Prospecting Agent Setup Guide

Documentation Index

Fetch the complete documentation index at: https://agents-docs.hockeystack.com/llms.txt

Use this file to discover all available pages before exploring further.

​

Before You Start

Make sure the following are in place before beginning:

• Your CRM (Salesforce or HubSpot) is connected and syncing
• A call recording tool is connected (Gong, etc)
• Website tracking is installed and collecting data
• You have admin access to your HockeyStack workspace

​

Step 1: Add Users

Where: Workspace Settings > Team > Team Members Invite your team members to the workspace first, so you can assign them to roles in the next step.

​

How to add users

1. In the Team tab of Workspace Settings, scroll to Team Members.
2. Click + Invite and add team members by email.
3. Assign each person to the appropriate role from Step 2.

Every rep who should receive prospecting tasks needs to be both invited to the workspace and assigned to a role.

​

Step 2: Create Roles

Where: Workspace Settings > Team > User Roles Roles define how your team is organized and which accounts each person owns. The Prospecting Agent uses roles to route tasks to the right reps.

​

How to set up roles

1. Click your workspace name in the top-left corner.
2. Select Workspace Settings, then go to the Team tab.
3. Under User Roles, click + Add Role.
4. For each role, fill in:
   • Role name — AE, SDR, etc
   • Ownership Field — the CRM field that determines which accounts belong to this role (for example, OwnerId)
   • Users — the team members you invited in Step 1
5. Click Save Roles.

Create a role for each team structure you have: AEs, SDRs, and CSMs each get their own role with the appropriate CRM ownership field.

The system only generates tasks for accounts owned by users in the motion’s assigned roles. Make sure every rep who should receive tasks is covered.

​

Step 3: Scoring

Where: Atlas > Scoring Scoring tells the agent which accounts to prioritize. You define the signals that matter, and the system scores accounts based on those signals.

​

How to configure scoring

1. Open the Scoring card in Atlas.
2. Create Scoring Categories. Each category is a signal used for intent scoring. Assign a weight to each:
   • Positive weights boost an account’s score (for example, pricing page visits, demo requests)
   • Negative weights lower a score (for example, closed-lost within 6 months, competitor domains)
3. Click + Create Category to add signals.
4. Optionally, add Custom Prompts to define rules for how the AI interprets signals. Use this to specify tone preferences, what to highlight or ignore, and any company-specific logic.
5. Save your configuration.

Start with 2 to 3 categories. You can add more once you see how scores are distributed across your pipeline.

​

Step 4: Target Account List (Sales Views)

Where: Companies > Sales Views Sales Views define which accounts are in scope for prospecting. Each motion you configure later points to one or more Sales Views.

​

How to create Sales Views

1. Navigate to the Companies page from the left navigation.
2. Create or select a Sales View that filters to your target accounts.
3. Use filters like deal stage, company properties, scoring thresholds, segment, or ownership to build each view.
4. Give each view a clear name, since you will reference it by name when setting up motions (for example, “Enterprise Target Accounts” or “SMB Outbound List”).

You can create multiple views for different motions.

Only accounts owned by a user in the motion’s assigned role will receive tasks. Accounts without a matching owner will be skipped.

​

Step 5: Personas

Where: Atlas > Personas Personas define the buyer roles you want to target. The agent uses personas to find the right contacts at each account and tailor messaging.

​

How to configure personas

Personas tab:

1. Open the Personas card in Atlas.
2. Click + Add Persona for each buyer type.
3. For each persona, define:
   • Name — a clear label like “Security Leadership” or “Technical Evaluator”
   • Department — the function this persona belongs to
   • Description — criteria the system uses to match contacts to this persona
4. Aim for 2 to 3 personas that represent your typical buying committee.

Stakeholder Mapping tab:

1. Switch to the Stakeholder Mapping tab.
2. Add one or more Company Views.
3. With this enabled, the system automatically discovers and maps contacts at target accounts to your defined personas.

Define your personas before configuring motions in Step 10. Motions reference personas directly.

​

Step 6: Signals (Research Agents)

Where: Atlas > Research Agents Research Agents are AI-powered signals the system evaluates for each company. They let you define custom research questions — the agent answers them for every account in scope and uses the results to inform task generation.

​

How to create a Research Agent

1. Navigate to Research Agents in Atlas.
2. Click + Add to create a new signal.
3. Fill in the following:
   • Scope — Company or Contact
   • Name — a clear label for the signal (for example, “Recent Sales Leader Hires”)
   • Prompt — the research question the AI answers for each company. Be specific about what you are looking for, any relevant time windows, and what a positive result looks like.
   • Output Format — how the AI returns its answer (for example, Freeform)
   • Importance — how much weight this signal carries in sorting (Low, Medium, or High)
   • Refresh Every (days) — how often the AI re-runs this signal. Leave blank for no expiration.
   • Context Agent — turn on to use the company’s pre-built context instead of live web research
   • Auto AI Research — turn on to automatically run this signal for companies in selected views
   • Company Views — which views this signal should run against
4. Click Run Now to trigger an immediate run, or let Auto AI Research handle it automatically.
5. Check the Distribution bar at the bottom to see how many companies returned a result.

​

Step 7: Business Context

Where: Atlas > Business Context Business Context gives the AI the information it needs to personalize outreach. Everything the agent generates draws from this.

​

How to configure business context

1. Open the Business Context card in Atlas.
2. In the Current Version tab, write a description of your company. Include:
   • What your company does and the products you sell
   • Who your target customers are
   • What makes you different from competitors
   • Common reasons buyers purchase from you
   • Any specific language or terminology you want the agent to use
3. Use the Documents tab to upload supporting materials like one-pagers, case studies, or competitive battle cards.
4. Click Save.

The more specific your business context, the more relevant the agent’s output will be.

​

Step 8: Email Examples

Where: Atlas > Email Examples Email Examples teach the agent your outreach style. The agent references these when generating email and LinkedIn tasks.

​

How to add email examples

1. Open the Email Examples card in Atlas.
2. Each example has three parts:
   • Instructions — the context for this email type (for example, “cold outbound to security leaders”)
   • Subject line — an example subject
   • Body — an example email that reflects your preferred tone and structure
3. Add examples for each scenario your team uses: cold outreach, follow-ups, re-engagement, event-based, competitive displacement, and so on.
4. Click Manage to add, edit, reorder, or remove examples.

Use real emails that have performed well.

​

Step 9: Products

Where: Atlas > Products If you sell multiple products, define them here so the agent can match outreach to the right product. Each motion maps to a specific product.

​

How to add products

1. In Atlas, open the Products section.
2. Click + Add Product.
3. Give each product a name and a description. The description should cover what the product does, who it is for, and why customers buy it. The agent reads this when generating outreach, so more detail leads to better results.
4. Click Save.

​

Step 10: Prospecting Task Generation

Where: Atlas > Prospecting Task Generation This is where everything comes together. Prospecting Task Generation combines your scoring, personas, business context, email examples, and target accounts into automated outbound tasks.

​

Overview

The configuration has three tabs:

​

Motions

Each motion is a distinct outreach strategy. Click + Add Motion to create one.

• Outreach Strategy Prompt — instructions for the AI describing how to approach outreach for this motion. Be specific about tone, angle, and context.
• Users — the team members who will receive tasks from this motion
• Personas — which buyer personas to target (from Step 5)
• Company Views — the target account list for this motion (from Step 4)
• Product — the product this motion is associated with (from Step 9)
• Task Types — which channels the agent can use: Email, LinkedIn Message, Call, Research

You can create multiple motions for different strategies. Use the Rank field to set priority when motions apply to the same account. Toggle each motion on or off with the switch next to its name.

​

Task Limits

Set a daily cap for each user role. Start conservatively and increase once reps are consistently completing their tasks.

​

Activity

The Activity tab shows the current state of your prospecting engine:

• Company Coverage — percentage of in-scope companies with active tasks
• Context Coverage — percentage of companies with generated context
• Active Tasks — total pending or scheduled tasks
• Acceptance Rate — percentage of tasks reps accept vs. dismiss
• Last Context Sync and Last Task Generation — timestamps for recent runs

Use Sync All Context to trigger a context sync. Use Generate All Tasks to manually trigger task generation.

Watch your Acceptance Rate after launch. If it drops below 70%, review your outreach strategy prompts, personas, or target account lists.

​

Setup Checklist

Once you have completed all steps, verify the following:

• [ ] Team members are invited to the workspace
• [ ] User roles are created with the correct CRM ownership fields and users assigned
• [ ] Scoring categories are defined with appropriate weights
• [ ] At least one Sales View is created with your target accounts
• [ ] 2 to 3 personas are defined and stakeholder mapping is enabled
• [ ] Signals (Research Agents) are configured and running
• [ ] Business context is populated with company details and supporting documents
• [ ] Email examples cover your key outreach scenarios
• [ ] Products are configured (if applicable)
• [ ] At least one prospecting motion is configured and turned on
• [ ] Task limits are set for each role
• [ ] Context sync and task generation have been triggered
• [ ] Acceptance rate is being monitored in the Activity tab

​

FAQ