Migrate from Certainly to Salesforce Service Cloud

Certainly Flows define dialogue trees with conditional branches, user prompts, and handoff rules. Service Cloud has no native chatbot dialogue builder. We map each Certainly Flow to a Salesforce Flow decision tree with equivalent branching logic, and the handoff triggers map to Omni-Channel routing rules. Conditional branches that use slot-filling or entity extraction require manual recreation in Flow variables. We document the full dialogue tree structure in a written handoff so the customer's admin or a Salesforce partner can rebuild the conversation logic. Flow conditions referencing dynamic entity values may require custom Apex or a third-party NLU integration to replicate the original conditional depth.

Certainly stores trained NLU intent classifiers with associated training utterances. These do not export as a deployable model. We export intent names, descriptions, and all associated utterances as a structured dataset. Migration to Service Cloud requires retraining Einstein AI or integrating a third-party NLU provider (such as Google Dialogflow or IBM Watson) because Service Cloud does not ingest pre-trained NLU models from external platforms. We document the complete intent-utterance corpus so the customer's AI team can retrain the destination model. Without retraining, intent classification accuracy degrades significantly in the new environment, reducing bot confidence scores and increasing fallback rates.

Custom entities in Certainly store value lists, synonyms, and pattern-based slot types. We map entity names and their values to Salesforce custom fields. Regex-based entities map to Salesforce Validation Rules with REGEX formulas. Multi-value entities map to Multi-Select Picklist fields. Entity synonyms require manual curation in the destination NLU provider's synonym management interface. We preserve the entity hierarchy (parent-child relationships) in a written mapping document so the customer can recreate the entity structure in their chosen NLU tool.

Static response texts, rich message templates, and conditional response variations export cleanly from Certainly as structured text data. These map directly to Salesforce Knowledge Article bodies (if Knowledge is licensed) or to Flow text template resources. Rich message components (buttons, cards, quick replies) do not have a native Service Cloud equivalent and require manual recreation as Lightning Web Components or Flow screen components. We deliver response texts in a structured format so the customer's developers can rebuild rich message components as custom Lightning components in Service Cloud.

Certainly's Zendesk connector stores ticket creation triggers, field mappings, and handoff rules that are not exported by standard data export tools. We document the complete Zendesk integration configuration during discovery, including which Flows trigger ticket creation, which Certainly fields map to Zendesk ticket fields, and what handoff conditions apply. This documentation allows the customer's admin to rebuild the connector in Service Cloud using the Salesforce-Zendesk native sync, MuleSoft Composer, or a custom Apex integration. We do not migrate the connector configuration as deployable code.

Historical chat transcripts export from Certainly as structured data including session IDs, timestamps, agent attribution, customer identifiers, message content, resolution status, and bot-versus-human attribution. Service Cloud does not have a native conversation log or transcript object. We archive conversation logs as timestamped CSV and JSON files, organized by date range and channel, and deliver them alongside the migration. The customer can import these into a custom object or a data warehouse for reporting purposes. We do not attempt a 1:1 object mapping for logs since no standard Service Cloud object accommodates the full transcript schema.

Bot-level settings including fallback behavior, greeting messages, session timeouts, and channel enable/disable flags store as configuration in Certainly. These do not export as structured records. We document all bot configuration values in a written configuration inventory so the customer's admin can set equivalent values in Service Cloud setup menus and Flow default variables. Some bot settings (such as conversation start triggers and session persistence logic) require custom Apex or Flow configuration to replicate.

Agent IDs, team assignments, and routing rule targets in Certainly must map to corresponding Salesforce User records and Omni-Channel Presence configurations. We extract agent assignments from routing rules and map them by email match to Salesforce Users. Skills and capacities defined in Certainly routing rules map to Omni-Channel Skills and Capacities in Service Cloud. Queues defined in Certainly routing map to Salesforce Queues. Any agent or queue without a Salesforce equivalent goes to a reconciliation queue for the customer's admin to provision.

Bot performance metrics including intent accuracy scores, conversation volumes, fallback rates, resolution times, and channel-specific KPIs export from Certainly as analytics snapshots. Service Cloud reporting does not ingest external analytics data natively. We archive all analytics data as CSV and JSON files organized by date range and metric type. The customer can import these into Salesforce Reports and Dashboards by creating custom report types on a custom analytics object, or use an external BI tool such as Tableau for long-term trend analysis.

Certainly integrations with external systems via webhook URLs, OAuth credentials, and API endpoints require documentation and manual recreation in Service Cloud. We document the full integration surface (endpoint URLs, authentication methods, payload schemas, trigger conditions) during discovery. The customer's admin or integration developer rebuilds these in Service Cloud using Named Credentials, Apex REST callouts, Platform Events, or MuleSoft Composer depending on the integration complexity.

Tags applied to conversations in Certainly for categorization and filtering migrate to a custom multi-select picklist field on the Case object. If the customer used tags for routing or SLA assignment, these map to Case Origin, Case Reason, or Omni-Channel Skill assignments in Service Cloud. We flag any tag-based routing logic that requires manual configuration of Omni-Channel routing attributes.

Certainly's fallback policies determine bot behavior when intent classification confidence is low or a user request cannot be matched. These map to Salesforce Flow default paths and Case Assignment Rules in Service Cloud. Unhandled intent scenarios that route to a human agent in Certainly map to Omni-Channel Work Item routing to the appropriate queue. We document the fallback thresholds and routing logic so the customer's admin can configure equivalent escalation behavior in Service Cloud Omni-Channel.