Skip to main content

Usage Statistics

The Usage Statistics page provides detailed insights into your AI SmartTalk consumption, helping you understand costs, optimize usage, and track AI activity across your organization.

Usage Statistics

Accessing Usage Statistics​

From the Organization page, click View Stats → on the usage overview card to access detailed statistics.

Page Overview​

The statistics page is organized into several sections:

  1. Usage Summary Bar - Quick overview of resource consumption
  2. Period Metrics - Key usage indicators
  3. Cost Breakdown - Where your AI credits go
  4. Traceability - Link costs to sources
  5. Daily Usage Chart - Consumption over time
  6. Recent Activity - Detailed action log

Usage Summary Bar​

At the top of the page, the same usage bar from the Organization overview displays your current consumption:

MetricDescriptionExample
AI TokensTotal tokens consumed0 / 200.0M (0%)
Documents / ProductsIndexed items0 / 100.0K (0%)
AgentsActive assistants1 / 5 (20%)
SeatsTeam members1 / 10 (10%)

Period Metrics​

Four cards display key metrics for the current billing period:

Tokens This Period​

FieldDescription
ValueTotal tokens consumed (e.g., "4.1K")
TrendPercentage change from previous period (e.g., "↑ 751%")

The trend indicator shows:

  • ↑ Green: Increase from previous period
  • ↓ Red: Decrease from previous period

AI Calls​

FieldDescription
ValueNumber of AI API calls (e.g., "56")
PeriodTime range (e.g., "30 days")

Average Per Call​

FieldDescription
ValueAverage tokens per AI call (e.g., "72")
Unit"tokens/call"

Daily Average​

FieldDescription
ValueAverage daily token consumption (e.g., "135")
Unit"tokens/day"

Cost by Action​

This section breaks down token consumption by action type, helping you understand what drives your AI usage.

Action Types​

ActionDescriptionExample
💬 ChatbotTokens used by chat conversations2.2K (55%)
⚡ WorkflowTokens consumed by SmartFlow automations1.2K (30%)
🔄 Client EmbeddingsTokens for document processing and indexing624 (15%)

Reading the Chart​

Each action shows:

  • Icon: Visual identifier for the action type
  • Name: Action category
  • Token Count: Absolute number of tokens
  • Percentage: Share of total consumption
  • Progress Bar: Visual representation (color-coded)

Color coding:

  • 🔴 Red: Chatbot interactions
  • 🟠 Orange: Workflow automations
  • 🟡 Yellow: Embeddings and processing

Cost by AI Model​

Understand which AI models are consuming your tokens:

Model Breakdown​

ModelDescriptionExample
GPT4O_MINIOpenAI's efficient GPT-4 model3.4K (85%)
OPENAI_ADAEmbedding model for document processing624 (15%)

Model Characteristics​

ModelUse CaseCost Efficiency
GPT4O_MINIConversations, AI responsesBalanced
OPENAI_ADADocument embeddings, searchHigh efficiency
GPT4OComplex reasoning tasksPremium
ClaudeAdvanced analysisPremium

Traceability​

The traceability section shows how well costs can be attributed to specific sources:

Source Types​

SourceDescriptionExample
🔗 SmartFlowsCosts linked to specific workflows6 (11%)
💬 MessagesCosts attributed to chat messages8 (14%)
📧 ChatsCosts tracked to conversation sessions54 (96%)

Understanding Traceability​

High traceability percentages indicate:

  • Well-structured workflows
  • Proper tagging and attribution
  • Easy cost analysis and optimization

Low traceability may suggest:

  • Background processing tasks
  • System maintenance operations
  • Untagged automated actions

Daily Usage Chart​

A line chart visualizes token consumption over time:

Chart Features​

ElementDescription
X-AxisDate range (e.g., "11 Dec" to "17 Dec")
Y-AxisToken count scale (e.g., 0 to 2.4K)
LineDaily consumption trend
AreaFilled area under the curve for visual clarity

Reading the Chart​

  • Peaks: High-activity days (product launches, campaigns)
  • Valleys: Low-activity periods (weekends, off-hours)
  • Trends: Overall consumption direction

Chart Insights​

Use the chart to:

  • Identify usage patterns
  • Plan for capacity
  • Detect anomalies
  • Optimize scheduling

Recent Activity​

The activity table provides a detailed log of all AI operations:

Recent Activity

Table Columns​

ColumnDescription
TimeTimestamp of the action (e.g., "17/12/2025 20:13")
ActionType of operation (Workflow, Chatbot, etc.)
SourceOrigin of the request with link
ModelAI model used
TokensTokens consumed
TraceLink to detailed trace

Activity Filters​

Filter the activity log by type:

FilterShows
AllAll activity types
SmartFlowWorkflow executions only
TracesTraced operations
MessagesChat messages only

Action Details​

Click on an action row to see:

  • Full request/response content
  • Execution duration
  • Error details (if any)
  • Related actions

The ↗️ Trace icon opens the full execution trace, showing:

  • Complete workflow path
  • Step-by-step execution
  • Variable values
  • Performance metrics

Exporting Data​

Available Exports​

FormatContains
CSVRaw activity data
PDFFormatted usage report
APIProgrammatic access

Export Use Cases​

  • Accounting and invoicing
  • Performance analysis
  • Compliance reporting
  • Capacity planning

Optimization Tips​

Reducing Token Consumption​

StrategyImpact
Use efficient AI modelsChoose GPT4O_MINI over GPT4O when possible
Optimize promptsShorter, clearer prompts use fewer tokens
Cache common responsesReduce redundant AI calls
Batch operationsCombine multiple requests

Improving Traceability​

StrategyBenefit
Tag workflowsEasy cost attribution
Use conversation contextLink related actions
Enable loggingComplete audit trail

Monitoring Best Practices​

  1. Set alerts for unusual consumption spikes
  2. Review weekly to catch trends early
  3. Compare periods to understand growth
  4. Analyze by model to optimize costs

Troubleshooting​

High Token Usage​

SymptomPossible CauseSolution
Sudden spikeLoop in workflowCheck SmartFlow for infinite loops
Gradual increaseGrowing user baseConsider upgrading plan
Consistent highInefficient promptsOptimize AI configurations

Missing Activity​

IssueSolution
Activity not showingCheck date range filter
Incomplete dataWait for data sync (up to 5 minutes)
Zero consumptionVerify assistant is active

Incorrect Attribution​

IssueSolution
Untraced costsAdd workflow tags
Wrong sourceCheck integration configuration

Ready to elevate your
user experience?

Deploy AI assistants that delight customers and scale with your business.

GDPR Compliant
Start Free TrialAccess Backoffice