Documentation

How HR teams use Harriet for day-to-day employee support and admin tasks.

Create tickets, follow threads, and keep employee requests organized.

Structure the Knowledge Base around employee questions, owners, and plain language so Harriet cites the right material.

Publish separate knowledge documents with the right groups and geography metadata for each audience.

Give managers who need global context membership in a narrow group that can see reference documents without opening them to all staff.

Personalisation comes from HRIS-backed attributes and roles—balance helpful defaults with least data exposure.

Open a ticket for human follow-up, escalations, or anything that needs a tracked thread beyond a chat answer.

Review tickets, chat themes, and search misses to decide what to publish next in the Knowledge Base.

A step-by-step introduction to building a real workflow—using new starter onboarding to cover triggers, authoring agent steps, assigning skills, reaching out to stakeholders, and pausing for replies.

Keep workforce questions scoped, aggregated, and aligned with what managers are allowed to know about their teams.

Harriet workflows are configured by your team—there are no locked-in templates, but common patterns for onboarding, leave, and data collection are straightforward to set up.

Workflows can be started by a new ticket, a chat request, a scheduled time, or a webhook from an external system—choose the trigger that fits your process.

Workflows automate multi-step tasks with triggers, agent steps, and optional human checkpoints—separate from ad-hoc chat.

Some workflow actions pause until an approver confirms them—often on the linked ticket—then the run continues automatically.

A ticket can trigger workflows, host approvals, and collect replies that resume a paused run—without duplicate noise loops.

Running a workflow from chat may create a support ticket by default; many other triggers reuse an existing ticket or run without one.

Workflows can pause for approvals, ticket replies, external replies, scheduled time, or sub-tasks—each resumes automatically when its condition is met.

Combine clear questions, ticket threading, and explicit resume behavior so pauses feel predictable to employees and HR.

Start with one configurable agent step when plain-language instructions and skills are enough; open Advanced mode (canvas editor) when you need multiple separate automated phases or branching. Delegation runs one conversation at a time.

The configurable agent step runs an AI agent inside a workflow—you write plain-language instructions, assign skills, and the agent completes the task, pausing where it needs human input.

Use the data-gathering agent step to have Harriet reach out by Slack DM or email, ask follow-up questions, and return a structured summary when the conversation is complete.

A workflow agent can hand off a bounded sub-task to a specialist sub-agent, wait for it to complete, and use the results to continue—enabling sophisticated multi-step automation without losing oversight.

Scheduled workflows run automatically at set times—useful for regular reports, proactive reminders, and recurring compliance checks.

Workflow agents maintain full conversation history within a run, so context from earlier steps is always available when the workflow resumes after any pause.

Complex workflows combine a trigger, one or more agent steps with skills, pause mechanisms for human input, and clear error paths—designed in layers and tested before going live.

How Harriet connects to your HR and productivity systems safely—including MCP connectors managed alongside other integrations.

Control who can manage knowledge, workflows, integrations, and sensitive data.

Step-by-step guide to Okta SSO (OIDC) and SCIM user provisioning in Harriet, including group push for roles.

How optional skills extend Harriet, how MCP connectors are governed, and how Endpoint AI review fits in for some organizations.

Use user groups on knowledge documents so only the right people see them in search.

Choose the right metadata on knowledge documents for access control versus filtering and discovery.

Combine group membership with geography and structure so local answers rank first but wider context stays available.

Split versions by audience and geography, write clear titles, and avoid one giant document that mixes jurisdictions.

Directory and HRIS integrations supply employee data Harriet can use for answers and automations, within your plan and permissions.

Map vendor-specific HRIS or webhook fields to Harriet so syncs and automations use the right attributes.

Who can run workforce analytics in Harriet depends on roles, features, and any metadata scoping your admins configure.

Step-by-step expectations for adding a new MCP connector in Company settings, syncing tools, and Endpoint AI review when enabled.

Example prompts for workforce analytics and what you need before they work reliably.

Two different ways GitHub and package registries show up in Harriet—sandboxed MCP connectors from npm-style commands vs GitHub-synced skill packages in Endpoint AI.

MCP connectors let Harriet call approved third-party tools through governed skills—create them with New connector in Company settings, sync tools, and control who can use them.

MCP tools are exposed to workflow agents through skills—IT controls which tools are visible, which require confirmation before executing, and which may be used by sub-agents.

MCP servers may use org secrets, per-user OAuth, or API keys—know what Harriet retains, how the connector wizard maps to these patterns, and how approvals work in workflows.

Prefer a workflow with one configurable agent step unless teams need multiple canvas steps, branching, or strict separation of phases. Delegation is always serial—no parallel sub-agents.

Calendar features depend on which calendar integration is enabled and which user has connected their calendar to Harriet.

Agent steps are the core of powerful workflows—choose skills carefully, configure delegation correctly, and always set a sub-agent timeout to avoid stuck runs.

Sub-agents are isolated by design—they only see explicitly delegated skills, cannot access parent workflow data, and cannot delegate further.

The connected user’s calendar scopes and sharing settings determine list, read, and create access.

When expenses are enabled, Harriet can help employees submit or check claims through your connected expense system, within that system’s rules.

Upload Office formats to the Knowledge Base, use them in automations where supported, and use template-based document generation—not full desktop editing in the assistant.

Dedicated public channels give Harriet a clear context so SSO, policies, and docs match the product or team the channel is for—not a generic default.

The embeddable Harriet widget lets visitors chat from your marketing or help sites with scoped knowledge, skills, and ticketing—without signing into Harriet.

A practical sequence for enabling Harriet across chat platforms, embeds, and optionally Endpoint AI–managed desktops.

Recommended install order for Endpoint AI — secret enrollment link, Harriet Desktop download, link the device, then optional Claude Desktop for organizations that use it.

Endpoint AI is Harriet’s deployment and governance surface for consistent Harriet access on desktops and managed endpoints, including skills and connector review.

Align endpoint deployments with the same skill catalog and restrictions admins approve in Harriet, then provision approved skills to teams or users.

Checklist before enabling integrations, MCP connectors, or new skills—including Endpoint AI review for connectors and package skills when your organization uses it.

Use the built-in Skilify skill on your endpoint to author Anthropic-style skill packages, push them into Harriet’s review queue, then have admins provision approved skills to teams or users.

After skills are approved, use the Endpoint AI console to assign Harriet skills or profiles to teams or users—including Quick assign from the Harriet skills list.

When to use Claude Desktop alongside Harriet, and how the default desktop target setting affects new devices.