Qtangent
Knot 03AI Feedback Analyst

Thousands of comments. Three clear priorities. Every one backed by real words.

The AI Feedback Analyst reads every survey comment, review, mention and call transcript in every language, finds the themes and root causes that drive your scores down, and links each finding to the actual customer comments.

AI Feedback Analyst · Detractor themes, Q3 6,240 detractor comments read
1Disbursal delays after sanction34%

Worst in 12 branches

2Unclear foreclosure charges21%

Rising since August policy change

3Call-centre wait times on Mondays14%

Peaks 10am to 1pm

Likely root cause, theme 1

Property valuation reports take 9.4 days on average in 12 branches, against 3.1 days elsewhere. 71% of delay comments mention “valuation” or “technical visit”.

Real comments behind theme 1

Sanction letter came fast, then nothing for 3 weeks.

NPS 3 · Thane West · Survey

Builder is asking for payment, bank keeps saying valuation pending.

NPS 2 · Pune Kothrud · Call

पैसे कब मिलेंगे, कोई जवाब नहीं देता

NPS 4 · Nashik Road · WhatsApp

Showing 3 of 2,121 linked comments

18,400

Comments read in one quarter, in minutes

Impact-ranked

Themes sorted by effect on NPS

100%

Of insights linked to real comments

What you get

Everything in AI Feedback Analyst.

Theme discovery

Finds themes on its own, then lets you rename, merge or split them to match how your business talks.

Impact ranking

Ranks themes by how many points each one costs your NPS or CSAT, not just by how often it appears.

Root cause tracing

Connects themes to branches, products, agents, policies and dates to point at the likely cause.

Real comments, always

Every number opens the comments behind it, with score, channel and location for each one.

Emerging issues

Flags themes that are new or rising fast, such as complaints after a policy change.

Shareable summaries

A plain-language weekly brief for leadership, with masked comments and clear next steps.

How it works

How the analyst turns voices into priorities.

  1. 1

    Read everything

    Surveys, reviews, mentions and transcripts in every language, as they arrive.

  2. 2

    Group and rank

    Comments are grouped into themes and ranked by their effect on your scores.

  3. 3

    Trace the cause

    Each theme is broken down by branch, product, team and time to find where it starts.

  4. 4

    Hand to people

    Your CX lead reviews, shares with owners and turns themes into actions.

Root causes

From “customers are unhappy” to “valuation reports are late in 12 branches”.

Theme volume tells you what people complain about. Root cause tracing tells you where and why. The analyst compares branches, products, dates and process data so operations can act on something specific.

  • Branch and team breakdowns for every theme
  • Change detection around policy and release dates
  • Links to operational data you connect
  • 01Branch and team breakdowns for every theme
  • 02Change detection around policy and release dates
  • 03Links to operational data you connect

Questions in plain words

Ask the analyst anything about your feedback.

Type a question such as “Why did NPS drop in Pune last month?” and get an answer with the themes, numbers and comments that explain it. Every answer shows its sources.

  • Answers cite the comments they used
  • Filters for date, branch, product and channel
  • Save any answer to a dashboard
  • 01Answers cite the comments they used
  • 02Filters for date, branch, product and channel
  • 03Save any answer to a dashboard

In practice

How teams use AI Feedback Analyst.

NBFC

6,240 detractor comments were grouped into themes. Disbursal delays explained 34% and traced to valuation reports in 12 branches.

Delay complaints down 46% next quarter

Insurer

The analyst flagged a new theme about claim document requests two weeks after a process change.

Checklist simplified, claim CSAT up 0.6

Edtech

Refund comments in Hinglish and Tamil were grouped with English ones for the first time.

Refund policy rewritten in plain language

Customer respect and responsible AI

The analyst explains. Your people decide.

Consent before contact

Surveys and call-backs go only to customers who agreed to be contacted. Every message carries an easy opt-out, and frequency limits apply per customer.

Every customer asked

Review requests go to all customers, never only happy ones, in line with Google and app store rules.

People decide, people call

AI finds themes and shows the real comments. Your team decides the actions and makes every call-back.

Masked personal details

Names, phone numbers, emails and account numbers inside comments are masked in shared dashboards and exports.

Encrypted, never sold

Data is encrypted in transit and at rest, processed under India's data protection rules and never sold or used to train public models.

Explainable themes

Every theme shows how it was formed, which comments sit inside it and how confident the analyst is.

The line is open

18,000 comments are waiting to be read.

Find your three things to fix this week.

Coming down the string

  • 3

    Thane West branch

    2 min ago

    Call back
  • 9

    Play Store review

    4 min ago

  • 5

    Call centre survey

    6 min ago