Getting Started

First of all, it's important to note that no matter what AI and whatnot is used, it's highly adviced to recheck everything, especially if any AI-produced artifacts are going into production.

That said, AI can also be a good companion and a tool that cuts the costs and time for a variety of tasks. And Tazama might be the case when one would consider using AI for mundane tasks.

Is something wrong with Tazama?

Actually - no. Tazama is a hidden gem amongst open source projects that deserves much more spotlight than it has. It's one of a few open source AML systems that could rival closed-source ones. Despite it's relatively young age, Tazama had already won it's place for many banking systems across Africa and Asia, which also empowers it's rapid development in close alignment with actual market needs.

Tazama is a very feature-rich ecosystem of microservices, and the contracts between them make it very powerful and quite easy for development and understanding. Tazama organization also has a ton of various documentation, including even end-user manuals for compliance officers and other roles.

Keeping attention is hard

Software engineering is always about pros and cons. Whilst the System is simple and elegant in design, microservices are infamous for being hard to keep up with - for example, each rule in tazama is more or less would be a separate container, which should be wrapped into a rule executer, which should be registered for rule director to run. And whilst the rule system and it's execution follows quite straightforward path of execution and hierarchy, other core services are harder to look for - connection and case managements, configurations, schemas and mappings for adapters are all something developers need to keep in mind.

And that's without accounting any existing codebase that is to be integrated with Tazama, if there is a need for it (e.g., setting up integrations, extending case management capabilities to use existing data that requires tazama core modifications, etc.)

How do we resolve this?

To make our development less painful and easier to manage, we can use nextdocs to delegate the scaffolding and search in the codebase and documentation onto something that is far more user-friendly.

Nextdocs offers several ways of interaction with the codebase and documentation:

Thus, even if you don't like idea of using AI agents to do the work, you can still efficiently search throughout the codebase without spending hours skimming throughout hundreds of files by yourself or relentlessly using grep trying to capture the correct query to search.

Do you want to try?

Then, check out our guides and tutorials:

[there should be links soon]