
End-to-end complaints management platform
Sustained regulatory compliance and fewer escalated complaints. More than four years of continuity with ongoing maintenance and evolution; 2,440 hours in the 2023 ecosystem phase and expansion to 3 partners.
We automate operational processes end to end: RPA robots where the task repeats, AI where documents need interpreting, and ongoing maintenance and evolution when the process changes.
The process automation service from ACL (ACL powered by DataArt), a Chilean technology services company in business since 1991, identifies which tasks in a process are worth automating, chooses the right layer for each one (RPA robots, system integration, or AI to interpret documents), and puts it into production with monitoring, exception handling, and ongoing maintenance and evolution.
It decides which tasks to automate, with which layer, and how they keep running in production.
As an RPA and process automation company, ACL starts with an assessment that measures hours, errors, and rework, tests the solution on a narrowly scoped process, and adjusts it through ongoing maintenance and evolution when the process changes.
Classic RPA automates the repeatable task and breaks when the process changes. That is why the first step is choosing the layer, and sometimes the answer is that you don’t need a robot.
| Layer | When it fits | ACL example |
|---|---|---|
| RPA robot | The task repeats with clear rules on systems without an API, such as external portals or legacy applications | Regulated banking: automated integration with UiPath between a complaints platform and external entities |
| Integration in the source system | The process lives in your own systems, with access to the data. Automating at the source is more stable than a robot working on the screen | Payments: automated transaction reconciliation that cut rejected transactions by 90%, delivered in 60 business days |
| AI | Something needs interpreting: documents, free text, request classification, or natural-language responses | Analytics: an AI assistant for querying a document repository in natural language. We cover it in artificial intelligence for enterprises |
| AI + automation | AI reads the input (a document, a request) and the robot or integration carries out the next steps | Higher education: applicant evaluation with Power Automate and Azure OpenAI, with analyst review |
The robot is one piece. The service covers what it needs to keep running when the process changes.
We measure hours, errors, and rework in the current process and prioritize where automation pays off and how quickly, before we write a single line of code.
We propose the optimized model and the layer for each task, and test it on a scoped segment of the process.
We develop robots and integrations with exception handling. A QA analyst tests them before they go to production.
We deploy the automations on your infrastructure.
We adjust the automations when the process changes, as on the banking platform we have maintained for more than four years.
We recommend the platform during the assessment, based on your process and your licensing. These are the RPA and automation solutions ACL has implemented.
Automated integration with external entities on a bank’s complaints platform, with ongoing maintenance and evolution for more than four years.
They connect systems and AI services, as in the applicant evaluation with Azure OpenAI for a higher education institution.
When the data is available in the source system, we automate there, as in the transaction reconciliation for a payments company.
Two ACL projects in banking and payments.

Sustained regulatory compliance and fewer escalated complaints. More than four years of continuity with ongoing maintenance and evolution; 2,440 hours in the 2023 ecosystem phase and expansion to 3 partners.

−90% rejected transactions, by detecting gaps in the authorizer’s rules, and an end to daily manual reconciliation. Delivered in 60 business days, on time and on budget.
Learn about our work in software and data for banking and financial services.
Which process takes up the most hours? Let’s start by measuring it.
Map the processes you can automateEvery automation has to leave evidence of what it did and of who is accountable for it.
Owner and monitoringEvery automation has an owner, monitoring, and exception handling, so a failure doesn’t go unnoticed.
Audit evidenceReports and digital signature for audits, as in the automated reconciliation for a payments company.
Regulated processesExperience with regulated processes, such as a bank’s complaints management under legal deadlines.
Certified securityACL operates with an ISO/IEC 27001:2022 information security management system and an ISO 9001:2015 quality management system. See our ISO 9001 and ISO/IEC 27001 certifications.
Companies that trust ACL








It includes everything the automation needs to work in production: assessment, design, proof of concept, build with QA, go-live with exception handling, and ongoing maintenance and evolution.
At ACL, the assessment prioritizes which tasks to automate and how quickly they pay for themselves, and the design chooses the layer for each one: RPA, integration, or AI, depending on the case.
Start with high-volume processes that have clear rules, digital data, and a visible cost when they fail, such as reconciliations, data loads between systems, or back-office validations. That was the case with the transaction reconciliation ACL automated for a payments company.
A process that changes every month should be redesigned before it is automated.
ACL has implemented RPA robots with UiPath and low-code automation flows with Microsoft Power Automate.
If you haven’t chosen a platform yet, the recommendation is part of the assessment, based on your process and your licensing.
As a reference, ACL delivered the automated transaction reconciliation for a payments company in 60 business days, on time and on budget.
The timeline depends on how many systems the process touches, whether they expose APIs, and how many exceptions need handling. That is why the assessment defines the scope before we build.
They are adjusted through ongoing maintenance and evolution, so the change doesn’t stop operations. That is why we design every automation with an owner, monitoring, and exception handling.
At a regulated bank, the platform with RPA integration that we built has had more than four years of ongoing maintenance and evolution, and in 2023 it expanded to an ecosystem of three partners.
Use RPA when the task repeats with clear rules and structured data, such as loading data between systems or reconciling records. AI comes in when you need to interpret a document, classify free text, or draft a response.
In many processes they work together: AI reads and the robot executes in your systems. If your project centers on AI, see our artificial intelligence for enterprises service.
It is measured against a baseline taken before automating: the process’s labor hours, errors, rework, and cycle time. After go-live, the same metrics are compared, net of licensing and maintenance costs.
In ACL’s assessment, we estimate that return before we build.
If the process starts with documents, such as invoices, forms, or contracts, AI reads them and hands structured data to the robot.
If AI is at the center of the project, such as an assistant or a decision engine, we handle it in our artificial intelligence for enterprises service.
AI agents are already part of how we build software: with SPIREX®, our agentic orchestration framework, every stage of development passes through a human gate (Human-in-the-Middle).
Automation and RPARPA software in Chile: how to choose the right platformProcess, integration, operational governance, and support before comparing platforms.
Automation and RPAHow to calculate RPA ROI in Chile: metrics, costs, and real casesWhich metrics to use, which costs to consider, and how to prioritize processes.
Automation and RPAProcess automation: what to automate first and where RPA + AI fitWhat to automate first, what to avoid, and where RPA and AI add value.Together we measure volume, rules, the systems involved, and exceptions, and we choose the layer: robot, integration, or AI.
Map the processes you can automate