Scaling AI in the Enterprise: Lessons from Intelligent Workflows and Automation with Nirmal Ganesh

As artificial intelligence moves beyond experimentation and into day-to-day business operations, organizations are increasingly focused on how AI can improve workflows, automate complex processes, and unlock value from enterprise content. From intelligent document processing to agentic workflows, the challenge is no longer whether AI can be deployed, but how to implement it in a way that is scalable, governed, and aligned with real business needs.

To explore these issues, AIJourn spoke with Nirmal Ganesh, Senior Director of Product Management at Box, where he leads the global strategy for the company’s Intelligent Workflow and Automation suite. With deep experience in enterprise software, digital transformation, workflow automation, and AI-driven business processes, Nirmal shares insights on the evolution of intelligent workflows, the realities of enterprise AI adoption, and what organizations should be doing today to prepare for the next generation of automation.

You began your career as a software engineer before moving into executive product leadership roles focused on enterprise AI and workflow automation. What experiences shaped that transition, and how has your engineering background influenced the way you approach innovation and leadership today?

You know, I started my career as an engineer because I just loved building things. I really enjoyed figuring out how things worked – how software systems connect, communicate, scale to solve real-world problems, and how the foundation has to be right before you can build on top of it. From there, I moved into Product leadership, and even now, my journey still shapes how I think about innovation and leadership.

 First, it’s the empathy for builders – I’ve been there. I know the dependencies and know what it takes to build something. Second, it’s problem-solving. When I was a builder, there was more of a rulebook. Now we are figuring out how many things are done at once, and my programming background helps me connect the dots. Third leadership, a strong leader sets the vision and the north star! I learned a lot of that in the trenches with great product managers and executives! Trial by fire. I’m standing on the shoulders of giants!

Looking back on efforts around agentic workflows, what do you believe were the most important innovations you helped introduce, and how have those contributions influenced the way organizations think about automation and AI-driven business processes today?

One of the most important innovations I introduced was Box Automate.  It was built to automate workflows for how work happens today, with agents, people, and systems working together in one place.

When we started building Box Automate, the goal wasn’t to replace traditional automation with AI. It was to create a system where both could work together. Some tasks need to follow the same path every time for compliance or operational reasons. Others require judgment and context. We wanted to support both within a single workflow.

By combining deterministic workflows with AI-driven decision-making, organizations can automate routine work while keeping humans involved where oversight matters. Agents operate within defined guardrails, including confidence thresholds and escalation paths, allowing teams to increase automation as confidence in the system grows.

We’ve seen legal teams streamline contract reviews, procurement teams reduce vendor assessment times, and HR teams scale onboarding processes. The result is a more practical approach to automation, one that combines rules, AI, and human oversight in a way organizations can understand, govern, and trust.

You have spent years working on enterprise workflow modernization and digital transformation initiatives. What are some of the biggest operational challenges organizations still face as they move away from manual and paper-based processes?

There’s a ton of institutional knowledge in enterprises, and most processes are built around it. For example, with payroll workflows, it’s easy to say, “if salary is above X, approve,” but that’s not how the real world works. Recruiting compensation, market pressure, and internal constraints,  a lot of which isn’t documented. Almost every process has formal rules and the internal context for how things are actually done. Translating that into a system? That’s the hardest part.

For a bank, risk, security, and compliance come before process automation. There’s often a strong ROI case, but other priorities win  so you’re left with processes that are still mostly manual or paper-based.

Intelligent document processing and AI-driven automation are changing how companies manage information and workflows. Where do you see the greatest opportunities for AI to improve efficiency while still keeping systems practical and scalable for enterprise teams?

For decades, enterprises have relied on structured data from systems like CRMs and ERPs, but the real opportunity for AI lies in the vast amount of unstructured content that makes up most enterprise information, including documents, emails, images, and videos. AI can now understand and extract insights from this content, enabling it to perform tasks that previously required human review, such as processing complex mortgage applications or large document sets. Intelligent Document Processing (IDP) and AI-driven automation make this possible by orchestrating workflows across systems, agents, and people.

 To scale successfully, organizations need governance built in from the start, including audit trails, human oversight, escalation paths, permission controls, and guardrails that ensure security, consistency, and accountability.

Earlier in your career, you helped scale digital enrollment and document workflow platforms into large enterprise solutions. What lessons from those experiences continue to influence how you think about product growth and long-term technology strategy?

Two things stick with me very deeply. First is having a very strong thesis on value to users: digital enrollment wins when you reduce cost, pain, and friction to onboard users! It’s like Amazon: cheaper prices, bigger selection, faster delivery. 

Second: differentiation drives scale. The products that actually scale are the ones with real differentiation, whether technical, process-oriented, brand-based, or some combination. That has to be front and center in your thinking about product growth and long-term technology strategy. Without differentiation, you may see initial success but struggle to build a durable platform.

As automation and AI become more integrated into enterprise operations, how do you think leadership expectations are evolving for product and technology executives managing global teams?

AI investments have increased dramatically, and naturally, executives are expected to deliver ROI. If the returns don’t show up, spending will drop drastically. That accountability is table stakes for product and technology leaders now.

At the same time, I have seen non-technical execs spend their weekends setting up and experimenting with agents and automation tools to improve their work.  That shifts what leaders must model, not only strategy and execution, but hands-on engagement with technology.

Personally, that’s exciting, and that energy has pulled me back to my roots. I have become a builder again, and I encourage everyone on my team to do the same! The separation between PM, design, and engineering is dissolving, product managers are building, and engineers are product managers. Global execs are expected to drive business impact while fostering this kind of experimentation and cross-functional collaboration across those teams. 

Looking ahead, how do you see enterprise workflows evolving over the next five years as AI, automation, and intelligent document systems become more deeply embedded into everyday business operations?

Nobody says no to faster, cheaper, and more efficient workflows, and AI makes all of that possible today.  Over the next five years, I only expect more automation. Things like more autonomous execution, cognitive document intelligence, invisible UI with hyper-personalization, and intelligence embedded directly into everyday business operations.

AI is here to stay and will become as integral as email, laptops, and phones. My generation could not work without those; the next generation will feel the same way about AI. Work will look different, but it will be better.