Every call makes the next one better.
Graded, curated, fine-tuned, tested, approved, measured. Every week, on your own calls.
Most voice bots are as good on day 300 as on day 1.
"It learns" is not a mechanism. This is.
A named pipeline, not a promise.
One week of learning, hung as seven works. Walk the wall.
The city calls
Recordings, transcripts, outcomes, human takeovers and compliance flags stream in as signals — 100% of calls, not a sample.
48,210 signals · 7 days illustrativeOne hundred calls, graded
Grounding, flow, compliance, tone. Failures are clustered: "repeated the EMI amount", "missed already-paid", "asked branch-or-link twice".
100% graded · 14 clusters illustrativeRedaction, then selection
Aadhaar, PAN, card and phone numbers masked in text and audio. Your best agents' calls and the corrected failures become the training set — approved by your reviewer.
PII 100% · 3,120 examples illustrativeYour weights
Speech and language models fine-tuned per tenant. No shared brain. The weights are yours, stored in your environment for VPC and on-prem.
candidate v42 · your tenant illustrativeTwelve hundred and forty scenarios
The candidate replays the golden set — barge-in, angry customer, already-paid, wrong number — with Hindi/English parity and frozen-line checks. Any drop blocks release.
1,240 scenarios · 0 red illustrativeA human signs
Named approval from your QA or compliance owner, staged rollout 5% → 25% → 100%, automatic rollback on a guardrail breach.
5% → 25% → 100% · 1-click back illustrativeSix weeks, the dip left in
Containment, promise-to-pay, right-party contact, complaints and cost per resolved call — tracked per release, reported every morning. Then the next week begins.
PTP +8.5 pts over 6 weeks illustrativeNext week's calls feed the next cycle.
Client memory carries over: scripts, glossary, frozen lines, escalation matrix.
Illustrative week for one collections cohort. replace with a live agent's counters
Recursive, not just repeated.
The per-turn check produces the call grade. Call grades produce the week's training set. The week's release produces the portfolio's lessons. And the evaluation model that grades all of it is retrained on your reviewers' corrections — the loop that grades the loop.
Your model. Your data. Nothing shared.
We fine-tune speech and language models on your graded, redacted calls. The improvement lives in weights you own.
- "Yeh call record ho rahi hai." — recording disclosure
- "Main AI assistant hoon." — AI disclosure
- Grievance redressal line and helpline number
- RBI Fair Practices Code calling-hour and tone rules
- IRDAI product-claim restrictions
- Objection handling and phrasing per language
- Intent recognition (new customer phrasings)
- Number, date and amount pronunciation
- Handle-time and turn-taking behaviour
v41 missed it. v42 learned it.
"Already paid" — flagged 212 times in a week. The next release looks up the UTR and closes the call.
- 00:00Maine kal UPI se pay kar diya tha.मैंने कल UPI से pay कर दिया था।
- 00:04Aapki 5 tareekh ki EMI ₹4,850 abhi tak pending hai. Kya aap aaj payment kar payenge?repeats the amount · misses "already paid"
- 00:00Maine kal UPI se pay kar diya tha.मैंने कल UPI से pay कर दिया था।
- 00:04Theek hai, dhanyavaad! Main abhi check kar rahi hoon… UTR mil gaya, payment 15 September ko receive hua hai. Aapko aage koi call nahi aayegi.stage: already_paid → utr_lookup → close ✓
Rising week over week, with the dip left in.
When v41 breached a live guardrail, canary caught it. We show that, because unseen failure cannot be trusted.
Illustrative shape of six weeks of weekly learning cycles on one early-bucket cohort; every release marked, the dip kept in. Pilot data replaces this. replace with cohort data + methodology note
Every version has a name, an approver and a way back.
frozen RBI disclosure · consent · grievance · approved client QA lead + FDE
regression 1,240 / 1,240 · ● live · canary → 100%
regression 1,198 / 1,198 · ▲ canary breached tone guardrail → auto-rollback 41 min → fixed → re-released
regression 1,150 / 1,150 · ✓ approved client compliance + FDE
Example release notes for illustration. replace with a live agent's notes
How a bank can accept a self-learning agent.
- Compliance lines never learn.
- Disclosures, consent and grievance lines are frozen, version-hashed strings.
- Humans approve every release.
- Named sign-off from your QA or compliance owner and our engineer, logged.
- Regression before release.
- Golden scenarios and last week's failures; any compliance drop blocks it.
- Canary, then scale.
- 5% → 25% → 100%, auto-rollback on breach, previous version always warm.
- Learn only from what you allow.
- Per-client boundaries, PII redacted before storage, pause per journey.
- Every change is explainable.
- Which calls taught it, who approved it, scores before and after — for RBI, IRDAI and audit.
Climb at your pace. Own every level.
Most BFSI deployments start at L3. The weekly cycle takes you to L4, under the policy your compliance office signed.