Nanonets Platform:Analytics

Track anything
about your workflows

Type a question in plain English. Analytics writes the query against your agent data and returns a chart. Save it, build a dashboard, and track usage and cost across every agent so nothing runs unchecked.

Nanonets
Agents
Tasks
Connectors
Documents
Analytics
Usage
Settings

Analytics › Dashboards › AP Command Center

LLM Logs

Days Payable Outstanding

32

Touchless Rate (STP)

94%

Discounts Captured

71%

Open Liability (GR/IR)

180

Blocked Invoices by Reason

On-Time Payment Trend

How it works

From question to insight in seconds

01Ask ANY question

Example queries

What's my touchless rate this month?Which vendors have the most exceptions?How long does invoice processing take?Show DPO trend over the last 6 months
What's my DPO trend this quarter?
02Get a chart
What's my DPO trend this quarter?

No SQL. No BI tool. Ask in plain English and the right chart comes back instantly.

DPO this quarter
03Inspect the querySQL

-- Not a black box.

-- Every answer has a query behind it.

-- Inspect it. Edit it. Build from scratch.

SELECT DATE_TRUNC('month', processed_at)

AVG(processing_time_hours) AS avg_dpo,

COUNT(*) AS invoices

FROM agent_runs

WHERE agent_id = 'ap-agent'

GROUP BY 1 ORDER BY 1

04Save and share

Group questions into a named dashboard and share with your team. No BI license required.

JFMAMJ
JFMAMJ
Shared with 3
Saved · Refreshes daily

What you can measure

Business goals and AI performance, in one place

Business goals

Did the workflow achieve its outcomes?

Track the operational metrics your team actually cares about: how much got done, how fast, and how much value moved through the system.

Total processedTotal invoices processed this month
Automation rate% of orders completed without human review
Processing timeAvg hours from receipt to completion
Value processedTotal dollar value of POs approved
Pending exceptionsOpen exceptions by type and agent
Top customers / vendorsTop 10 vendors by invoice volume
Status distributionCompleted vs pending vs rejected breakdown
SLA performance% of documents processed within target time
AI performance

How well is the AI performing?

Monitor model confidence, exception rates, and step-level timing. Catch degradation before it becomes a problem. Full LLM call logs available for every run.

Confidence scoresAvg extraction confidence by document type
Exception rate% of runs that required human review
Human review rateWhich step generates the most exceptions
Extraction accuracyFields extracted correctly vs flagged
Model performanceConfidence trend over time
Step-level timingWhich agent step takes the longest
Retry rateHow often does the agent retry a step
LLM call logFull trace of every model call per run

Usage

Know exactly what AI is costing

Every AI action is metered. The Usage tab shows your credit balance, projected depletion, and a full breakdown of spend by agent, step type, and time period. You can see the bill before it arrives.

  • Live credit balance with projected depletion
  • Cost broken down by agent and step type
  • Filter by date range to isolate a specific period
  • Top agents by credit spend at a glance

Billing

Track credits consumed across agents and tasks

Credit Balance

17,578 credits used

Growth Plan
Renews Jun 28
Agents: All
Step Type: All
34,763 credits
Dec 25Jan 26Feb 26Mar 26Apr 26May 26

For technical teams

Natural language for everyone. SQL when you need it.

Non-technical teams get plain-language queries and instant charts. Technical teams can inspect and edit the generated SQL, browse the underlying analytics tables, and build custom queries from scratch.

  • View and edit the AI-generated SQL for any question
  • Browse the analytics tables your agent writes to
  • Full LLM call trace log per agent run
  • Export raw data for use in external BI tools
SQL Query
SaveCopy

SELECT

DATE_TRUNC('month', processed_at) AS month,

COUNT(*) AS total_processed,

AVG(processing_time_hours) AS avg_hours,

SUM(CASE WHEN human_reviewed = false

THEN 1 ELSE 0 END) * 100.0 /

COUNT(*) AS automation_rate

FROM agent_runs

WHERE agent_id = 'ap-agent'

AND processed_at >= NOW() - INTERVAL '6 months'

GROUP BY 1

ORDER BY 1

6 rows · 12ms
Analytics Tables

See it run on your process, with your documents.

Start free. No credit card. Or talk to our team about your workflow.