Shriram Finance
less than 3–4 hours — Onboarding Turn-Around-Time (TAT)
Independently reportedBanking & finance · Process automation · Classical ML · In production
Process automation
Halo Exploration (HE) · Odin AI
Vendor-reportedEnergy & utilitiesPilotClassical MLLive 2026
Vendor-reported. The customer is named and the numbers are quoted from the source page, but the account comes from the vendor. No independent confirmation.
Reported by skyfrac.com (vendor self-report). Checked against the source page on 2026-09-05. 5 of 5 figures below appear on that page word for word.
The problem
Halo Exploration (HE) needed to understand why its wells were underperforming. Conventional completion levers explained ≤12% of the production spread. Seven wells had seven different bottlenecks, and the levers that actually matter (landing depth, rate execution, stage size) were tangled together, making optimization impossible.
What was deployed
- Vendor
- Odin AI
- Products
- Odin AI, CFrac, Screen-Out Model
- Technique
- Classical ML
- Build or buy
- Bought off the shelf
- Deployment
- SaaS
- Data used
- pumping data, live pumping data
- Scale
- 7 wells
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.
93.8%production variance explained by CFracfrom ≤12%
“CFrac alone explains 93.8% of production variance — every traditional driver explains ≤12%.”
Quoted word for word from the source
0.46correlation (R²) between stage CFrac and distance to S2 bench
“Across all 95 stages of 103/10-20, stage CFrac tracks how close each stage lands to the S2 bench — R² = 0.46 at p ≈ 6.6e-14.”
Quoted word for word from the source
0.94correlation (R²) between CFrac and 6-month production6-month
“And because CFrac in turn predicts 6-month production (R² = 0.94), it doubles as a measurement instrument”
Quoted word for word from the source
~17%CFrac explained by rate-hold quality
“In 103/10-20, how cleanly a stage held design rate explains ~17% of its CFrac.”
Quoted word for word from the source
77%Production variance explained by CFrac-normalized fluid intensityfrom 10%
“Raw fluid intensity explains 10% of variance and trends negative. CFrac per unit fluid intensity explains 77%.”
Quoted word for word from the source
Difficulties and limits
- Findings on rate-execution and stage-level CFrac decomposition (including the ~17% rate-hold-quality figure) are explicitly marked 'Draft · pending data QC' — not yet finalized
Sources
- skyfrac.com — vendor self-report
Similar deployments
GE Aerospace
Independently reportedManufacturing · Process automation · Classical ML · Scaled
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
Amazon · Amazon
10% more — shopper spend
Independently reportedRetail & e-commerce · Process automation · Computer vision · Scaled · 2024
icare · SAS
increased by 15% — accuracy
Independently reportedInsurance · Process automation · Classical ML · Scaled · 2019