Download the Readiness Checklist
RBI Model Risk Management Guidance June 2026

RBI wants your models to be fair and transparent for your customers

The first model governance update in 24 years applies to every bank, NBFC, and housing finance company from AI scoring engines to pricing spreadsheets.
Book a 15-Minute Readiness Review

Access the Readiness Checklist

How iTuring.ai is built for this

RBI wants your models to be fair, transparent and human-supervised by design.

iTuring.ai is fair, transparent, and human-supervised by design

Fair from day one

Business decisions taken by models shouldn't be discriminatory in any way.

Transparent at every stage

Requires defined explainability thresholds and enhanced controls, so explainability gaps don't disrupt your business.

Human in the loop

Requires human-in-the-loop arrangements, override and kill-switch mechanisms, and periodic human review of model-driven decisions.

Model Definition and Scope · Para 7(3)

Your spreadsheets and bureau rules are models now

RBI’s definition does not ask what your team calls a tool. It asks whether something takes an input, applies logic, and produces an output that shapes a business decision.
“…algorithms, analytics, decision-based rules, and computational tools which have a material impact on decision-making, irrespective of whether such tools are recognised as models by the RE [Regulated Entity].” – Para 7(3)

What this means for you

Pricing spreadsheets, underwriting grids, and VBA macros are all in scope.

Pricing spreadsheets, underwriting grids, and VBA macros are all in scope.

Institutions counting 20 models often find 150 to 300 once they apply Para 7(3).

Inventory Requirements · Para 23

Retire a model.
Keep ten years of records

Decommissioning a model does not end your obligation. Every retired model stays in your inventory for at least ten years, or until it stops serving as a backup or benchmark reference, whichever is later.
“Decommissioned models shall remain in the inventory for a period of at least ten years from the date of decommissioning or from when they cease to serve as a backup reference, whichever is later.” — Para 23

What this means for your institution

Removing records at retirement leaves a gap you cannot recover at examination.

Every retired model needs full records maintained for the entire retention period.

Ten years runs from decommissioning or from when the model stops serving as a reference, whichever is later.

Vendor Risk and Accountability · Para 45 and Para 46

Your vendor's model certificate does not cover you

Every Regulated Entity (RE) that acquires or relies on a third-party model remains fully accountable for its outcomes. The vendor’s validation certificate carries no weight under this guidance.

“An RE shall remain fully accountable for all outputs and decisions generated by third-party models…notwithstanding any validation, certification, or assurance provided by the third-party provider.” — Para 45

What this means for you

Your institution must independently validate all third-party models

Contracts must give you access to model methodology, data lineage, and validation records.

An unannounced vendor update that changes model outputs restarts your validation and Board/RMCB notification obligations.

These are 3 of 64 clauses. Every chapter of the guidance carries obligations your institution needs to map against its current model stack.

Proven in enterprise financial institutions

18 financial institutions run their model governance on iTuring.ai

Manual processes leave gaps in inventories, validation, drift detection, and Board reporting that surface at examination. iTuring replaces the scramble with continuous, automated governance.
Fundi
NCB
indostar
Munich Re
Aadhar Housing Finance

Two paths, one platform

Built to automate model risk and ML governance

iTuring MRM

For risk, audit, and compliance teams

  • Relevancy Score gives early warning before performance drops.
  • Three-layer drift analysis runs across model, feature, and data simultaneously.
  • Pinpoints the cause: data drift, feature degradation, or accuracy drop.
  • Generates a full investigation with board-ready evidence in minutes.
See the MRM Platform

iTuring Model Gov

For analytics, ML, and engineering teams

  • Register models built in iTuring, in Python or SAS, or sourced from vendors.
  • Maker-checker approvals with accountability chains.
  • Complete audit trails: lineage, approvals, change history.
  • Access code, data, documents, models, and lifecycle status across your portfolio.
See the Model Gov Platform

Ongoing Monitoring · Para 54 and Para 56

Fixing bias once is not enough

Every time your model ingests new data, the risk of bias grows. RBI requires continuous monitoring, not a one-time review at deployment.

Para 54 requires institutions to identify and address bias and discriminatory outputs on an ongoing basis.

Para 56 sets enhanced controls and more frequent monitoring for models with dynamic or automatic updates.

A model that was clean at deployment may be biased today.

Your compliance obligation does not end at go-live.

FAQs

What compliance teams ask first

Does this apply to all NBFCs and banks?
Yes, to every Regulated Entity (RE), from commercial and co-operative banks to NBFCs of all sizes. Requirements scale with risk tier, but core governance rules apply to all.
Yes. Any spreadsheet, cutoff rule, underwriting grid, or macro that turns inputs into a business decision counts as a model, whatever your team calls it.
No. Your institution stays fully accountable for third-party outcomes. A vendor certificate does not replace independent validation. You validate every vendor model yourself, before and after deployment.
At least ten years from decommissioning, or from when the model stops serving as a backup reference, whichever is later.
Using an uninventoried model is a direct compliance breach under Para 21.
Customer-facing AI must disclose it is an AI system on first contact, offer a path to a human agent, and support an immediate kill-switch. Para 59 sets these requirements.
It approves the Model Risk Management Framework, reviews model tiering and validation reports, and signs off on policy exceptions. Reports reach the Board within three months of completion.

Ready to see iTuring in action?

Get a free readiness review, identify gaps, and see what compliance looks like at scale.
Book a 15-Minute Readiness Review

© 2026-27 iTuring.ai — All rights reserved

Tarika Bhutani

Senior Director – Sales and Marketing Operations

Tarika is a market development leader driving global growth through strategic partnerships and go-to-market initiatives.

 

She focuses on expanding enterprise adoption of AI solutions across international markets, working closely with partners and clients to enable data-driven transformation.

 

Her work centres on scaling enterprise AI through partner-led growth and direct customer engagement, supporting organisations in implementing impactful, data-driven solutions worldwide.

Vipin Johnson

Vice President – Customer Acquisition

Description Goes Here

Rajnish Ranjan

Vice President, Head – Data Science

Rajnish brings over two decades of experience leading data-driven transformation across Fortune 500 organisations.

 

His career spans senior roles at HSBC, Zafin, Cisco, TCS, Nielsen, iQuanti, Symphony, Supervalu, and Harman, delivering measurable cost savings, operational efficiencies, and revenue growth.

 

With experience across banking, retail, telecom, pharma, CPG, and digital marketing, he leads cross-functional teams at iTuring.ai to deliver advanced analytics, machine learning, and AI solutions.

Aishwarya Hegde

VP Operations & Content Head

Aishwarya has been instrumental in building iTuring.ai from inception and continues to manage core operations across the organisation. Her responsibilities span project operations, financial planning, and evaluating future expansion opportunities.

 

Prior to iTuring.ai, she worked with Market Probe and WNS Research & Analytics, delivering high-impact decision support and actionable analytics for IBM with a record of zero errors.

 

Aishwarya holds a postgraduate degree in Data Science and Machine Learning from Manipal University.

Bryan McLachlan

Managing Director – Africa

Bryan has 30 years of experience driving innovation and growth across technology, banking, insurance, and retail.

 

Prior to iTuring.ai, he held executive leadership roles at Instant Life, AIG, Nedbank, FNB, and TransUnion. He focuses on enabling enterprises to adopt AI and machine learning within trusted, governed, and risk-managed frameworks.

 

Bryan holds a Master’s degree in Commerce from the University of Johannesburg.

Mohammed Nawas M P

Co-Founder, VP Product Development

Nawas brings 20 years of experience in designing and delivering cloud-native software and data systems. He has held senior technology roles at HCL, Radisys, Kyocera, and Mindtree, leading large development teams and complex product builds.

 

At iTuring.ai, he oversees product roadmap and customer delivery, applying cloud-first thinking, deep systems expertise, and a focus on building robust, scalable AI solutions that challenge industry norms.

 

He is a graduate of Rajiv Gandhi Institute of Technology.

Amit Kumar
Co-Founder, VP Product Engineering

Amit is a technology architect with over 18 years of experience designing data-intensive systems and enterprise analytics platforms. He has built highly scalable products across open architecture models and virtualised infrastructure, aligning deep technical detail with business requirements for AI and ML solutions.

 

Prior to iTuring.ai, he held senior technical roles at Radisys and Aricent. Amit leads platform architecture with a focus on governance, lineage, and traceability.

 

He holds a First Class with Distinction BTech in Computer Science from Cochin University.

Valsan Ponnachath

President, COO and Co-founder

Valsan brings over two decades of global leadership across sales, professional services, and product operations in technology and SaaS enterprises.

 

Prior to iTuring.ai, he held senior executive roles at Fiserv, Cisco, and Sun Microsystems, most recently serving as Senior Vice President at Fiserv overseeing global system integration and international professional services. Based in California, he leads iTuring.ai’s growth in the Americas.

 

Valsan holds an MBA from the University of Nebraska and a BE in Computer Science from Bangalore University.

Suman Singh

Founder & CEO

Before founding iTuring.ai in 2018, Suman led analytics at Zafin and Fiserv as CAO and General Manager Analytics, delivering enterprise-scale solutions still running in production.

 

His work includes fraud detection systems saving clients over $19M, patented Customer Relationship Score methodology, and price optimisation recognised by the INFORMS Edelman Award (2014). He has authored multiple research papers and pioneered the data-to-value approach.

 

Suman holds a Master’s in Statistics from CCS HAU and a Bachelor’s in Agricultural Engineering from BHU.