DeepSea Management Consultancy

For banks and lenders

Data and AI for banks, taken from idea to production.

DeepSea Management Consultancy is a founder-led advisory for banks and lenders. We design credit risk management end to end, build analytics teams and the governance around them, and take data, machine learning and agentic AI from strategy to production.

25+ years
in data, analytics and AI across global financial services
13 countries
covered by credit risk models overseen from one regional team
14 million
active bank customers served by predictive analytics
4–6 weeks
to deliver analytics that used to take 6–8 months

Figures from the founder's career in banking. Details below.

Services

Six areas where we work with banks and lenders

  • Credit risk management, end to end

    Risk models, application and behavioral scorecards, lending strategies and decision engines, and the tools that run them, across the credit life cycle from origination to collections.

  • Data and AI strategy

    Strategy, roadmap and investment priorities agreed with executive leadership, tied to the decisions the bank wants to improve.

  • Analytics and AI governance

    A governance framework for data, models and AI that the whole organization can follow, covering how analytics is built, approved and monitored.

  • Analytics teams, built from scratch

    The design and operating model for a central analytics team that serves every department, and training that turns business teams from users of analytics into builders.

  • Generative and agentic AI

    Agentic workflows on large language models that carry a piece of analytical work end to end, from raw data to a written recommendation.

  • AnalyticsOps and intelligent automation

    Automated data pipelines, model deployment and scheduled retraining, so that models reach production in weeks, plus automation with a decision layer between operational tasks.

Product

In development

RAROC Engine

A web-based pricing tool for credit products, built end to end on Claude. It prices each credit contract on its risk-adjusted return on capital (RAROC), taking into account every component that plays a role in the price.

Most banks today price to a target net spread built from a handful of inputs: a cost of capital and hurdle rate from Treasury, a single cost of risk from Risk Management or Collections, and a flat operating expense from Operations. These are bank-wide averages. Two loans with the same spread can consume very different amounts of capital, and the price does not show it. Low-risk customers end up overpaying and leave, while high-risk customers are underpriced and stay.

Pricing on RAROC asks for more: funding cost, expected loss, capital consumption, operating cost, fees, utilization and collateral, each calculated for the specific product or portfolio so that prices can differentiate. That is a change in how a bank works, and it stalls because every parameter belongs to a different department and those teams rarely sit together. RAROC Engine makes the change practical. A bank can start from the averages it has today, see which parameters move the price most, and refine those first.

RAROC Engine mock-up: parameter sets by department on the left, and RAROC, economic profit, break-even rate and an income breakdown on the right
  • One set of agreed parameters

    Each department maintains and approves only its own inputs, in its own time, with a full history of what changed.

  • An answer in seconds

    RAROC against the hurdle, economic profit, capital required and the break-even price for every deal.

  • What-if and scenarios

    See how the return moves with price, default risk, collateral or utilization, and compare pricing scenarios side by side.

  • Ask in plain language

    Claude answers what-if questions and writes the pricing commentary, while the numbers come from fixed, auditable formulas.

Open the mock-up Early mock-up with example data.

Approach

How an engagement runs

  1. Assess

    Your data, models, platform and team, with quick wins tested along the way.

  2. Prioritize

    A roadmap with a step-by-step implementation plan.

  3. Build to production

    Models, pipelines and AI workflows, deployed where decisions are made.

  4. Hand over

    Your teams trained to build on it themselves.

Founder

Hakan Yılmaz

Hakan is a former Chief Data & Analytics Officer of Yapı Kredi Bank and the founder of the first centralized Analytics Center of Excellence in Turkish banking, at Garanti BBVA. He has spent more than 25 years building credit risk, analytics and AI functions inside banks across Europe, the Middle East and Africa.

Contact

Tell us what you are trying to solve

A short note about your bank, your data and the decision you want to improve is enough to start.

Emailinfo@deepseaconsultancy.com

Phone+971 55 770 47 71

LinkedInlinkedin.com/in/hakanmaz

CompanyDEEPSEA MANAGEMENT CONSULTANCY - FZCO
Free zone company licensed by the Dubai Integrated Economic Zones Authority (IFZA), license no. 71995
Unit 001-71995, IFZA Business Park, Building A1, Dubai Digital Park, Dubai Silicon Oasis, Dubai, United Arab Emirates