Core Concepts
The building blocks of Scout. Here’s what you need to know.
Agents
AI assistants that take action across your tools. Give an agent a goal, it figures out the steps.
In Scout: Create in Studio → Agents. Connect your tools, write instructions, and let it run.
What makes them different:
| Chatbots | Scout Agents |
|---|---|
| Answer questions | Take action |
| Wait for instructions | Figure out the steps |
| Single conversation | Remember context across sessions |
| One tool | Connect to all your tools |
Example: “Research our top 5 competitors and summarize their pricing.”
The agent finds competitors, searches for pricing, handles incomplete data, and delivers a summary, all without step-by-step instructions.
Instructions
Natural language you write to guide an agent’s behavior. Instructions define what an agent should do, how it should act, and what to prioritize.
In Scout: Write instructions when creating or editing an agent in Studio → Agents.
What makes instructions effective:
- Be specific about the goal, not just the task
- Describe the desired output format
- Give context about who the agent is helping
- Explain how to handle edge cases or missing data
Vague: “Research competitors.”
Specific: “Research our top 5 SaaS competitors. For each, find their pricing page, free tier limits, and top 3 integrations. If pricing isn’t publicly listed, note that. Return results as a table with one row per competitor.”
The second version defines scope, sets expectations for incomplete data, and specifies the output format. That’s what gets you a useful result on the first run.
Instructions vs. prompts: Instructions are persistent. They shape every run of the agent. Think of them as a job description, not a one-time request.
Collections & Tables
Structured data storage for your agents. Think of it as building your AI’s personal library.
Collections are top-level containers. Tables store structured data within collections.
In Scout: Create in Collections. Define schemas with columns like text, number, select, datetime, relation.
What you can do:
- Store documents, customer data, product info
- Let agents search and retrieve information
- Sync from external sources (websites, Notion, Google Drive)
Example: Upload your product catalog. Agents can answer customer questions about specs, pricing, availability.
Drive
File storage for documents, images, and assets your agents need.
In Scout: Upload in Drive. Organize into folders. Agents read and write files.
Use cases:
- Upload PDFs for agents to analyze
- Store images for processing
- Share assets across workflows
Syncs
Import data from external sources automatically.
In Scout: Configure in Syncs. Set up once, stay updated.
Supported sources:
- Websites (crawl and scrape)
- Sitemap (bulk page import)
- Notion (pages and databases)
- Google Drive (docs, sheets, slides)
- Microsoft 365 (SharePoint, OneDrive)
- Laserfiche (enterprise documents)
How it works: Point Syncs at your source, map fields to your Collection tables, and we keep it updated.
Observability
See what your agents are doing. Every action logged, every decision traceable.
Activity Logs:
- When agents ran
- What they completed
- How long each step took
- Success or failure
Execution Traces:
- Which tools were called and when
- What decisions were made
- Why the agent chose an approach
Tool Usage:
- Reads: Documents, databases, APIs queried
- Writes: What was created, updated, sent
- External calls: Third-party services contacted
This is how you build trust. Not mystery. Visibility.
Workflows
Automated sequences triggered by events or schedules. No code required.
In Scout: Build in Studio → Workflows. Connect blocks to create logic flows.
Triggers:
- Webhooks from external services
- Schedules (cron)
- Events (agent completes, file uploaded)
- Native integrations (Calendar, Email, Slack)
Blocks (the building blocks):
- Action Blocks: API calls, data storage, notifications
- Agent Blocks: Run AI agents within workflows
- Condition Blocks: Branch logic (if/else)
- Transform Blocks: Modify and restructure data
Example: New lead from website → Research company → Enrich CRM record → Draft outreach → Notify sales rep.
Key Terms
| Term | What It Means |
|---|---|
| Agent | AI assistant that takes action across tools |
| Workflow | Automated sequence triggered by events or schedules |
| Collection | Top-level data container |
| Table | Structured data within a Collection |
| Database | Collections + Tables (your AI’s personal library) |
| Drive | File storage for documents and assets |
| Sync | Automatic data import from external sources |
| Instructions | Natural language “programming” for agents |
| Context | Data, files, and history agents work with |
Next Steps
- Quick Start — Build your first agent
- Workflows Overview — Automate with workflows
- Collections Guide — Store data for agents
- Integrations — Connect the tools your agents use
- Agent Templates — Start with presets