Soliq · Document AIDocument AI for KYC: minutes, not days
A digital lender cut KYC turnaround from 4–7 days to 12 minutes with a Soliq document-AI pipeline: 99.4% accuracy, 81% less manual review.
A stylised excerpt of the production UI.
The problem we walked into.
KYC documents arrived in 23 formats from 18 source bureaus. Manual reviewers (three teams, eight hours/day, 27 reviewers) kept up but cost ₹X/application and bottlenecked at 50k apps/month. Volume target: 200k/month within 6 months.
How we attacked it.
- 01
Document classification first: route each file to a specialised extraction model.
- 02
Layout-aware extraction (PAN, Aadhaar, payslips, bank statements, address proofs, employment letters).
- 03
Confidence scoring per field. > 0.96: auto-accept. 0.85–0.96: light human review. < 0.85: full review.
- 04
Cross-document consistency checks (name across PAN/Aadhaar/payslip; address across utility/bank).
- 05
PII redaction before any log enters our pipeline; on-prem deployment for production workloads.
- 06
Continuous eval set; quality drift alarmed within 24 hours.
The tools, picked deliberately.
“The eval discipline alone justified the engagement. Other vendors gave us a demo and walked away; Techliphant set us up to run this ourselves.”
Want the full case PDF?
We share the unabridged story: architecture diagrams, eval set excerpts, runbooks, all under mutual NDA.
More cases.
Ready when you are
Let's build something exceptional.
Tell us about your business, your stack, and the problem you are trying to solve. We respond with a clear next step usually a 30-minute discovery call, no fluff.