WHO IS THIS FOR?

Exploring the Principles and Impact of Human-Defined Workflows in the Age of AI-Driven Automation

As artificial intelligence becomes more deeply embedded in business operations, the challenge is no longer simply whether to automate, but how — and who gets to shape the logic behind that automation. While AI can accelerate and scale routine processes through “set-and-forget” automation, this approach often assumes frictionless execution is always the goal. In reality, such automation can suffer from shallow reasoning, hidden assumptions, and a lack of transparency — especially in domains that require contextual judgment and domain-specific nuance.

In contrast, the emerging practice of human-defined workflows offers an alternative vision — one where AI becomes a co-creator, not just an executor. These workflows are not static instructions but expressive design artifacts. They encode human intention: what should happen, how it should unfold, why it matters, and who has a say. At their best, they become dynamic tools for sense-making and decision-support, reflecting differences in domain expertise, values, and experience. The workflow becomes not just a means of execution but a medium of authorship.

This is especially relevant for small and medium-sized enterprises (SMEs) — particularly in manufacturing — where traditional ERP platforms often fail to accommodate the nuance and diversity of smaller operations. Instead of forcing SMEs into rigid, one-size-fits-all systems, we propose that human-defined workflows offer a more flexible and scalable alternative: one that SMEs can design, own, and evolve to match their unique context.

Today, productivity gaps remain real. Manufacturing SMEs trail larger peers by 10–15% in efficiency, especially in functions like order entry, quality checks, and reporting. We hypothesize that AI workflows, when applied even to just 20% of these tasks, can yield a 3–4% overall productivity lift. For a $10 million SME, that’s hundreds of thousands in annual value — with no added headcount. And it’s just the beginning. Proven, well-designed workflows don’t just close gaps — they become launchpads for new forms of innovation, surfacing adjacent issues, compound inefficiencies, and emerging needs.

Research Aims and Methodology

This research initiative will investigate human-defined workflows as both a design method and a framework for responsible AI collaboration. Our core objectives are:

• To define and differentiate human-defined workflows from conventional automation strategies.

• To examine how human agency, judgment, and variability shape AI outcomes.

• To explore workflow design as a legitimate form of authorship in high-context business domains.

This work will be delivered as a field-based, experiential learning project. Student teams will work directly with SME partners to co-design, document, and iterate AI-powered workflows across diverse business functions — from onboarding to forecasting. We will generate and track detailed use cases, measure productivity impacts, and collect reflective case studies. Our broader goal is to build a modular, shareable library of workflow templates — enabling regenerative reuse and shared learning across the SME ecosystem.

Significance and Impact

At its core, this research challenges a narrow vision of AI that equates progress with human replacement. Instead, we explore AI’s real value: amplifying human domain expertise. The focus shifts from replacing labor to designing better logic — workflows that align with business realities, stakeholder needs, and organizational values.

This human-first, systems-level approach is especially critical in areas where ethical judgment, complexity, and ambiguity prevail. By building workflows that are transparent, testable, and tuned to real-world use, we foster a new paradigm: AI not as a black box, but as a strategic collaborator. The result is a new architecture for value creation — one that invites more people into the design process and gives SMEs a powerful lever for growth and adaptability.

In sum, this research positions human-defined workflows as both a practical methodology and a philosophical stance — one that honors human insight while making AI work better for more of us.