aiqbench analysis
This article addresses the critical need for deterministic decision-making and accountability within enterprise agentic architectures, proposing a hybrid approach that combines Large Language Models with formal DMN decision models. By integrating guardrails and agent skills, this framework allows organizations to maintain governance over high-stakes logic while leveraging AI capabilities, directly tackling the issue of non-deterministic outputs in production environments.
Presentation: Decision Models in Agentic Architectures: From Production to Agent Skills
Alex Porcelli discusses the critical gap in enterprise AI: non-deterministic output and lack of accountability in high-stakes decisions. He shares how integrating DMN decision models with LLMs, agent skills, and NeMo guardrails creates auditable, deterministic agentic architectures - allowing business leaders to own decision logic while engineers maintain robust architectural governance. By Alex Porcelli
Alex Porcelli: I want to start this and take a moment to reflect what we as a community of IT have been through the last 12 months. Actually, even less than 12 months. If you had asked me my opinions about AI in the fall last year, it would be completely different to what it is today. Back in the days, not long ago of course, I was using AI, being productive using AI, but it was mostly doing summarization, content writing, scripts, tests, configuration files. I used to be proud, "But AI does not touch my backend code. That is running critical systems for my clients. That is on me, not AI." Something magical happened during the holidays. It was super smart from the labs, new models came out. With the downtime, I had the opportunity to explore the technology more. I believe some of you had a similar experience. The models were not that powerful as those today, but I quickly realized that has the potential to change my work, my day by day. After 30 years doing software development, beginning of this year, I started coding differently. I do not type anymore. I always interact with my…
Excerpt shown under the publisher's RSS terms · 200 words · full text and figures are on the publisher's site
Related coverage
[AINews] Meta Muse Wonder & Wonder Full: MSL’s Claude 5.5-tier Frontier Model
Latent Space · Oct 13, 2026
Presentation: Autonomous Data Products for the Autonomous Era: Rethinking Data Architecture for GenAI
InfoQ · Jul 24, 2026
[AINews] High Taste Testers Ep. 1
Latent Space · Oct 17, 2026
Presentation: From Retrieval to Reasoning: Building Production-Ready Agentic AI Systems with Knowledge Graphs
InfoQ · Sep 12, 2026