Digital Workers

Overview

Digital Workers in SalesOS are automated agents that execute repetitive CRM tasks on your behalf. They operate around the clock, handling data entry, follow-up sequences, record enrichment, notifications, and reporting without manual intervention. By offloading routine work to Digital Workers, your sales team can focus on relationship-building and high-value activities that require human judgment.

Unlike generic workflow automation, Digital Workers in SalesOS are purpose-built for CRM operations. They understand your data model, respect permission boundaries, and integrate natively with every module in the platform.

Deterministic Bots vs AI Agents

SalesOS offers two categories of Digital Workers, each suited to different use cases:

Rule-Based Bots

Rule-based bots follow deterministic logic: if a condition is met, they execute a predefined action. They are predictable, auditable, and ideal for tasks with clear business rules.

AI-Powered Agents

AI agents leverage machine learning to make decisions in ambiguous situations. They analyze patterns, classify intent, and take contextual actions that go beyond simple rule matching.

Most organizations use a combination of both types, with rule-based bots handling structured tasks and AI agents tackling unstructured or judgment-heavy operations.

Types of Digital Workers

Data Hygiene Workers

Data hygiene workers continuously scan your CRM records for inconsistencies, duplicates, missing fields, and formatting issues. They can automatically standardize phone numbers, normalize company names, merge duplicate contacts, and flag incomplete records for review.

Follow-Up Workers

Follow-up workers monitor deal stages, activity timelines, and engagement signals to ensure no prospect falls through the cracks. They can send templated emails, create follow-up tasks, or alert reps when a contact has gone silent for a configurable period.

Enrichment Workers

Enrichment workers augment your records with external data. When a new lead or account is created, they pull firmographic data, technographic signals, social profiles, and news mentions from integrated providers to give reps a complete picture before outreach.

Notification Workers

Notification workers monitor events across your CRM and deliver real-time alerts to the right people. They handle deal stage changes, large deal alerts, renewal reminders, quota attainment milestones, and escalation triggers.

Reporting Workers

Reporting workers generate and distribute reports on a schedule. They compile pipeline snapshots, activity summaries, forecast updates, and performance scorecards, then deliver them via email, Slack, or the in-app notification center.

Creating a Digital Worker

To create a new Digital Worker, navigate to Settings > Automation > Digital Workers and select Create Worker.

Step 1: Define the Trigger

The trigger determines when the worker activates. Available trigger types include:

• Record Created — fires when a new record is created in a specified object
• Record Updated — fires when a field on an existing record changes
• Time-Based — fires on a schedule (hourly, daily, weekly) or after a time delay
• Event-Based — fires when a system event occurs (email opened, form submitted, meeting booked)
• Threshold Crossed — fires when a metric crosses a defined boundary

Step 2: Set Conditions

Conditions filter which records or events the worker acts on. You can combine multiple conditions with AND/OR logic:

• Field equals / does not equal a value
• Field contains / starts with / ends with
• Field is empty / is not empty
• Record owner matches a team or role
• Time since last activity exceeds N days

Step 3: Configure Actions

Actions define what the worker does when triggered and conditions are met. You can chain multiple actions in sequence:

• Create a record (lead, task, activity, note)
• Update fields on the triggering record or a related record
• Send an email or SMS notification
• Assign the record to a user or queue
• Enrich the record from an external provider
• Add the contact to a sequence
• Post a message to a Slack channel
• Log an activity

Worker Templates

SalesOS ships with pre-built templates to accelerate setup:

Scheduling Options

Digital Workers support three scheduling modes:

Continuous (Real-Time)

The worker monitors events in real time and executes immediately when conditions are met. Ideal for notifications, assignments, and time-sensitive actions.

Interval-Based

The worker runs on a recurring schedule — every 15 minutes, hourly, daily, or weekly. Best for batch operations like data hygiene scans, report generation, and enrichment runs.

Event-Triggered with Delay

The worker waits for a specific event, then pauses for a configurable delay before acting. Useful for follow-up workflows where you want to give the prospect time to respond before escalating.

Worker Actions in Detail

Create and Update Records

Workers can create new records in any standard or custom object. They can also update fields on existing records, including setting values, incrementing counters, appending to multi-select fields, and clearing stale data.

Send Notifications

Workers can deliver notifications through multiple channels: in-app alerts, email, SMS, Slack messages, and Microsoft Teams posts. Each notification supports dynamic merge fields drawn from the triggering record.

Assign and Route

Workers can reassign record ownership based on round-robin logic, territory rules, capacity limits, or lead score thresholds. This ensures records always land with the right rep without manual intervention.

Enrich Data

Workers can call external enrichment providers (Clearbit, ZoomInfo, Apollo, or custom APIs) to populate missing fields. Enrichment actions respect rate limits and credit budgets configured at the org level.

Execute Sequences

Workers can add contacts to existing sales sequences, remove them from sequences, or advance them to the next step. This enables dynamic enrollment based on real-time behavior.

Monitoring Performance

The Digital Workers dashboard provides visibility into every worker's performance:

• Execution Count — total runs in the selected time period
• Success Rate — percentage of executions that completed without error
• Average Duration — time from trigger to action completion
• Records Affected — number of records created, updated, or enriched
• Error Rate — percentage of executions that failed or timed out

Filter the dashboard by worker type, status (active/paused/errored), and date range to identify underperforming workers quickly.

Worker Logs and Audit Trail

Every Digital Worker execution is logged with full detail:

• Timestamp of trigger event
• Record ID and object type that triggered the worker
• Conditions evaluated (pass/fail for each)
• Actions executed with input/output values
• Duration of each action step
• Final status (success, partial, failed, skipped)
• Error message and stack trace (if failed)

Logs are retained for 90 days by default. You can export logs to CSV or push them to an external logging platform via webhook.

The audit trail integrates with the SalesOS platform-wide audit log, so compliance teams can trace every automated change back to the worker configuration that caused it.

Capacity Limits

To protect system performance and prevent runaway automation, Digital Workers operate within capacity limits:

When a worker approaches its execution limit, it queues remaining triggers and processes them in the next available window. Admins receive alerts when workers are consistently hitting capacity.

Error Handling

Retry Logic

When an action fails (network timeout, API error, validation failure), the worker retries up to three times with exponential backoff (1 second, 5 seconds, 30 seconds). You can configure retry count and backoff intervals per worker.

Escalation

If all retries fail, the worker can escalate by:

• Pausing itself and alerting the admin
• Creating a task for manual intervention
• Logging the failure and continuing to the next record
• Falling back to an alternative action

Circuit Breaker

If a worker's error rate exceeds 50% over a 10-minute window, the system automatically pauses it and sends an alert. This prevents cascading failures and protects downstream systems from repeated bad requests.

Permissions and Security

Digital Workers execute with the permissions of the user who created them, unless explicitly configured with a service account. This means:

• Workers respect role-based access controls
• Workers cannot access records outside their creator's visibility
• Admins can assign workers to service accounts for elevated access
• All worker actions are attributed in the audit log

Best Practices

• Start simple. Begin with one or two high-impact workers (stale lead alerts, new lead enrichment) before building complex chains.
• Use descriptive names. Name workers by their function and scope (e.g., "Enterprise Lead Enrichment - APAC") so the team can identify them at a glance.
• Set conservative conditions initially. Over-broad conditions can cause unexpected mass updates. Test with narrow filters and expand once validated.
• Monitor weekly. Review the performance dashboard every week to catch rising error rates or declining execution counts before they become problems.
• Version your workers. When modifying a worker, clone it first and test the new version in parallel before deactivating the original.
• Document escalation paths. Every worker should have a clear owner and a defined escalation path for when things go wrong.
• Respect rate limits. If workers call external APIs, configure appropriate intervals to avoid exhausting provider credits or triggering throttling.
• Combine deterministic and AI workers thoughtfully. Use rule-based bots for clear-cut operations and reserve AI agents for tasks that genuinely benefit from contextual judgment.
• Clean up unused workers. Audit your worker list quarterly and deactivate any that are paused, erroring consistently, or no longer aligned with current processes.
• Test in sandbox first. Use the SalesOS sandbox environment to validate new workers against test data before deploying to production.