If Batch Call is where you plan, Call History is where you learn. This page shows every attempt, who called whom, what the user did, and how your agent performed. Read it once, and you know which scripts work, which numbers struggle, and where to fix effort today.
The filters on top
Call Type
Pick inbound, outbound, outbound scheduled, outbound API, or agent test. Use it to isolate one stream, for example only API calls from your CRM.Include batch
Type a batch name to see calls from a specific campaign. Handy when you want a clean view for one upload.Agent Type
Focus on one bot at a time. This helps when A or B testing voices or prompts.Disconnect reason
Filter by outcomes such as user cut the call, user requested callback, voicemail detected, user unavailable, invalid number. This is gold for diagnosis, since you can tie reasons to script lines and time slots.Duration
Buckets like not connected, 1 to 10s, 11 to 30s, 31 to 60s, over 60s. Long calls are great for quality review. Very short ones often signal the opener needs work.Date range pickers
Choose start and end dates to compare days or weeks.Phone number search
Paste a number to pull its history for support or audits.Search, Reset Filters, Export as Excel
Apply or clear your view. Export gives your team a shareable file for BI or compliance.
The table columns, what each tells you
Type
Inbound or outbound, plus the exact flavor. Quick way to separate planned campaigns from callbacks.From, To
The caller and callee numbers. Use this to check routing and to spot repeated dials to the same person.Agent
The bot that handled the call, for example Heltar Outbound SDR - Hindi. When you pair this with duration and disposition, you can judge each agent’s script and tone.Time
Timestamp of the attempt. Useful for time-of-day analysis. If early afternoon beats evening, you will see it here.Disposition
Short text on what happened, for example user cut the call, user requested callback. This is your first stop for fixes. If many users ask for a later call, enable auto rescheduling in Call Settings. If many hang up in under 10 seconds, tighten the opener.Duration
Seconds on the line. Longer is not always better, but it often means the user stayed engaged. Review long calls for training and QA.Model
The LLM used. Correlate cost and quality by model to pick the right tier for each use case.Voice Provider
For example Eleven Labs. If a region prefers a calmer voice, test it and check results here.Transcriber
For example: Deepgram. If transcripts miss terms in a language, try a different model and measure again.
How this page helps the business?
Faster diagnosis
Filter by disconnect reason, then read the first 10 examples. You will know exactly what to change in the prompt, voice speed, or call window.Agent tuning
Compare two agents on the same batch. Keep the one with higher connected to converted ratio. Retire what does not work.Number health checks
Filter by From number. If one line shows more user rejected or user unavailable, switch the next batch to a cleaner number, then update Spam Status on the Phone Numbers page.Compliance and support
Search a phone number to answer user queries or to prove opt outs. Export to Excel when you need an audit trail.Cost control
Model column plus duration tells you where money goes. Use mini models for short, structured flows, and full models for complex cases.
A simple workflow that works
After each batch, open Call History and filter by that batch.
Sort by Duration and skim the long calls for quality.
Check Disposition counts. Fix one thing at a time, opener, timing, or reschedule rule.
Export and share a short note with actions. Run the next batch with those changes.
Call History in Heltar is not just a log. It is your daily scoreboard. Use it for quick reads, make one improvement per day, and your answer rates and conversions will keep climbing. Contact us for a demo if this is what your business needs.
