GE Aerospace
Independently reportedManufacturing · Process automation · Classical ML · Scaled
Process automation
Shriram Finance
Independently reportedBanking & financeIndiaIn productionEnterpriseClassical ML
Independently reported. An independent or peer-reviewed source reports this result, alongside the vendor.
Reported by enterpriseai.economictimes.indiatimes.com (independent reporting). Checked against the source page on 2026-09-05. 9 of 9 figures below appear on that page word for word.
The problem
Onboarding for two-wheeler loans was paper-heavy, involved multiple systems, and took 24-48 hours.
What was deployed
- Products
- Ziva 2.0, Shriram One
- Technique
- Classical ML
- Build or buy
- Built in-house or to order
- Deployment
- Custom build
- Data used
- KYC data, credit bureau checks, thousands of customer data points for dynamic pricing (rate, LTV, tenure)
- Scale
- Shriram One: 2 crore downloads, 1.4 crore active users; Ziva 2.0 live across two-wheeler branches, now extending to Green Finance and Commercial Vehicles
What changed
Each row is quoted from the source. Figures we could not find on the page in those words are marked — they are kept, not deleted, so you can judge them.
less than 3–4 hoursOnboarding Turn-Around-Time (TAT)from 24–48 hours
“Ziva 2.0 has reduced the onboarding TAT (turn-around-time) to less than 3–4 hours.”
Quoted word for word from the source
Disbursal throughput per executive
“a 15–20% jump in disbursal throughput per executive in branches where Ziva has gone live.”
Quoted word for word from the source · round percentage, no baseline given
2 crore (~20 million)Shriram One app downloads
“2 crore downloads and 1.4 crore active users”
Quoted word for word from the source
1.4 crore (~14 million)Shriram One active users
“2 crore downloads and 1.4 crore active users”
Quoted word for word from the source
40:60 (existing:new)Customer mix (existing vs new)from 80:20 (existing:new)
“shifting the mix from 80:20 (existing:new) to 40:60 in favour of new customers”
Quoted word for word from the source
20-25 minutesPaperless eKYC journey processing timefrom 24–48 hours
“Paperless, eKYC-led journeys have also reduced processes that earlier took 24–48 hours to just 20–25 minutes.”
Quoted word for word from the source
5%Digital channel contribution to disbursals
“Digital channels now contribute 5% of total disbursals across products driven by higher-quality, pre-screened leads”
Quoted word for word from the source
NPA (non-performing asset) performance improvement
“early results “show a 20% improvement in NPA (non-performing asset) performance.””
Quoted word for word from the source · round percentage, no baseline given
Chatbot-driven reduction in team workload
“Our AI-enabled chatbot now manages a large portion of customer queries, reducing team workload by nearly 60%”
Quoted word for word from the source · round percentage, no baseline given
ROI as stated: The hard numbers underline the ROI: higher straight-through approvals, faster disbursals, and a 15–20% jump in disbursal throughput per executive in branches where Ziva has gone live.
“Ziva 2.0 has reduced the onboarding TAT (turn-around-time) to less than 3–4 hours.”
Difficulties and limits
- AI governance risks — data quality, fairness, model explainability and unintended customer impacts (Kumar's own stated concern)
- Model deviation allowed: marketing executives can adjust parameters with approval when local trade dynamics or seasonality warrant deviation from the model's suggestion
- A regional repayment-behaviour mismatch against a model threshold was flagged by the branch team and required model recalibration before performance normalized
Sources
- enterpriseai.economictimes.indiatimes.com — independent reporting, 2025-12-02
Similar deployments
Broadcom · Moveworks
88% — autonomous IT resolution rate
Independently reportedTechnology & software · Process automation · NLP · Scaled
Apollo Tyres
under 10 minutes — Time for root cause analysis
Independently reportedManufacturing · Process automation · AI agents · In production
icare · SAS
increased by 15% — accuracy
Independently reportedInsurance · Process automation · Classical ML · Scaled · 2019
Amazon · Amazon
10% more — shopper spend
Independently reportedRetail & e-commerce · Process automation · Computer vision · Scaled · 2024
Met Office · AWS
under five minutes — Time to complete textual forecast generation
Independently reportedPublic sector · Process automation · LLM · Pilot