Discover how DHL Group increased connectivity across the enterprise and automated multiple processes end-to-end
AI is transforming the way organizations do business. But while AI is certainly boosting productivity, it isn’t foolproof: AI can still have contextual misunderstandings, produce biased results or even hallucinate to create false answers if there aren’t controls in place to prevent it.
This calls for guard rails and human intervention to ensure accuracy and build trust in this evolving technology.
Bizagi’s CEO, Gustavo Gomez, is a big advocate of AI. He oversees the development of AI capabilities at Bizagi and actively promotes the benefits of AI to other enterprise leaders. Gustavo encourages organizations to look beyond the hype and carefully consider how they are deploying the technology; including promoting cultural change that reinforces how to use AI properly and reassuring employees that AI is not stealing their jobs, but enhancing them.
This means ensuring humans retain their important role in tasks supported by AI. “Would you allow an airport to be run by AI today? No, no, no! Because they hallucinate, they make mistakes, they invent stuff. Even though the reliability is increasing dramatically, there are certain processes where error is not allowed,” Bizagi CEO, Gustavo Gomez said on the Analysis.Tech podcast.
But does that mean AI cannot improve critical processes? Absolutely not. Many organizations are using AI to make operational decisions that are then checked by a human.
“What we are seeing is what we call the ‘maker-checker paradigm’ where AI creates a summary of a legal case, but then the subject matter expert takes the input and checks it is, and it becomes official. So, there is a symbiotic collaboration between the AI and the human validation: the expert validation… It will make us hugely more productive and the capabilities for automation around the AI are enormous and we're just trying is starting to scratch the surface right now.”
Process automation ensures a structured and systematic approach to implementing the maker-checker model. By embedding AI outputs into predefined workflows, organizations can create seamless collaboration between AI systems and human supervisors.
Mapping out the process first will standardize workflows so that once AI has completed its task, it automatically assigns the ‘checking’ part of the task to human employees. This ensures there is always a clear handoff between AI and the checker.
This also reduces errors because a human is always assigned to check the work completed by the AI to ensure there are no hallucinations or biased results. This then promotes continuous improvement, as errors can be fed back to AI to alter future actions and prevent the same mistake from reoccurring, further streamlining the process.
Process automation also creates accountability as every decision and correction is documented within the process. This transparency is critical in regulated industries like finance, healthcare, and law, where accountability and audits are a necessity.
You can use process automation to ensure the maker-checker paradigm when using AI Agents. Here we are using the example of a customer complaint to a bank:
Prior to using AI, once the bank receives the complaint email from a customer, it is ingested by an RPA bot and a complaint case is initiated in Bizagi. The process includes two humans: a Data Analyst to read the complaint, categorize it, and obfuscate any private information; this takes about 10 minutes. The Analyst then passes on the obfuscated complaint to the Lending Manager to read, and create and send a response, which would take about 20 minutes: half an hour in total.
With the addition of an AI agent, Bizagi passes the complaint email over to an AI Agent to obfuscate the data and generate a new email that hides the private information. This is passed to a second AI Agent to read the obfuscated email and generate a suggested email response. These are the ‘makers’. The email response is then passed to the Lending Manager, the ‘checker’ who reviews the suggested email from the AI Agent, makes any necessary changes, and sends it to the customer. This will only take a matter of minutes, saving the bank significant time and money.
The maker-checker process can be seen in action at financial services company FinTrU. As part of the Know Your Customer (KYC) process within their Intelligent Document Processing product, TrU Label, FinTrU created a maker-checker paradigm. AI does the brunt work of scouring through the document, then a human checks for accuracy and provides feedback. The human annotator or ‘checker’ effectively marks the homework of the AI. By highlighting and correcting the AI’s errors, it can also learn and adapt for future scenarios. In some instances, it also passes the work on to a QC to ensure thorough evaluation.
The sensitive nature of financial governance procedures such as KYC, means any tech-based solutions must be accurate, which is why there has been industry-wide hesitation to rely on AI technology. But by using this maker-checker paradigm, you still have the reassurance of a human checking the processes, but much of the up-front effort is taken out. As the AI becomes smarter and more advanced, the QC will reduce over time. The end goal is a straight-through process without human intervention, but this is dependent on future AI developments.
“What we’re finding here in this maker-checker paradigm, is how do we shrink the maker checker need over time to get straight through processing? ” explains FinTrU’s head of product, Steven Hewlett-Light.
“Everybody tries to get to the end game of straight-through processing straight away. But the objective is to really reduce the QC time more and more over time, so as your AI gets smarter, your QC time reduces.”
Watch the full webinar, Enabling Change to Support Financial Institutions: How FinTrU Transformed IDP and KYC, where Steven Hewlett-Light explains how he and his team are using AI to support vital financial compliance processes.
The concept of maker-checker is already proving its worth in industries like financial services, legal operations, and data-driven reporting.
Process automation helps organizations to harness the full potential of AI without sacrificing accuracy or trust. It paves the way for collaborative working between humans and AI, creating more value together than either could achieve alone.
As Bizagi CEO, Gustavo Gomez, says: “It will make us hugely more productive.” Process automation isn’t just the facilitator for the maker-checker paradigm, it’s the engine driving collaboration at the heart of modern enterprise.
