How Colombia’s largest bank deploys AI that regulators can actually trust

Regulated industries across Latin America and beyond are navigating a trust-centered AI inflection point — one defined less by raw capability and more by whether AI can be explained, audited and trusted at scale.

Financial institutions racing to adopt AI without governance guardrails are accumulating risk faster than value. Bancolombia S.A., Colombia’s largest bank with more than 36 million customers and about 50% of the country’s financial transactions flowing through it, chose a different path — embedding process orchestration, human oversight and full audit trails into every AI deployment from the start, according to Alejandro Arias (pictured, left), continuous value team leader and project manager at Bancolombia.

“For the bank, if you cannot explain a decision, you cannot involve AI in that decision,” Arias said. “Transparency is not negotiable for Bancolombia. Trust is not only a value, it’s a principle.”

Arias and Leonardo Vivas (left), senior field sales manager at Appian Corp., spoke with theCUBE’s Dave Vellante and Alison Kosik at Appian World 2026, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed trust-centered AI adoption, process automation and how Bancolombia is scaling document intelligence without sacrificing governance. (* Disclosure below.)

Trust-centered AI demands transparency and human-in-the-loop control

In highly regulated financial environments, process complexity and disconnected legacy systems are the primary barriers to safe AI deployment. Bancolombia tackled that reality head-on by consolidating fragmented architecture before introducing automation.

“At the beginning we found many challenges — very disparate databases and disconnected systems within the architecture,” Vivas said. “We put all the Appian platform into production using AI, [Robotic Process Automation] and all the features to work with humans to improve the speed and the processes of all documents — reducing time spent in manual processes, almost [by] 70%.”

The result was a fully operational document automation system built in roughly 13 to 14 weeks, now processing 80,000 documents a month with high accuracy rates. But the transformation required more than technology — it demanded cultural change and skill development for the operations team, Arias noted. The broader strategic vision is also a warning for banks that are rushing AI adoption without that foundation.

“Many financial institutions are trying to make a rush,” Arias said. “But the reality is that some of the banks that I know are making so [many] mistakes because of that, because they’re trying to tell everyone, ‘You need to use AI to do all the things.’ Appian and the way that we work with AI ensures that the documents, information and processes are [part of] the process itself, not outside. That protects, speeds and secures the way you use the information within the company.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of Appian World 2026:

(* Disclosure: TheCUBE is a paid media partner for Appian World. Sponsors of theCUBE’s event coverage do not have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE