From Spreadsheets to Strategic Finance: How BANTRAB Built Advanced Profitability and Cost Management on Oracle, Accelerated by AI

Jun 20

At a glance

Client	Banco de los Trabajadores (BANTRAB), Guatemala
Sector	Banking / Financial Services
Partner	Asher & Company
Engagement since	2021
Technology	Oracle Enterprise Profitability & Cost Management (EPCM)
Methodology	Decision-to-Value™ (D2V™)
AI layer	Sáasil — from early ChatGPT experimentation to Claude in production
Headline outcome	Model build cycle cut from 16 weeks to 10; decision processes that took up to 4 weeks now run in one.

The challenge

Every bank can tell you whether it is compliant. Far fewer can tell you, with rigor, where the institution actually makes and loses money — by product, by channel, by client segment, by activity. Regulatory reporting answers “are we in order?” It almost never answers the question that drives strategy: “where do we act?”

Closing that gap takes more than a report. It takes an advanced profitability and cost management model — one that traces costs through activities to the products and segments that drive them, holds up under scrutiny, and keeps evolving as the business does.

Ambition was never the constraint. The constraint was the distance between the people who understand the business and the technical architecture that has to encode it. The bank’s logic lives in the heads of its financial experts; the Oracle platform needs that logic expressed precisely, consistently, and fast enough to keep pace with decisions. That distance — model iteration after model iteration — is where most cost-and-profitability programs lose momentum and stall.

The solution

Since 2021, Asher & Company has worked alongside BANTRAB to design and build these models on Oracle EPCM, anchored in Asher’s proprietary Decision-to-Value™ (D2V™) methodology — a disciplined path from framing the business question, to modeling it, to evaluating it, to putting it in front of the people who decide.

The work is built so the model is never a black box. Cost flows, activity logic, and allocation rules are made explicit and auditable inside the Oracle architecture. That discipline is exactly what lets a profitability model survive contact with a bank’s reality: regulatory scrutiny, organizational change, and a standing demand for traceability.

The AI layer: Sáasil

When the AI wave arrived, Asher and BANTRAB did what disciplined teams do — they experimented before they committed. The first exploration ran on ChatGPT. Over a year ago the work moved to Claude, which became the foundation of Sáasil: an AI layer that sits between the financial experts and the application.

Sáasil does not replace the expert or the model. Its job is to collapse the distance between them. In practice, it:

Accelerates communication with impact — translating between the language of the business and the precision the Oracle architecture demands, so intent becomes implementation faster.
Helps model the business — reasoning through cost and profitability logic alongside the team, surfacing edge cases and inconsistencies before they reach production.
Prepares relevant data — structuring and readying the inputs the Oracle EPCM architecture needs, cutting the manual handling that slows every iteration.

Choosing Claude was deliberate for a regulated-banking context, where accuracy, reliability, and auditability are non-negotiable — the qualities that matter most the moment an AI layer touches financial logic.

Results

37%

Faster model deployment

75%

Faster decision cycles

64%

Less manual data prep

37% faster model deployment — reducing profitability model build cycles from 16 weeks to 10 weeks, enabling the bank to respond more quickly to new business questions and organizational changes.
75% faster decision support cycles — shortening the time required to move from a business question to a validated, model-backed answer from up to four weeks to one week.
64% reduction in manual data preparation effort — significantly decreasing time spent collecting, validating, and preparing inputs for profitability and cost analyses.
Greater transparency and traceability — providing a clear, auditable view of how costs flow across activities, products, channels, and customer segments within the Oracle EPCM architecture.
Enhanced multidimensional profitability visibility — enabling management to evaluate performance across customers, products, channels, activities, and organizational units using a consistent profitability framework.

In their words

“For a long time the hard part wasn’t knowing where the bank made money — it was getting that knowledge into a model fast enough to act on it. Working with Asher on Oracle EPCM, with Sáasil in the loop, the time between a question and a defensible answer has dropped sharply. The model stopped being something we wait on. It’s an instrument we use.”

Edy Campa, Head of Profitability (Responsable de Rentabilidad), BANTRAB

What’s next

The collaboration continues to deepen the profitability and cost management models on Oracle EPCM, with Sáasil expanding the team’s capacity to model, validate, and act — turning financial architecture into a faster instrument for decisions.

See how Asher can bring Decision-to-Value™ to your finance function. Talk to our team →

Peter San Martin