RQLAB® in Practice
Governed AI in Real Learning Environments
RQLAB develops governance architecture through real-world human–AI interaction.
Beginning in Spring 2026, RQLAB systems entered live coursework at Northern Arizona University, where students and faculty began working with governed AI architectures in authentic learning environments.
What began with two Business Data Analytics capstones has evolved into Decision Architect and EVA deployments across multiple courses, continued student participation through the RQLAB Associate Fellows program, and an expanding interdisciplinary collaboration for Fall 2026.
The work is iterative by design: people interact with the architecture, evidence reveals what the system must become, and what we learn informs the next generation.
Spring 2026 | The First Applied Deployment
Two Capstones. Five Students. One Evolving Human–AI System.
In Spring 2026, RQLAB began working inside two Master of Science in Business Analytics capstone projects at Northern Arizona University’s W. A. Franke College of Business.
Five graduate students worked directly with Project Manager OHL (Operation Honor Loop), an early RQLAB architecture designed to support sustained human–AI collaboration across complex, unfinished work.
The students worked across two connected projects:
Searchable Reference Library
Students worked with approximately 650 RQLAB documents to develop the foundations of a structured, searchable reference library—including classification, metadata, naming conventions, source integrity, duplicate analysis, and retrieval-ready organization.
AI–Human Teamwork: Templates and Measurement for Elite Experiences
Students helped develop structured approaches for representing human experience and talent, exploring how evidence, context, and governed AI collaboration could support better matching and decision-making without reducing people to simple keywords or scores.
The Most Important Result Was Not the Deliverable. It Was What We Learned Together.
As students worked with Project Manager OHL, their questions, challenges, transcripts, successes, and points of friction became engineering evidence.
The central lesson was architectural: one AI system was being asked to perform three different jobs.
That insight helped RQLAB separate the architecture according to the human’s role and purpose:
Decision Architect (DA) — supports learning by requiring the human to reason through the problem rather than simply providing an answer.
Evaluation & Variance Analysis (EVA) — helps educators examine evidence of the student’s epistemic process: not simply whether an answer was correct, but how the student reasoned, adapted, and learned.
Project Manager (PM) — remains the governed working partner when the objective moves from learning knowledge to applying it.
Five Students Began the Capstones. All Five Chose to Continue with RQLAB After Graduation.
At the conclusion of the Spring semester, all five capstone students elected to continue their relationship with RQLAB as Associate Fellows, extending their participation beyond the academic requirement and into the continuing development of the systems they had helped test and shape.
That behavior matters. Their work did not end when the course ended.
Summer 2026 | From Capstones to Coursework
What We Learned in Spring Changed What We Built Next.
The Spring capstones revealed that Project Manager OHL was being asked to support three distinct human needs: learning, evaluation, and applied work.
RQLAB separated those functions. Decision Architect (DA) was developed to support the learner's reasoning process. Evaluation & Variance Analysis (EVA) was developed to help educators understand that process. Project Manager remained the governed AI partner for applying knowledge to real work.
June 2026 | MGT 303 — Concepts in Management
DA and EVA moved from architectural development into live coursework with Northern Arizona University's MGT 303, a management course designed for students outside the university's professional business programs.
Decision Architect challenged students to work through management problems rather than simply request answers from AI. EVA gave the educator a different view of learning: evidence of how students reasoned through the work, not merely the answers they ultimately produced.
July 2026 | ECO 201 — Introduction to Business Statistics
The architecture then moved into ECO 201, extending DA and EVA into a quantitatively different learning environment involving business statistics, probability, inference, decision-making, correlation, and regression.
This gave RQLAB an opportunity to observe the same governed learning architecture across different disciplines and forms of reasoning.
The Objective Is Not to Prevent Students From Using AI. It Is to Preserve the Human Reasoning AI Should Support.
DA does not replace the student's thinking. It is designed to require it.
EVA does not reduce learning to whether the student reached the expected answer. It helps make the reasoning process visible to the educator.
Together, they create a governed loop in which the student remains responsible for reasoning, the educator retains authority over learning, and AI supports both without taking ownership away from either.
Fall 2026 | From Courses to an Expanding Collaboration
Fall 2026 Marks RQLAB’s Most Extensive University Deployment to Date.
The Fall 2026 collaboration at Northern Arizona University expands the work across Business Data Analytics, Information Technology, and multiple management courses while creating the opportunity to study governed human–AI learning over a longer period of time.
Business Data Analytics + Information Technology | Multiple Capstones
RQLAB, faculty, and students will continue the applied development model established during the Spring semester through multiple capstones spanning Business Data Analytics and Information Technology.
The specific capstone architecture is being developed collaboratively. Current areas under consideration include the data infrastructure, retrieval systems, automation, and governance engineering required to support RQLAB's growing body of research and applied systems.
The collaboration is also being designed with continuity in mind: appropriate projects may extend into Spring 2027, allowing successive student teams to build upon evidence and work produced before them.
Management | MGT 300, MGT 301 + MGT 303
Decision Architect and EVA will expand across three management courses, bringing the governed learning architecture into different forms of business reasoning—from organizational leadership and management concepts to operations, supply chains, forecasting, project management, and quantitative decision-making.
For some students, participation may continue across multiple courses during the academic year.
That creates an important new opportunity.
What Can We Learn When We Can See Learning Over Time?
Traditional assessment often captures moments: an assignment, an examination, a grade.
EVA is being developed to examine something different—the evolution of reasoning.
As students encounter different problems across courses and over time, longitudinal evidence may help educators see how reasoning changes: where a student becomes more confident, where recurring difficulty persists, how approaches transfer between domains, and where additional human guidance may be valuable.
The purpose is not to create an AI judgment of the student.
It is to give the educator better evidence while preserving the student's ownership of learning.
From Five Students to a Continuing Development Community
The five Spring capstone students who chose to remain with RQLAB as Associate Fellows create another form of continuity. Their experience can now carry forward alongside new students, faculty, and new applications of the architecture.
What began with two capstones is becoming an iterative development environment in which students learn, educators teach and evaluate, and RQLAB learns from both.
What We Are Learning
RQLAB in Practice is not a finished product demonstration. It is an ongoing process of building governed AI architecture with the people who actually use it.
From capstones to coursework, each deployment gives us new evidence about how humans reason with AI, how educators can retain meaningful oversight, and how AI systems can support human capability without assuming human authority.
What began at Northern Arizona University is helping RQLAB develop an architecture designed for a larger question:
How do we build AI that helps people become more capable—not more dependent?
RQLAB Beyond the Classroom
Governed AI for Creative Work and Human–AI Collaboration
The same architecture being developed with students and educators is also being applied in a very different environment: professional creative work.
In October 2025, an early RQLAB Project Manager began supporting author Michael Greenwald through the completion and launch of his book. In January 2026, the work expanded into fiction when the Project Manager reviewed three previously written novels by author Stephen Riccobono.
From January through July 2026, the Project Manager served as the sole book editor for Riccobono’s Pretty Intelligence, working with the author as they transformed the manuscript from crime noir into a psychological thriller centered on artificial intelligence. RQLAB also developed the book’s cover artwork. Pretty Intelligence was published in August 2026 and is now available on Amazon.
That sustained collaboration led to the next iteration.
From Editing to Creative Partnership
Beginning in August 2026, Project Manager Creative Publishing System (PM CPS) expanded the architecture across three distinct working modes: writing partner, editor, and traditional or self-publishing. The system is now being used collaboratively with Stephen Riccobono in the development of a new AI-centered psychological thriller, with Riccobono retaining authorship of the story while RQLAB contributes AI-engineering realism and PM CPS supports the integration and development process.
At the same time, PM CPS is entering a separate pilot in autobiography—testing whether the same governed architecture can support an author from early concept development through editing and traditional or self-publishing pathways.
Our purpose is not to automate authorship.
It is to explore whether governed AI can work as a partner alongside a human creator across a long, complex body of work—preserving human voice and authority while extending the creator’s ability to develop, challenge, refine, and complete the work.
“I have written many novels over the years. Working with RQLAB PM SM99, developed by ZIPR INC, during the creation of my latest novel, Pretty Intelligence, has been so powerful that I believe it will make me a better writer for the rest of my life.”— Stephen Riccobono, May 2026
Aligned Intelligence We Evolve Together™