Owner, Admin, Member, Viewer)

    Users with **Org Member** or **Workspace Member** roles can create, edit, and run tasks — which covers the full day-to-day content workflow. Optional additional access — **Brand Manager** (manage rules) and **Designer** (manage design systems, personas, and audiences) — can be added on top of a user's organization role.

    Org Admins and Workspace Admins handle configuration and permissions; everyone else just runs work.

    [Roles & Permissions →](/docs/integrations-administration-and-setup/managing-access/roles)
  </Accordion>

  <Accordion title="Is the platform easy to use?">
    Yes — Gradial is designed to be very easy to use. You describe what you want done in plain language, and the agent handles the rest. No CMS navigation, no manual handoffs between tools, no technical knowledge required to run tasks.

    Two things make it especially approachable:

    * **Natural language input** — describe a task the way you'd describe it to a colleague. "Update the hero headline on this page to reflect our summer campaign" is a complete task brief.
    * **Grady** — the in-app AI assistant is available throughout the platform to answer questions in chat, explain features, and help you get more out of any workflow without leaving what you're doing.

    Training is covered as part of onboarding (Phase 4), and most users are comfortable running standard content updates within their first session.

    [Navigating the Platform →](/docs/getting-started/navigating-the-platform) · [Prompting an AI Agent →](/docs/getting-started/ask-grady)
  </Accordion>
</AccordionGroup>

***

## How Agents Work

<AccordionGroup>
  <Accordion title="What actually happens when I create a task?">
    When you create a task in Gradial (from the interface, a ticket, or a brief), here is what happens:

    1. An agent reads the task description and identifies the target content and systems involved.
    2. The agent retrieves relevant context — assets from the DAM, brand guidelines, templates, and content patterns for the relevant page type.
    3. The agent executes the work — authoring, updating, tagging, or building — inside your connected systems.
    4. Governance runs automatically: QA, accessibility (WCAG 2.2 AA), brand, and compliance checks are applied before the output reaches a human reviewer.
    5. Your team reviews the output — with full control over how much oversight is applied, from full review to auto-approve for trusted task types.
    6. The task is completed, status is updated in connected systems, and an audit trail is created.

    For bulk operations, this cycle runs in parallel across all items simultaneously.

    [How Gradial Works →](/docs/getting-started/how-gradial-works)
  </Accordion>

  <Accordion title="What kinds of work can I hand to Gradial?">
    You don't choose an agent or a mode — you describe the job and Grady does it, drawing on the right expertise and the right connected systems. The jobs teams hand to Gradial most:

    | Job | What it covers |
    | - | - |
    | **Update content that's already live** | Changes across web pages, from a single field to bulk operations across a whole site |
    | **Build something new** | Net-new pages from a reference, brief, Word doc, Figma file, or prompt |
    | **Make assets findable and usable** | Ingest, tag, name, govern and retrieve assets across connected DAMs |
    | **Check the work before anyone sees it** | Brand, accessibility, compliance and custom quality checks on every output |
    | **Build and send the campaign** | Emails, templates, variants, journeys and audiences in connected marketing platforms |
    | **Turn a brief into executed work** | One campaign brief becomes every deliverable it implies, across every channel |
    | **Localize for every market** | Multi-language and multi-region execution across pages, assets and video |
    | **Find out what's working, then fix it** | AI-search visibility, CX audits, SEO and analytics — and the content changes that act on them |
    | **Publish to channels** | Organic social, paid campaign structures, and commerce and product content |

    For jobs that span several of these, Gradial sequences the work automatically — running in parallel what can run in parallel, holding what has to wait, and passing outputs between steps without a handoff.

    [The jobs Gradial does →](/docs/getting-started/how-gradial-works) · [Customer Use Cases →](/docs/gradial-prompt-library/customer-use-cases)
  </Accordion>

  <Accordion title="Can Gradial handle bulk operations?">
    Yes. Bulk operations are one of Gradial's core strengths. A single task can run across hundreds of pages simultaneously — with the same QA and governance checks applied to every item in the batch.

    Common bulk operations include find-and-replace across a site section, applying a metadata schema update to a large asset library, running accessibility checks across all pages in a locale, and creating localized variants of a campaign landing page across multiple markets.

    [Rework Sections and Reusable Content →](/docs/updating-existing-experiences-i-e-content-updates/medium-content-updates-in-gradial)
  </Accordion>

  <Accordion title="How much human review is required?">
    It depends on how your team configures oversight — and how far along you are in the learning curve.

    Teams typically follow this progression:

    * **Early:** Agents execute with frequent human review. The platform is calibrating to your standards.
    * **Established:** Agents handle a wider range of tasks with lighter oversight. Human review shifts toward edge cases.
    * **Advanced:** Agents operate autonomously on routine jobs. Teams focus on directing outcomes, not reviewing execution.

    For any task type, you can configure full review, spot-check, or auto-approve. The level of autonomy is always in your control.
  </Accordion>

  <Accordion title="How do I get better results from agents?">
    The single biggest lever is the quality of your task description. A clear objective, the specific change required, and one concrete example produces dramatically better output than a vague request.

    Things that help:

    * Describe the *goal*, not just the action (e.g., "Update the hero headline to reflect our Q3 campaign theme — concise, action-oriented, under 8 words" rather than "Change the headline")
    * Reference a specific source URL for content updates
    * Attach relevant brand guidelines or approved copy as context
    * Configure golden pages so agents have a known-good structural reference for new page creation

    [Prompting an AI Agent →](/docs/getting-started/ask-grady)
  </Accordion>

  <Accordion title="What is Grady?">
    Grady is Gradial's in-app AI assistant, available throughout the platform to help you get more out of the product. You can ask Grady questions about how features work, get guided help creating tasks, or ask for context on any part of the Gradial workflow — directly inside the app, without leaving what you're doing.
  </Accordion>
</AccordionGroup>

***

## Integrations

<AccordionGroup>
  <Accordion title="What platforms and tools does Gradial support?">
    Gradial supports a broad and growing range of endpoints — CMS platforms, DAMs, ticketing systems, email and marketing automation tools, design tools, analytics platforms, and more. The integration ecosystem is always evolving as new connectors are added.

    For the current full list of supported integrations and setup guides, see the integrations documentation.

    [Integrations Overview →](/docs/integrations-administration-and-setup/integrations/get-started-with-integrations)
  </Accordion>

  <Accordion title="Can Gradial connect to internal systems or custom tools?">
    Yes. Gradial supports **native MCP (Model Context Protocol) integration**, which allows teams to connect proprietary knowledge bases, internal AI context layers, or custom workflows directly.

    This means you can:

    * Connect internal systems that don't have a standard API
    * Add proprietary brand or product knowledge as AI context
    * Kick off Gradial tasks from Claude, Slack, or other MCP-connected tools without opening the Gradial interface

    [Custom MCP Integration →](/docs/integrations-administration-and-setup/integrations/custom-mcp)
  </Accordion>

  <Accordion title="Do I need to replace my existing tools to use Gradial?">
    No. Gradial is designed as an overlay that works *across* your existing stack — not a replacement for it. Your CMS, DAM, ticketing system, and email platform stay in place. Gradial provides the execution layer that connects them.
  </Accordion>
</AccordionGroup>

***

## Quality & Governance

<AccordionGroup>
  <Accordion title="How does Gradial handle content quality?">
    Every output Gradial produces is automatically checked against your organization's defined quality standards before it reaches a human reviewer. This includes:

    * **Brand accuracy** — tone, voice, terminology, and visual standards
    * **Accessibility** — WCAG 2.2 AA compliance on every page and component
    * **SEO** — metadata, structure, and content quality rules
    * **Regulatory and compliance** — custom legal, industry, or internal compliance requirements
    * **Custom QA rules** — anything specific to your organization's standards

    These checks run on 100% of outputs — not sampled occasionally.

    [QA Reports →](/docs/quality-assurance-and-accessibility/qa-report) · [Accessibility Reports →](/docs/quality-assurance-and-accessibility/accessibility-report)
  </Accordion>

  <Accordion title="How does Gradial learn my organization's brand standards?">
    Gradial builds organizational context across four areas over time:

    * **Brand voice and editorial standards** — from brand guidelines, tone documents, and editorial standards you bring in
    * **QA thresholds** — your organization's specific accessibility, regulatory, and brand accuracy criteria
    * **Content patterns** — how your pages are structured, how assets are named, how campaigns are briefed
    * **Workflow and team preferences** — review levels, approval routing, publishing automation

    As agents complete more tasks and your team reviews outputs, the platform develops a practical working understanding of your standards that goes beyond what a style guide alone can capture.
  </Accordion>

  <Accordion title="What is a golden page?">
    A golden page is an exemplary page in your CMS that agents reference as a structural and quality standard when creating new pages or templates. By designating golden pages, you give agents a known-good reference — ensuring new pages match the quality, structure, and component patterns already approved for your site.

    [Golden Pages →](/docs/creating-new-experiences/golden-pages)
  </Accordion>

  <Accordion title="Can I set custom QA rules?">
    Yes. Gradial's Rules & Skills Manager lets you define organization-specific quality rules — custom brand checks, terminology requirements, legal disclaimers, regulatory compliance rules, or any criteria your team needs to enforce consistently.

    [Rules & Skills →](/docs/marketing-brain/managing-skills)
  </Accordion>
</AccordionGroup>

***

## Security & Data

<AccordionGroup>
  <Accordion title="Where does Gradial store content?">
    Gradial does not store your content permanently. It reads from and writes to your connected systems — your CMS, DAM, and email platform — which remain your systems of record. Task inputs and outputs are processed in transit; Gradial does not build a separate content repository.
  </Accordion>

  <Accordion title="What permissions does Gradial need in my CMS?">
    Gradial requires a service account with read/write permissions scoped to the content paths agents are expected to work in. For AEM specifically, this means JCR path-level permissions, which are documented in detail in the integration guide.

    Gradial follows the principle of least privilege — service accounts are scoped to only what is required.

    [AEM Integration: Permissions →](/docs/integrations-administration-and-setup/integrations/aem/aem-on-prem) · [AEM Cloud →](/docs/integrations-administration-and-setup/integrations/aem/aem-cloud)
  </Accordion>

  <Accordion title="What IP addresses does Gradial use? Do I need to allowlist them?">
    If your organization's network policy requires IP allowlisting for outbound connections from Gradial to your internal systems, consult the IP allowlist documentation for the current list of Gradial IP ranges.

    [IP Allowlist →](/docs/integrations-administration-and-setup/ip-allowlist)
  </Accordion>

  <Accordion title="Is Gradial SOC 2 compliant?">
    For detailed security documentation, compliance certifications, and data processing agreements, visit [gradial.com/security](https://www.gradial.com/security) or contact your Gradial account team via the [support portal](https://portal.usepylon.com/gradial).
  </Accordion>
</AccordionGroup>

***

## Troubleshooting & Support

<AccordionGroup>
  <Accordion title="A task didn't produce the result I expected. What should I do?">
    Before re-running the task, check the prompt first. The most common cause of unexpected output is an underspecified task description. Add:

    * The specific change you want (not just "update this page")
    * The source URL or content item
    * Any brand or style context the agent needs
    * One example of the output format or tone you want

    If the task description is clear and results are still inconsistent, check whether the relevant golden pages or QA rules are configured — agents produce better outputs when they have structural references to work from.

    [Prompting Guide →](/docs/getting-started/ask-grady)
  </Accordion>

  <Accordion title="My integration isn't working. Where do I start?">
    Start by verifying:

    1. The service account credentials are still valid and haven't expired
    2. The service account has the correct permissions on the relevant paths
    3. Your organization's network/firewall allows outbound connections from Gradial's IP ranges
    4. The integration is listed as active in **Settings → Integrations**

    For AEM specifically, OAuth tokens expire and need to be refreshed — this is the most common cause of intermittent failures.

    [AEM Troubleshooting →](/docs/integrations-administration-and-setup/integrations/aem/aem-overview) · [IP Allowlist →](/docs/integrations-administration-and-setup/ip-allowlist)
  </Accordion>

  <Accordion title="Security & Privacy">
    Gradial's security and privacy practices are documented at:

    * [gradial.com/security](https://www.gradial.com/security) — security architecture, certifications, and data handling practices
    * [gradial.com/privacy](https://www.gradial.com/privacy) — privacy policy and data processing information

    For security review documentation, compliance reports, or data processing agreements, contact your Gradial account team or reach out via the [support portal](https://portal.usepylon.com/gradial).
  </Accordion>

  <Accordion title="How do I get help from the Gradial team?">
    * **Support portal:** [portal.usepylon.com/gradial](https://portal.usepylon.com/gradial) — submit tickets and track status
    * **In-app:** Use the **Support** link in the Gradial app to open a ticket directly
    * **Grady:** Ask Grady (the in-app AI assistant) for help with platform questions without leaving your current task

    [Support Portal →](/support-portal)
  </Accordion>

  <Accordion title="Where can I see what's been released recently?">
    The release notes page is updated with each major release — new features, improvements, and changes.

    [Release Notes →](/release-notes/overview)
  </Accordion>
</AccordionGroup>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.