User Guide

AI Sales Assistant — User Guide

AI Sales Assistant is Xfatora’s AI-assisted calling and qualification module for first-contact conversations. It supports inbound and outbound calling workflows, campaign launch control, reusable assistant instructions, smart qualification, and post-call processing.

This module is designed to improve consistency in the first conversation layer. It does not replace CRM and Sales ownership, and it is not positioned as a full phone-system replacement.

Roles & permissions

Recommended ownership model:

  • AI Sales Admin
  • Configures: General Settings, Inbound Call Assistant Settings, Outbound Call Assistant Settings, Assistant Instructions, System Prompt Templates, Quick Start Templates, Knowledge Bases links, Smart Qualification defaults, and Post-Call Processing rules.
  • Launches: approves production-ready campaign settings and test readiness.
  • Reviews: quality of summaries, transcripts, qualification outcomes, and governance controls.
  • Should not change freely: qualification criteria and post-call mappings without RevOps alignment.
  • Campaign Manager
  • Configures: AI Campaigns, Campaign Setup fields, Upload CSV File input, Launch Schedule, Max Call Attempts, and campaign-specific instructions.
  • Launches: launch now / schedule later for approved campaigns.
  • Reviews: status, progress, completed/failed calls, and campaign-level quality.
  • Should not change freely: global prompt templates or organization-wide qualification rules.
  • Sales/RevOps Reviewer
  • Configures: qualification thresholds and post-call source/status mapping policy with admin.
  • Launches: usually no direct launch ownership.
  • Reviews: qualification discipline, lead handoff quality to CRM, success-rate and duration trends.
  • Should not change freely: live campaign behavior without change control.
  • Read-only Leadership Reviewer
  • Configures: none.
  • Launches: none.
  • Reviews: campaign statistics, quality trends, and operational consistency.
  • Should not change freely: prompts, qualification criteria, or post-call statuses.

Ownership recommendation:

  • Prompt standards: AI Sales Admin.
  • Qualification criteria: RevOps + AI Sales Admin.
  • Post-call source/status rules: RevOps + CRM owner.

Setup checklist

Use this checklist before production rollout:

  1. AI provider readiness
  • Confirm assistant service is available for your environment.
  • Confirm responsible owner for model/prompt governance.
  1. Scope definition: inbound vs outbound
  • Decide whether phase-1 rollout includes inbound, outbound, or both.
  • Define target call use cases by team.
  1. Greetings and instructions
  • Draft Opening Greeting Message (inbound).
  • Draft Opening Introduction Message (outbound).
  • Define Sales Assistant Instructions & Approach.
  1. CSV format and template
  • Download CSV Template.
  • Validate required columns in source files.
  • Run preview before campaign launch.
  1. Launch schedule and attempt rules
  • Set launch now or schedule later policy.
  • Define launch date/time windows.
  • Set Max Call Attempts guardrails.
  1. Qualification criteria
  • Enable Smart Qualification only after criteria are documented.
  • Define Qualification Criteria aligned to sales stages.
  1. Post-call source/status mapping
  • Define Post-Call Lead Source values.
  • Define Post-Call Lead Status values.
  • Validate CRM follow-up ownership for each outcome.
  1. Knowledge-base selection
  • Select approved Knowledge Bases for assistant grounding.
  • Remove outdated or conflicting references.
  1. Phone and web testing before go-live
  • Run Test Call to phone path.
  • Run Test Assistant (Web) with Start Conversation / End Conversation.
  • Capture test findings and required prompt changes.
  1. Governance for transcripts and recordings
  • Define retention and access policy for transcripts and call records.
  • Align leadership/legal expectations for review access.

Key workflows

Workflow 1: Inbound setup

  1. Open Inbound Call Assistant Settings.
  2. Set Opening Greeting Message.
  3. Define assistant personality and approved boundaries.
  4. Run a test conversation.
  5. Adjust instructions based on summary/transcript quality.
  6. Approve and publish inbound handling behavior.

Workflow 2: Outbound campaign setup

  1. Create AI Campaigns entry.
  2. Complete Campaign Setup (name, description, target use case).
  3. Set Outbound Call Assistant Settings and Opening Introduction Message.
  4. Add Sales Assistant Instructions & Approach.
  5. Confirm launch controls and readiness checklist.

Workflow 3: CSV import and launch

  1. Download CSV Template.
  2. Prepare contact-list file and validate required columns.
  3. Upload CSV File and inspect preview.
  4. Set Max Call Attempts.
  5. Choose launch now or schedule later.
  6. Launch and monitor progress dashboard.

Workflow 4: Smart qualification

  1. Enable Smart Qualification.
  2. Define Qualification Criteria.
  3. Run calls for selected campaign.
  4. Review qualified vs non-qualified outcomes.
  5. Validate lead quality with Sales/RevOps.
  6. Adjust criteria if too broad or too strict.

Workflow 5: Call review

  1. Open AI Call logs.
  2. Open Call Details for selected records.
  3. Review Summary and Transcripts.
  4. Check ended reason and call duration.
  5. Confirm post-call source/status mapping quality.
  6. Feed improvements into prompts or campaign settings.

Workflow 6: Test-call validation

  1. Run Test Call (phone path).
  2. Run Test Assistant (Web).
  3. Validate tone, qualification behavior, and information quality.
  4. Document defects and improve prompts/greetings.
  5. Re-test before launch approval.

Workflow 7: Knowledge-base refinement

  1. Select Knowledge Bases relevant to campaign use case.
  2. Connect references to assistant behavior.
  3. Re-test factual responses.
  4. Review call summaries for consistency.
  5. Update knowledge and prompt controls as needed.

Reports

Campaign statistics

  • Contains: total campaigns, active campaigns, total calls made, progress indicators.
  • Used by: Campaign Manager, Leadership.
  • Supports decisions: campaign pacing, resource allocation, rollout confidence.

AI Call logs

  • Contains: call-level list, direction (Inbound / Outbound), duration, ended reason, linked lead context.
  • Used by: Campaign Manager, RevOps Reviewer.
  • Supports decisions: call quality review, follow-up prioritization, issue detection.

Summary and Transcripts

  • Contains: condensed call outcome narrative and transcript-level detail.
  • Used by: AI Sales Admin, QA/RevOps.
  • Supports decisions: prompt improvements, coaching insights, qualification consistency.

Qualification outcomes

  • Contains: smart-qualification pass/fail patterns and configured criteria impact.
  • Used by: RevOps Reviewer, Sales leaders.
  • Supports decisions: tighten/relax qualification thresholds and campaign targeting.

Duration and success-rate review

  • Contains: average call duration and success-rate trend by campaign or period.
  • Used by: Campaign Manager, Leadership.
  • Supports decisions: script quality changes, timing strategy, retry policies.

Launch and test-review outputs

  • Contains: pre-launch test notes, approval status, known issues list.
  • Used by: AI Sales Admin, Campaign Manager.
  • Supports decisions: go-live approval, rollback conditions, post-launch monitoring plan.

Troubleshooting / FAQ

CSV file invalid

  • Confirm the file is CSV and encoded correctly.
  • Re-download and use Download CSV Template.
  • Remove unsupported formatting and re-upload.

No valid phone column

  • Confirm required phone column is present and populated.
  • Normalize values before Upload CSV File.

Campaign launches but quality is poor

  • Review Opening Introduction Message and Assistant Instructions.
  • Re-check Quick Start Templates / System Prompt Templates selection.
  • Run test calls and iterate before scaling volume.

Summaries/transcripts are not useful

  • Tighten instructions for summary style and qualification objective.
  • Review transcript quality against expected question flow.
  • Add explicit post-call outcome requirements.

Qualification criteria too broad or too strict

  • Revisit Qualification Criteria and business thresholds.
  • Compare qualified outcomes with actual sales acceptance.
  • Calibrate criteria with RevOps and campaign owners.

Post-call lead status mapping is confusing

  • Align Post-Call Lead Source and Post-Call Lead Status with CRM stages.
  • Publish a short mapping matrix for campaign managers.

Knowledge base is not improving answers

  • Remove irrelevant knowledge bases.
  • Add approved, campaign-specific sources.
  • Re-test with both Test Call and Test Assistant (Web).

Team confusion between AI assistant and CRM ownership

  • Clarify scope: AI Sales Assistant handles first conversation + qualification layer.
  • CRM and Sales remains the system of record for account ownership, pipeline, and closing workflows.

Need help with this section? Contact our team for guided setup support.