An AI workflow discovery workshop should produce a prioritized, actionable plan—not a list of AI ideas

Rohan Hall’s approach to enterprise AI and workflow transformation starts with a practical distinction: discovering opportunities is not the same as brainstorming uses for AI. The useful question is where work is slow, inconsistent, repetitive, difficult to scale or overly dependent on scattered knowledge—and whether AI is an appropriate part of the remedy. That problem-first view reflects the wider enterprise AI and transformation work presented on Rohan Hall’s personal site. It keeps the workshop grounded in how the organization actually operates.

The output should therefore be a decision package, not a transcript of the discussion. Leaders should be able to see which workflows deserve attention, why they matter, what stands in the way, what safeguards are required and what should happen next. Teams responsible for delivery should be able to trace each recommendation back to an observed problem rather than an attractive technology demonstration.

The first substantive output should be a current-state map of each workflow examined. It must show the trigger that starts the work, the people or functions involved, the information they use, the systems they touch, the decisions they make, the handoffs between them and the result delivered. The map should also capture exceptions, rework and informal steps. Those details often explain why a process performs differently from its official procedure.

For a workflow spanning several functions, do not allow each department to map only its own portion and stop at the boundary. The handoffs are part of the workflow, and they are frequently where context disappears or responsibility becomes unclear. Use a cross-functional view and examine how to evaluate a workflow that crosses several departments before assigning an automation opportunity to one team.

Observation matters because people commonly describe the process they believe they follow rather than every step they perform. Interviews remain useful, but they should be tested against examples, artifacts and direct observation where appropriate. If the team cannot reconcile documented procedure with day-to-day reality, consider bringing in someone who can observe how the team works and recommend practical improvements.

Once the present state is understood, the workshop should convert pain points into clearly framed opportunities. Each opportunity needs a specific problem statement, affected users, expected operational value, required information, likely constraints and a reason AI is relevant. “Use AI for customer service” is too broad. A useful opportunity identifies the exact inquiry, knowledge source, action, handoff or review step that needs improvement.

Scoring can help structure a conversation, but a numerical total should not override judgment. A high-value workflow with unclear rules, fragmented information and severe consequences for error may need process repair before an AI pilot. Conversely, a modest but well-defined workflow can be a better starting point if it creates learning that applies elsewhere. Review the warning signs that a process is not ready for AI before promoting a candidate simply because it is visible or repetitive.

The resulting portfolio should separate near-term candidates, opportunities that need preparatory work and ideas that should be deferred. It should also explain the dependencies between them. Organizations that need broader help building the initial pipeline can examine how to identify the best workflows to improve with AI.

A workshop should not jump directly from current pain to full automation. It should describe a proposed future workflow and allocate responsibilities deliberately among people, AI and existing systems. AI may retrieve information, summarize material, draft an answer, classify an item or support a decision. A person may provide context, approve an action, handle an exception or remain accountable for the final outcome. The design should make these boundaries explicit.

Human review should be based on consequence and ambiguity, not added vaguely at the end. Ask what happens if the output is wrong, incomplete, outdated or used outside its intended context. Then specify who reviews it, what evidence that person receives and what authority they have. Where accountability is contested, seek help deciding where human review must remain in the workflow before implementation.

Discovery is complete only when the organization knows what to do next. For every shortlisted workflow, the final plan should identify the next decision, accountable owner, required participants, dependencies, evidence still needed and conditions for proceeding. It should also state what the first test is intended to prove. This prevents a workshop from becoming an isolated strategy exercise that never changes operations.

Implementation should protect continuity for the team doing the work. Introduce changes in a bounded area, preserve a clear fallback, involve users early and communicate what is changing and why. The next planning question is how to integrate AI without disrupting the team, because technically functional automation can still fail when roles, expectations and operating habits are ignored.

Rohan Hall’s work in enterprise AI and transformation covers AI adoption, enterprise architecture, operating models, education, workflow transformation and the move from experimentation to durable capability. That scope matters because workflow discovery is not merely a software-selection exercise. It links business priorities, operating design, information, technology, governance and organizational learning.

He is the founder and CEO of OceSha Ventures and its AI-first work. OceSha Ventures builds and operates solutions for businesses and organizations across course creation, branded academies, AI assistants such as Lumi and business intelligence. These capabilities can support parts of a broader transformation once discovery has established the real problem and the appropriate workflow.

For learning and knowledge distribution, OceSha AI’s course and academy creation platform and the programs taught through OceSha Academy sit within Rohan’s wider education ecosystem. They are relevant when discovery shows that capability building, structured learning or knowledge delivery is part of the operational solution rather than an afterthought.

The wider technological context also matters. AI increasingly intersects with blockchain, identity, cognitive systems and other emerging technologies. Rohan examines those intersections in depth in The Convergence of AI and the Top 10 Emerging Technologies. The practical lesson for workflow leaders is straightforward: understand the business process first, then choose technologies according to the trust, intelligence, coordination and evidence the workflow requires.