AI ROI Starts with Better Processes

Artificial intelligence has quickly become one of the main priorities for organizations. Companies are experimenting with Microsoft 365 Copilot, AI agents, intelligent document processing and generative AI.

Alongside all this experimentation, an increasingly important question is emerging:

Are AI investments actually creating measurable business value?

In many organizations, the challenge starts before AI even enters the picture. It starts with the process itself.

Look beyond your core business systems

Most organizations already use ERP, CRM, HRM, MES or other specialized business applications. Yet many everyday internal processes still happen outside these systems.

Requests are submitted by email. Statuses are tracked in Excel. Documents are stored in different folders. Approvals require manual follow-ups. Employees copy information between systems. Reports are assembled manually.

Individually, these activities may not seem critical. Multiplied across departments and hundreds of employees, however, they create significant hidden inefficiencies.

This is also where AI initiatives can run into problems. If a process is fragmented across emails, spreadsheets, documents and individual employees, there is no clear workflow and often no reliable source of structured data. Without that foundation, adding AI does not automatically make the process better.

Start with the process, not the technology

Instead of starting with "Where can we use AI?", we recommend asking "Where are our employees losing time?"

Look for processes in which employees regularly:

• move information between emails, spreadsheets and business systems;
• search for the latest version of a document or piece of information;
• manually remind colleagues about deadlines;
• prepare recurring reports;
• enter the same data more than once;
• ask one another about the status of requests or tasks.

These are often excellent candidates for digitalization. Once a process is structured, an organization can move through four stages:

Digitize

Create one clear digital process with defined responsibilities, statuses, information and documents. Employees know where to start the process, what needs to happen next and where to find the relevant information.

Automate

Remove repetitive administrative work. Notifications, reminders, approvals, document generation and transfers of information between systems can often happen automatically.

Measure

A digital process starts producing reliable data. Organizations can measure processing times, workload, bottlenecks, overdue activities and other indicators that were previously difficult to see.

Add AI

With a structured process and reliable data in place, AI can be introduced where it creates additional value. It can help extract information from documents, classify requests, summarize cases, search large volumes of information, suggest next steps or support employees in making decisions.

The distinction matters: AI is no longer an isolated experiment. It becomes part of an actual business process.

What does this look like in practice?

Hisense Europe provides a good example. Its R&D team managed patent research requests using forms sent by employees via email. This made it difficult to maintain a clear overview of incoming requests, monitor their status, allocate work across the team and prepare reports efficiently.

Together with Hisense, BizIT digitalized the process using Microsoft Power Platform.

The new application provides a structured overview of requests and their statuses, automatically informs stakeholders when a status changes and generates Word reports from information already collected in the process. The captured data can also be used for up-to-date analysis and reporting.

What changed was not simply the technology. The process itself became structured, transparent and measurable. This creates a much stronger foundation for future automation, analytics and AI.

As Elvir Čaušević, Director of Shared R&D at Hisense Europe, explained:

"The tools within Microsoft Power Platform have impressed us, as they allow us to digitalize internal processes that we previously managed via Excel and email relatively quickly."

Sometimes automation creates more value than AI

One of the most important questions companies should ask is not "How much AI are we using?", but "Where is technology creating measurable business value?"

Sometimes the answer will be an AI agent. In another process, automatic document generation may eliminate hours of repetitive work. Elsewhere, a simple application that gives everyone a clear view of responsibilities and statuses may create the biggest improvement.

AI should therefore not be the goal. Better business processes should be the goal. AI is one of the tools that can help us get there.

From digital transformation to intelligent processes

Digital transformation and artificial intelligence should not be treated as two separate initiatives. There is a natural progression:

Digitize → Automate → Measure → Add AI

The stronger an organization's digital foundation, the easier it becomes to introduce AI into everyday work in a way that employees actually use and management can measure.

Before launching the next AI pilot, take another look at the processes already running across your organization, especially those that still depend on Excel files, emails, manual approvals and repetitive administrative work.

The next valuable AI opportunity may begin by simply building a better process.