Human + AI: Shifting the Paradigm in Modern Business Operations

In today’s rapidly evolving corporate landscape, knowing how to work with artificial intelligence (AI) isn't just a resume builder, it’s a core business competency. Inside the top-ranked MBA programs at Pepperdine Graziadio Business School, students aren't just reading about the future of technology. They are actively engineering it.

In ISTM 619: Technology and Operations Management, a cornerstone course of the MBA Information Systems curriculum taught by Erik Krogh, PhD, students recently tackled a hands-on AI Agent project. The assignment challenged them to select a business discipline, partner with an advanced AI tool, and complete a complex analytical task. Moving far beyond basic prompting, these future leaders designed multi-step agent workflows, critically audited AI outputs for hallucinations, and re-evaluated the division of labor between humans and machines.

The course revealed a fundamental truth about the future of work: The true power of AI lies not in replacing humans, but in elevating the uniquely human skills, such as critical judgment, ethics, and local intuition, that a machine simply cannot replicate.

Two Full-Time MBA students specializing in finance recently shared their experiences from the front lines of the AI frontier.

Aziz Abdurahimov (graduating November 2026): Focused on corporate finance and emerging markets powered by Anthropic Claude.

Kyle Pientka (graduating August 2026): Centered his research on commercial real estate analysis powered by OpenAI ChatGPT.

Mastering the Workflow: How Do You Move Beyond Basic Prompt Engineering?

The project required students to move past simple questions and instead engineer complex, contextual workflows. Both students quickly learned that an AI is only as good as the parameters a human sets for it.

Context Engineering vs. Prompt Engineering

For Aziz Abdurahimov, the breakthrough came when he realized that vague prompts lead directly to AI fabrications.

Key Takeaway: "Claude doesn't hallucinate because it's dumb. It hallucinating because vague prompts force it to fill gaps." — Aziz Abdurahimov

Aziz pivoted his strategy toward context engineering rather than just prompt engineering. He restructured his queries to give the AI a specific professional persona:

"I told Claude: 'You're conducting due diligence for a PE firm.' Nothing else changed in the analysis, but the tone shifted. It became more skeptical, risk-focused, and professional. The role itself triggered better reasoning. For the financial model, I added explicit meta-instructions: 'Flag any assumptions that materially impact valuation.' Suddenly Claude started critiquing itself. By Stage 5, I wasn't designing better prompts. I was designing better contexts."

Data Integration in Real Estate Analysis

Meanwhile, Kyle Pientka targeted the commercial real estate sector, looking for ways to streamline how analysts review potential property opportunities. Kyle chose ChatGPT's paid tier due to its robust ability to integrate with Microsoft formats and run specialized data tools.

"My prompt engineering strategy was based upon what I had learned throughout the term," Kyle shared. "I understood I needed to be able to verify accuracy of the results, utilize different AI platforms, and to be as specific as possible wherever possible. I based my prompt structure by utilizing long and complex sentences to ensure that the AI platforms would work with my concept properly and not create hallucinations."

Playing the "AI Auditor": Fact-Checking the Machine

A major part of the curriculum involves teaching students to become rigorous AI auditors. Because large language models are programmed to sound confident, they often present false information with total conviction.

Aziz found that while Claude was exceptional at structuring financial frameworks and mapping out ESG parameters, it fell short on strict factual compliance.

"Claude assigned a 'BB+ / Medium-High Risk' rating to a fictional company with full supporting analysis, which was completely invented," Aziz warned. "It also generated credible-looking market percentages for Vietnam's capacity that I couldn't verify. Claude had filled gaps with plausible fabrications. The dangerous part? It looked professional doing it. That’s what made the project valuable, not just that I caught errors, but that I developed a fact-checking reflex I'll apply to every AI output going forward."

The Limitations of LLMs with Mathematics

Kyle encountered a different, yet equally vital limitation: basic arithmetic errors. While ChatGPT excelled at generating working formulas inside Excel documents via Python backend scripting, it faltered when relying on pure language patterns for math.

"I’ve seen outputs that should be basic mathematical calculator functions that it answers incorrectly because it derives its data from the data source and selects one popular answer rather than doing the actual calculation," Kyle noted.

Expert Insight on AI Reliability: Very few generative AI tools will stay effective without an adequate amount of RLHF (Reinforcement Learning from Human Feedback) to verify outputs on a consistent basis.

The New Division of Labor: Where Human Soft Skills Triumph

The project ultimately forced students to rethink the future division of labor. Aziz and Kyle both started the assignment assuming AI would handle routine administrative tasks while humans managed the strategic thinking. They both walked away with an entirely different perspective.

The Quantitative vs. Qualitative Paradox in AI

Aziz discovered that the machine actually handled complex quantitative equations perfectly, but completely lacked the qualitative skepticism necessary for real business strategy.

"I expected Claude to struggle with quantitative work and excel at qualitative judgment. I had it backwards," Aziz said. "The financial calculations were mathematically perfect. But the investment thesis? It sounded compelling on the surface but lacked teeth. It didn't ask the obvious follow-up questions about market curtailment issues.

AI has absorbed procedural knowledge far better than critical thinking patterns. An analyst who just knows Excel is done. But critical judgment, professional skepticism, and the ability to read between the lines? That's precisely what AI can't do."

Protecting the Core and Leading Change

Kyle approached the division of labor from an organizational management and security perspective. He realized that implementing AI isn't just about speed; it requires massive structural oversight and a human-centric approach to risk.

"I am a risk-averse individual and worried a lot about cybersecurity for myself, my colleagues, and my organizational data," Kyle remarked. "I learned how to protect my use of these AI agents to set up privacy regulations to avoid data being uploaded into the LLMs and available to the public.

When you start implementing a new system, you need to be specific and adaptable throughout your organizations’ Strategy, Structure and Culture, Technology, and Individual Roles. Without the employee aspect, it’s a recipe for large mistakes if those utilizing the tools do not have the proper knowledge to verify the process."

Value for the Post-MBA Career and the Pepperdine Advantage

As both students prepare to transition into leadership roles post-graduation, their time in Dr. Krogh’s classroom has given them an undeniable edge. They aren't afraid of AI replacing them because they know exactly how to guide the machine.

Aziz’s Blueprint: Build Before You Are Ready

For Aziz, the agility of the Pepperdine Graziadio community provided the perfect ecosystem to take his classroom knowledge and immediately build real-world tools, leading him to create projects like PepperBot, STOCK PRO, and Folio.

"My number one piece of advice? Build something before you're ready," Aziz urged. "The reason Pepperdine is the ideal place is because you can actually execute. At a mega-university, you're anonymous. At Pepperdine, you're visible. Faculty actually know you, and there is an underlying ethos about ethics and impact. It's not just 'How do we use AI?' It's 'Should we use AI this way? What are we responsible for?'"

Kyle’s Blueprint: Never Stop Learning

For Kyle, the intimate campus environment fosters a community of continuous adaptation, which is the ultimate safeguard against obsolescence.

"Ignorance of change is the antithesis of self-improvement," Kyle stated. "Pepperdine is not the largest campus, it is intimate, it is a community that supports you from all angles from your peers, your staff, your professors, and even those who haven’t met you yet. We continue to push the boundaries of learning on Artificial Intelligence and do not hesitate to teach it in the classroom. Adaptability and knowledge of what is going on in real time are the ways you can support yourself and your community. Never stop learning."

Conclusion

In the MBA programs at Pepperdine Graziadio, the future of business isn't a distant concept; it's a daily practice. By mastering the synergy between artificial intelligence and human leadership, Graziadio MBAs are preparing to enter the workforce not just as technically proficient managers, but as the critical thinkers the modern corporate world desperately needs.