Buddhi Labs
Intelligence → Systems → Outcomes

Forecast revenue. Quantify risk. Plan growth. Govern the decision.

Cohort Brain is a decision intelligence platform for banks and fintechs. It models consumer portfolios at the cohort level — projecting P&L, losses, and growth under multiple macro scenarios — so finance, risk, and acquisition teams can plan from the same model and defend the same numbers.

Start a Conversation Explore Products
P&L forecasting at the cohort level
Loss projections under macro stress
Board-ready outputs with full traceability
Flagship product
Cohort Brain
Simulate consumer portfolios at the cohort level — projecting revenue, losses, and growth under real macro conditions — so every function plans from the same model.
  • Revenue and P&L forecasting grounded in cohort behavior
  • Loss and delinquency projections by vintage, segment, and channel
  • Stress testing under multiple macro rate paths
  • Origination planning with downstream economics visibility
  • Variance explained through drivers — not just "actuals differed"
  • Board outputs traceable to the assumptions that produced them
Request a demo →

Cohort Brain

Cohort Brain (CB) is a decision intelligence platform for consumer financial portfolios. It projects P&L, credit losses, and portfolio growth at the cohort level — under multiple macro scenarios — using a roll-forward engine grounded in months-on-book behavioral dynamics. Today it supports credit cards, consumer deposits (Savings / HISA), and BNPL portfolios, with a simulation engine built to extend to additional products where risk and behavior unfold over time.

What this means: Finance gets a defensible forecast. Risk gets vintage-level loss visibility and guardrails. Growth gets origination plans connected to downstream economics. Leadership gets board-ready outputs — all from one model with shared, explicit assumptions.

The problem

Most institutions run three separate planning engines: Finance forecasts the P&L. Risk models losses. Growth plans acquisition. These models rarely speak the same language — until reconciliation time.

Finance
Forecasts live in spreadsheets or generic FP&A tools—slow to update and hard to explain.
Risk
Behavioral assumptions sit in separate models or black boxes—difficult to inspect, govern, or align on.
Growth
Acquisition plans evolve independently from risk capacity and funding realities—creating disconnects downstream.
Inputs are slow to change. What-if analysis is manual. Macro shifts require rebuilding models. Board decks are assembled from disconnected sources. Confidence in the numbers suffers — and coordination becomes negotiation instead of clarity.

This fragmentation makes it hard to answer critical questions:

  • What happens if we scale faster through a different channel mix?
  • How does risk evolve as cohorts season?
  • What assumptions drove this forecast—and what changed since last time?
  • How do different macro pathways (rates, unemployment, cost of funds) change outcomes?
  • Why do external outputs (board / investor materials) drift from internal models over time?
Cohort Brain replaces fragmented forecasting with a cohort-based roll-forward engine where assumptions are explicit, scenarios are comparable, and decisions are traceable.

Core capabilities

CB models the business cohort-by-cohort — projecting how revenue, losses, and growth evolve as customers age and macro conditions shift.

1) Cohort-based forecasting
Project exposure, revenue, and losses month-by-month — driven by behavioral curves for default, churn, utilization, paydown, and draw.
2) Behavioral risk modeling
Drivers are curves over time, segmented by risk band and acquisition channel—visible, reviewable, governable.
3) Acquisition planning
Plan origination by segment and channel — with visibility into how each cohort contributes to margin, loss, and exposure over time.
4) Scenarios & stress testing
Best/Base/Worst cases using transparent scenario overlays—adjust behavioral, economic, and external pathways (e.g., rates, cost of funds)—repeatable and comparable.
5) Decision-ready outputs
Board summaries, scenario comparisons, and variance attribution — outputs that explain what changed and why, not just what the numbers are.
6) Governance & traceability
Assumptions owned by functions, forecast runs logged, changes traceable over time—confidence in the numbers and the process.

Three products. Full portfolio economics.

Each product has its own behavioral mechanics, revenue drivers, and risk profile — but all run on the same cohort-based simulation engine, producing P&L, loss, and growth projections from shared assumptions.

Credit Cards
Revolving credit simulation across utilization, paydown, churn, default, and revenue drivers — modeled month-by-month by risk band and acquisition channel.
  • Interest, interchange, fees, losses, servicing
  • Risk guardrails and concentration limits
  • Channel-level acquisition and mix planning
Consumer Deposits (Savings / HISA)
Balance flow simulation across inflows, outflows, attrition, and rate sensitivity — segmented by balance tier and acquisition source.
  • Deposit growth and decay dynamics
  • Rate-path and funding sensitivity
  • Balance-tier mix planning
BNPL
Fixed-tenor installment modeling with deterministic amortization and behavioral overlays — built for short-duration cohort visibility and merchant economics.
  • Vintage curves by months-on-book
  • Default, prepayment, and late behavior
  • Merchant economics and take-rate tracking

What does Cohort Brain do for you?

Different leaders ask different questions. Cohort Brain gives each function the answer they need — revenue upside, downside protection, coordinated planning — from the same model and the same assumptions.

Portfolio Owners
Cards · Deposits · BNPL
Make growth-vs-risk trade-offs explicit — before locking a plan, not after.
  • Compare scenarios by loss-adjusted return, not just top-line growth
  • Pressure-test pricing, channel mix, and origination changes at the cohort level
  • Identify which segments and vintages are driving margin — and which are eroding it
  • Model macro rate paths and see their impact on portfolio economics, not just risk metrics
CFO / FP&A Leaders
Produce a forecast that finance, risk, and growth all recognize — without reconciling three models.
  • Eliminate version conflicts: one forecast engine, one set of assumptions, one output
  • Cut scenario cycle time from weeks to hours with pre-built macro pathways
  • Explain variance through drivers — volume, mix, pricing, behavior — not just "actuals came in different"
  • Generate board and investor outputs that trace directly to the assumptions that produced them
CRO / Risk Leaders
Strengthen downside governance without becoming a bottleneck to growth teams.
  • Run stress scenarios on the same model growth and finance use — comparable assumptions, traceable results
  • Enforce guardrails tied to risk appetite: loss ceilings, concentration limits, and first-payment default thresholds
  • Monitor vintage behavior by MOB, segment, and channel — not just portfolio averages
  • See how risk drivers propagate into P&L outcomes, not just standalone loss estimates
Acquisition / Growth Heads
Plan origination with full visibility into what each cohort will cost, earn, and lose over time.
  • See how channel and segment mix decisions flow through to losses, margin, and exposure over 12–60 months
  • Evaluate cohort-level unit economics and payback periods before committing budget
  • Stress-test growth plans against rising unemployment or rate changes — not just base-case assumptions
  • Coordinate with risk and finance on one shared plan instead of defending a separate model
One model. Same assumptions. Every function sees the same forecast — so the conversation moves from reconciliation to decision.

Credit Card Portfolio Modeling

Cohort Brain models revolving credit portfolios month-by-month, capturing utilization, paydown, churn, new draw, fraud, and default behavior across risk bands and acquisition channels.

Behavior Simulation

  • Months-on-book default and loss curves
  • Revolver vs transactor mix dynamics
  • Utilization and paydown behavior
  • Fraud and risk-band concentration guardrails

Economics & Planning

  • Interest income, interchange, fees
  • Credit loss forecasting
  • CAC and payback analysis by channel
  • Rewards impact on long-term unit economics

Consumer Deposit Modeling

Deposit portfolios are modeled through balance-tier segmentation and behavioral flow dynamics, enabling rate sensitivity analysis and funding planning under multiple macro paths.

Balance Dynamics

  • Inflow and withdrawal behavior curves
  • Attrition by balance tier
  • Rate elasticity modeling
  • Acquisition by tier and channel

Economics & Sensitivity

  • Interest expense modeling
  • Funding transfer pricing alignment
  • Net interest margin sensitivity
  • Macro rate-path scenario overlays

BNPL Installment Modeling

Designed for short-duration portfolios, BNPL modeling captures deterministic amortization with behavioral overlays for default, prepayment, and payment timing — enabling fast cohort insight.

Installment Simulation

  • 3, 6, and 12-month tenor modeling
  • Vintage performance tracking
  • First-payment default monitoring
  • Rapid cohort runoff visibility

Merchant Economics

  • Merchant discount revenue tracking
  • Average order value sensitivity
  • Take-rate analysis by merchant category
  • Origination cohort profitability mapping

Scenario & Macro Planning

Macro assumptions are first-class inputs. Interest rates, unemployment, funding costs, and stress overlays flow directly into behavioral curves and portfolio economics.

Multiple Macro Paths
Define base, upside, and stress scenarios with explicit rate paths and economic assumptions.
Behavioral Sensitivity
Default, churn, balance decay, and funding costs adjust under each macro scenario.
Side-by-Side Comparison
Compare exposure, revenue, and loss outcomes across scenarios with clear variance attribution.
Every scenario uses the same cohort engine. Only assumptions change — not the structure of the model.

Governance & Board Readiness

Cohort Brain is built for institutional environments where assumptions must be explicit, scenarios comparable, and outputs defensible.

Versioned Forecasts
Each forecast run stores its assumptions, acquisition plan, macro inputs, and results as a locked package.
Guardrails & Risk Limits
Risk appetite thresholds can be defined and evaluated before forecasts are finalized.
Traceable Assumptions
Every output can be traced back to the behavioral curves and inputs that produced it.
Designed to support executive reviews, board discussions, and institutional reporting workflows.

How teams use it

From assumptions to board-ready outputs in a single workflow — with every input explicit, every scenario comparable, and every result traceable to the drivers that produced it.

1
Select product and horizon
Choose credit cards, deposits, or BNPL. Define start period, forecast horizon, and macro scenarios.
2
Review assumptions
Inspect and adjust behavioral curves, scenario overlays, and external pathways where applicable.
3
Build the acquisition plan
Define inflows by segment and channel—plans directly create cohorts and drive outcomes.
4
Run the forecast
Preflight checks validate inputs, then the roll-forward engine simulates monthly portfolio evolution.
5
Analyze scenarios and drivers
Compare scenarios side-by-side and see which assumptions and drivers explain the deltas.
6
Decide and deliver
Lock the forecast. Generate board summaries with full assumption traceability. Every output links back to the behavioral curves, macro paths, and planning inputs that produced it.
Outputs: monthly P&L projections, loss forecasts, scenario comparisons, guardrail status, and driver attribution — all linked back to explicit assumptions.

Trust

Every assumption is owned. Every change is logged. Every output traces back to the inputs that produced it — so the forecast is defensible, not just directional.

1
Guardrails
Forecast outputs can be evaluated against defined thresholds and policies—flagged, blocked, or approved based on guardrail results.
2
Traceability
Forecast runs are logged and comparable over time, making it clear what changed, why it changed, and which assumptions drove the outcome.
3
Leadership & Boards
Outputs are decision-ready and grounded in documented assumptions — so external narratives stay linked to internal logic.
Built so teams can move from reconciliation to decision — with confidence in the numbers and the process.

Who it’s for

Built for banks and fintechs where cohort behavior drives portfolio economics — and where finance, risk, and growth need to plan from the same model.

Primary users

  • Executive teams (growth, capital, risk)
  • Finance leaders (forecasting + performance management)
  • Risk teams (credit and portfolio behavior)
  • Marketing leaders (acquisition strategy)

Stakeholders

  • Boards and investors seeking clarity and accountability
  • Cross-functional operators aligned on shared assumptions
Best fit: Banks, fintechs, and neobanks with credit card, deposit, or BNPL portfolios where cohort behavior materially drives P&L, losses, and growth outcomes.

How Cohort Brain is different

Not generic FP&A
FP&A tools forecast revenue. They don't model cohort-level credit behavior, vintage loss curves, or macro-driven risk dynamics.
Not a BI dashboard
Dashboards describe the past. Cohort Brain simulates the future — under stress, under growth, under different rate paths.
Not a siloed risk model
Risk models quantify loss. They don't connect to acquisition plans, P&L forecasts, or board outputs. Cohort Brain does.

Start a conversation

If your teams forecast revenue, model losses, and plan origination in separate models — and reconcile them quarterly — we should talk.

hello@buddhilabs.ca

No pitch decks. No sequences. Just a focused discussion.