You know what work should be automated, and you know what steps need specific software to run that work. Sometimes, these steps depend on software that isn't part of your standard environment. That could be an internal software package (a bundle of code your team has already written and maintains), a command line tool your security team pre-approved, or an open source library that does the job really well.
Workflows, and ultimately the work itself, slow down when these fundamentals live outside your standard environment, and you’ve likely had to build a manual workaround or reinstall packages when a step needs one.
Give each step everything it needs to run
This slowdown is solved by running the step in an environment that already has your internal software and tools installed. That environment comes from a custom container image. It’s a package of the exact software and dependencies a step needs. Bring your own, from a public or private registry — a place where container images are stored so they can be pulled and reused. The step still works like any other: it takes inputs, runs in a sandbox, and passes results along.
Everything is packaged once, and your steps use it each time the workflow runs. The output is then passed to the next step, and the workflow keeps running.
A real-world example: automating fraud checks
Take a fraud team within a fintech organization. They score transactions for risk as part of a larger triage workflow. The team is required to run every transaction through their internal scoring package, so the result determines what should happen next.
Their workflow runs fine until it reaches the scoring step. Because the package isn’t part of the standard environment, someone on the fraud team needs to run the check separately and then manually copy the score into the workflow so it can continue to run.
With the scoring package installed in a container image, the scoring step runs inside that environment. It takes the transaction as input and outputs a risk score that determines the decision to approve, review, or block. The handoff between steps now happens much faster and without any manual effort.
Same controls, fewer workarounds
You already know which steps depend on your libraries, tools, and software. Run those steps in a custom container image to save your team's time and effort on repeatable tasks. Steps start faster and run on software your security team has already approved. And all the controls you’ve set for workflows like testing, reviewing, and access are built in and consistent.
There shouldn’t be manual blockers or workarounds between your team and the work you need to get done. Using a custom container image in Tines 3B puts what you need, where you need it.
Watch a short demo of the fraud example in action, or head to our docs to learn how to customize a step’s container image with its own Dockerfile.
