Getting Started
Getting Started
This guide walks you through signing up for Groundtruth, creating your first engagement, running it, and sharing the results with your client.
1. Sign Up
Navigate to the platform registration page and create your account:
- Company Name -- The name of your organization. This becomes your tenant.
- Your Name -- Your full name.
- Email -- Your work email address.
- Password -- Minimum 8 characters.
After registering, you are assigned the Owner role for your tenant. You can invite team members later from the Team Management settings.
2. Choose a Plan
Select the plan that fits your usage. You can change plans at any time from Billing settings.
| Plan | Price | Engagements | Features |
|---|---|---|---|
| Starter | $99/month | Up to 5 | Core platform, email notifications, API access |
| Professional | $299/month | Up to 25 | Everything in Starter + client portal, custom branding, priority support |
| Enterprise | Custom pricing | Unlimited | Everything in Professional + SSO/SAML, dedicated support, custom SLAs |
Payment is handled through Stripe. You will be redirected to a secure checkout page to enter your payment details.
3. Create Your First Engagement
From the dashboard, click New Engagement and fill in the details:
- Name -- A descriptive name for the engagement (e.g., "Q3 Go-to-Market Strategy").
- Client -- The client company name this engagement is for.
- Brief -- A detailed description of what you need. The more context you provide, the better the AI crew performs. Include goals, constraints, target audience, and any relevant background.
- Budget -- Maximum LLM spend for this engagement in USD. Defaults to $5.00. The engine will stop if the budget is exhausted.
Use a Template (Optional)
Instead of configuring from scratch, select one of the five system templates:
| Template | Best For |
|---|---|
brand_digital_presence | Brand strategy, digital channels, content planning |
business_transformation | Organizational change, process redesign, operating models |
go_to_market | Launch strategy, market analysis, positioning, pricing |
product_strategy | Product vision, roadmap, feature prioritization, user research |
research_deep_dive | Market research, competitive analysis, trend reports |
Templates pre-fill the task list with tested configurations. You can customize them after selection. See Templates for more on creating your own reusable templates.
Configure Tasks
Each engagement consists of tasks arranged in a DAG (directed acyclic graph). For each task, specify:
- Description -- What the task should accomplish.
- Expected Output -- The format and content you expect (e.g., "A 2,000-word market analysis report in markdown").
- Dependencies -- Which tasks must complete before this one starts. Tasks without dependencies run in parallel.
The engine assigns each task to the most appropriate AI agent based on department and specialization.
4. Run the Engagement
Click Start on the engagement detail page. The engine begins executing tasks according to the DAG:
- Activity Feed -- Watch task progress in real time via server-sent events (SSE). Each task logs its status as it moves through queued, running, and completed states.
- Cost Ticker -- Monitor LLM spend as it accumulates. The engagement pauses automatically if the budget limit is reached.
- DAG View -- See which tasks are running in parallel and which are waiting on dependencies.
A typical engagement takes 10 to 60 minutes depending on the number and complexity of tasks. You can pause and resume at any time. See Engagements for details on run controls.
5. Review Deliverables
Once tasks complete, their outputs appear in the Deliverables tab:
- Each deliverable is rendered as formatted markdown.
- Quality Scores -- The observer agent scores each deliverable on relevance, depth, clarity, and actionability.
- Version History -- If you re-run a task, previous versions are preserved. Use the diff viewer to compare versions side by side.
See Deliverables for the full versioning and comparison workflow.
6. Approve or Request Revisions
Each deliverable enters a Pending Review state. You have three options:
- Approve -- Mark the deliverable as final.
- Request Revision -- Add notes describing what needs to change. This triggers a targeted re-run of just that task with your feedback injected as context.
- Reject -- Reject the deliverable with an explanation.
The approval workflow ensures nothing is marked as final until you have reviewed it. See Deliverables for the full approval process.
7. Export as PDF
Generate a polished PDF report from any engagement:
- Full Engagement Report -- All deliverables compiled into a single branded document with a cover page, table of contents, and quality summary.
- Single Deliverable -- Export an individual deliverable as a standalone PDF.
PDFs include your tenant branding (logo and primary color) if configured in Team Management settings.
8. Share with Clients via Portal
Share deliverables with your clients without requiring them to create an account:
- Go to the engagement detail page and click Share.
- A unique portal link is generated with a secure token.
- Send the link to your client.
From the portal, your client can:
- View all approved deliverables.
- Leave comments on specific deliverables.
- Approve or request revisions (if you enable client review).
- Download deliverables as PDF.
The portal is read-only and branded with your company logo and colors. No login required. See Client Portal for configuration options.
Next Steps
- Engagements -- Advanced engagement configuration and run management.
- Templates -- Save and reuse engagement configurations.
- File Uploads -- Attach reference documents to provide additional context to the AI crew.
- API Reference -- Integrate Groundtruth into your CI/CD pipelines, Slack bots, or internal tools.
- Analytics -- Track engagement performance, spend, and quality trends over time.