Capabilities & Advantages

Multiple deployment options



We can deliver our engine as a:

  • PaaS (platform-as-a-service), running in your cloud, or in your on-premise infrastructure
  • SaaS (software-as-a-service), running fully in our AWS cloud

You can choose which deployment approach suits you the best:

  • Banks & fintechs usually prefer PaaS option. That ensures that they are in full control over their decisioning. We as a vendor have no access to the production & personal data of their customers
  • Some clients don't use sensitive data for their decisioning. Those clients might be ok with SaaS option & sending data to a vendor's cloud

Other decision engines

  • Traditional decision engines have been deployed to client's on-premise infrastructure. Cloud deployment hasn't been supported at all
  • Recently, many vendors have done a shift and they now offer SaaS decisioning only. Clients are sending sensitive data to the vendor's cloud, and the decision engine runs there

Use case flexibility



You can use TaranDM for any data-driven use case. It can help you with decisioning across the customer's lifecycle, such as:

  • credit risk decisioning (loan origination, credit scoring, fraud detection, …)
  • customer management (recommendations, offer personalization, cross-sell & up-sell, …)
  • collection strategies
  • and many others

Both real-time and batch decisioning are supported

Other decision engines

  • Most decision engines are built for a specific use case, or a set of use cases
  • The engine doesn't support additional use cases, or you have to buy another software solution from the vendor

Covering decisioning end-to-end



Covers any decisioning process in its entirety, including:

  • data integrations to other internal & external systems
  • data pre-processing & aggregations
  • decision flow management via no-code, web-based GUI (including data source calls)
  • out-of-the-box simulation capabilities for business logic changes
  • comprehensive reporting of the decisioning on all levels of granularity

Other decision engines

Don't have capabilities to cover all steps, for example:

  • clients have to build & maintain data source adapters by themselves
  • staging of data source calls cannot be configured within the engine
  • data pre-preprocessing has to be done by the client's IT/engineering team
  • no in-built simulation capability




No-code, web-based GUI, in which the business users can:

  • create & manage decisioning strategies (including data source calls)
  • push new or changed strategies from sandbox to production via a few clicks
  • set up A/B tests
  • run various simulations & view the results
  • access past decisions, reports & dashboards

Other decision engines

Often, some of the capabilities are:

  • not available via GUI at all (simulations, …) configured in other systems (data source calls, …)
  • more back-end driven (push from sandbox to production, A/B tests, …)

Fast and easy data integrations



Includes pre-built generic data source adapters, which:

  • can be used for fast and easy integrations of new data sources via API or SQL
  • have embedded caching functionality to enhance speed and reduce costs. That is useful, for example, to avoid repetitive queries to the credit bureau

Other decision engines

Either of the following scenarios usually happens:

  • No data source adapters included. Your IT team has to integrate all data sources and maintain the adapters in the long-term
  • Some data source adapters are available, but they are for specific data sources only. You cannot re-use them to connect to any new data sources

Data pre-processing capability



  • Can work directly with raw, large data, such as transaction data or credit bureau report
  • Data pre-processing & aggregations can be executed within the engine
  • There is little to no dependence on internal IT/engineering team

Other decision engines

  • Data pre-processing has to be done by IT/engineering team before the data enters the engine. That can take weeks to months
  • Often, it ends up with a mismatch between the initial business specification and its implementation

AI/ML models



  • Deployment of AI/ML models from analytical environment to production via one click
  • Native support of various:
    • model types (logistic regression, gradient boosting, neural networks, …)
    • model formats (Python export, PMML, csv, …)
  • Optional, Jupyter-based TaranDM Model Development Workflow for analytics & fully-/semi-automated model development

Other decision engines

  • Sometimes they support only logistic regression models
  • You need to manually encode predictor binning and transformations
  • You might need an external module for deployment of some AI/ML techniques

Embedded A/B testing framework



  • Provides an embedded A/B testing (Champion Challenger) framework
  • You can define Champion Challenger strategies and their sampling directly in the GUI
  • Champion Challenger flags are automatically propagated downstream to the reporting & dashboards

Other decision engines

  • Often, no out-of-the-box support for A/B testing
  • All such tests have to be reconciled and evaluated manually or in the back-end

In-built simulation tools



Multiple out-of-the-box simulation tools in the GUI:

  • You can simulate individual decisions, either by tweaking past decisions, or by creating new ones
  • When you change decisioning logic, you can simulate the impact of your changes in one click. You immediately see the impact of the change on approval rate, risk grades, and so on.

Other decision engines

  • Most other vendors don’t have any in-built simulation tools
  • Business users have to simulate & validate their changes outside the engine

Decision optimization



  • Includes Decision Optimizer block which clients use to solve difficult & complex business problems
  • The goal can be to maximize revenue or profit, subject to business constraints or KPIs
  • Full set of optimization methods covered, from linear, mixed-integer, non-linear to constrained optimization

Other decision engines

  • Most other vendors don’t have such capability
  • Very few vendors have it, but they offer it as a standalone software product. You need to license it & implement it separately

Decision reporting & dashboards



Automated reporting layer, which includes:

  • Aggregated information (number of decisions per use case/strategy, decision outcomes, …)
  • Full detail of every decision made
  • Usage of data sources
  • Evaluation of A/B tests
  • Simulation results
  • And more

Other decision engines

  • Usually less comprehensive reporting
  • Some reports have to be created manually or outside the engine

Modern technology stack & open standards



  • Containerized, microservices-based application
  • Open-source technologies, no proprietary languages or formats
  • Minimized vendor lock-in. Clients can do their own customizations

Other decision engines

  • Often use proprietary languages or formats (for example, when exporting a decisioning strategy)
  • Significant vendor lock-in. Customizations not possible, or they can be done only by the vendor, which is costly and takes time

Business logic re-usability



  • You can re-use any block of the decisioning logic (such as set of rules) across multiple use cases or strategies
  • No duplication or copying & pasting is needed
  • GUI conveniently shows in which use cases/strategies is the logic used
  • If you want to change a shared logic, you do it at a single place. The change is automatically propagated everywhere where the logic is used

Other decision engines

  • You cannot easily re-use decisioning logic across use cases or strategies
  • Using the same logic at multiple places requires duplication, copying & pasting. That makes the creation & maintenance of strategies more difficult and error-prone

Fraud detection



  • Optional Antifraud module to detect fraudulent loan applications
  • The module stores & automatically calculates concentration checks and cross checks to prevent fraud
  • Hundreds of pre-defined aggregations are included. Client can customizable further and add new aggregations

Other decision engines

  • Many vendors don't have any fraud detection capability
  • Some vendors have it, but they offer it as a standalone software product. You need to license it & implement it separately

Competitive pricing & favorable commercial terms



  • Thanks to its architectural & deployment design, TaranDM can be implemented in a short period of time
  • Client can cancel the contract anytime. We are confident about TaranDM. We know that clients like both the platform as well as the services we provide

Other decision engines

  • Implementations can take many months or even years
  • Clients have to commit into multi-year minimum license period. They can get stuck with the engine when they are not satisfied with the performance of the vendor

If you are interested in TaranDM demo, we will be happy to show you more

Let's Talk About You

We help our clients to automate their complex, data-driven decision.

You can contact us at or via this form.

icon Linkedin

Follow us on LinkedIn to get the latest news.