Archeiothiki participated in the 4th GenAI Summit, held on 24 November at the Stavros Niarchos Foundation Cultural Center, contributing to a high-level discussion on return on investment, scalability, and governance in Generative AI adoption.
As part of “CEO Panel 2: GenAI ROI, Investments, Expectations”, Andreas Papadakis, Co-Founder of Archeiothiki, reflected on the gap between AI experimentation and measurable business value, highlighting a clear shift across the industry toward practicality.
“Value lies in taking boring high-volume tasks and handling them beautifully.”
“There is a delusion, because returns have not occurred. But, we found returns. We found huge returns, because we deal with streamlining of processes. Safe return on investment is clearly with streamlining,” Andreas noted, adding that “there is a shift to practicality”.
He emphasized that real value in enterprise AI lies in automating high-volume, repetitive processes at scale:
“Value lies in taking boring, high-volume tasks and handling them beautifully. A few weeks ago, we handled a project that was aiming at automating a process with two million documents, and this was to be delivered in three weeks”.
During the panel, Andreas made special reference to Archeiothiki’s proprietary Intelligent Document Processing platform, AI Flows, outlining its role within enterprise workflows:
“It ingests information. You start from unstructured information, you structure it, and AI proposes what to do”.
Addressing governance and compliance, Andreas underlined the continued importance of human oversight in AI-driven systems:
“We are not GenAI fanatics. We believe that human governance and human-in-the-loop are still important. It’s important for compliance purposes. We need clear rules for governance reasons”.
Engaging with industry leaders including Peggy Antonakou (Google), George Tsarouchas (Dialectica), and Tey Bannerman (Former McKinsey Partner), the discussion explored why 46% of enterprise GenAI initiatives were shut down last year before scaling and what ultimately defines success.
“For the past five years, everybody was experimenting with AI, on how to generate revenue, trying to do fancy things. Returns have not occurred. But, we found returns. We found huge returns, because we deal with streamlining of processes. Safe return on investment lies clearly in process streamlining and cost-cutting rather than in revenue generation”.
Challenging conventional success metrics, Andreas concluded:
“It may take ages to streamline a process at 99.9% whereas you can streamline at the same time dozens of processes at 92%. To that respect, the ultimate metric will not be not accuracy, but how many problems we are solving today that we couldn’t solve 12 months ago, and in volumes. ‘Success will be measured in volumes of transactions handled, not in accuracy’”.
