Every support call is checked against your quality checklist, the same day

Each recorded call is written out, masked and scored against your checklist, with the score in the Freshdesk ticket. Team leads hear about serious misses the same day and coach from the exact moment.

Used day to day by
  • Quality team
  • Support team leads
  • Customer service head

How it runs with AI

The AI writes out every recorded call, in the language spoken, masks personal details and scores the call against your checklist, noting the minute each point was met or missed. The score goes into the Freshdesk ticket, and calls that miss a critical item reach the team lead on Teams the same day.

How it is done today

The quality team listens to a small sample of support calls, picked at random, and scores each one against a checklist in Excel. Agents hear about a miss weeks later, and a wrong promise on any other call goes unheard.

Step by step

Where the work happens, what the AI does, and where it hands over to a person.

  1. Step 1, AI

    Writes out each recorded call and masks phone numbers, addresses and account numbers

    Needs to know: Details to mask, Languages callers use

    In Call recordings

    Exception to step 1

    Sends calls with poor audio to the quality team, who listen and score them

  2. Step 2, AI

    Scores the call against your checklist and the ticket, noting the minute of each point

    Needs to know: Quality checklist, Which items are critical

    In Call recordings , Freshdesk

  3. Step 3, AI

    Records the score in the Freshdesk ticket, with a link to each missed moment

    Needs to know: Score access rules

    In Freshdesk

  4. Step 4, AI

    Alerts the team lead on Teams the same day when a call misses a critical item

    Needs to know: Team lead by agent

    In Microsoft Teams

  5. Step 5, Person

    Listens to the marked moments, confirms or corrects the score and coaches the agent

    In Freshdesk , Call recordings

What stays with a person

  • Listens to the marked moments, confirms or corrects the score and coaches the agent

How caller details are protected

Call recordings and transcripts are personal data. The set-up handles them this way:

  • Notice. Callers hear that the call is recorded, and that quality checks are one of the reasons, before they reach an agent. The notice says in plain language what is collected and why, and your legal team confirms the wording.
  • Masking. Phone numbers, addresses, account numbers and any card or Aadhaar number are masked in the transcript before the scoring step or any person reads it.
  • Access. Agents see their own scored calls and team leads see their team’s. Full recordings stay with the quality team.
  • Deletion. Transcripts are deleted after the period your policy sets.

Apps in this example

  • Call recordings
  • Freshdesk
  • Microsoft Teams

The same flow works with the apps you use.

This fits you if

  • Your support team takes more than 5,000 calls a month
  • Your quality team listens to a small sample today, often 1 to 3 in every 100 calls
  • You already record calls and score them against a written checklist
  • Calls come in more than one language, across centres or shifts

What to watch for

  • Before it goes live, the quality team scores 200 past calls by hand, in each language, and compares their scores with the AI's item by item
  • A number read out in words, like "nine eight four five", is harder to spot, so a person checks the masking on a sample of transcripts every week
  • Agents know their calls are scored by software, see their own scores and can ask the quality team to review any of them
  • Warnings, ratings and any step about an agent's job stay with team leads and HR. The score is where the coaching conversation starts

Read more

Talk to us about this for your company

Explore AI agents
Book an intro