Pricing has always been one of the most sensitive decisions in any business operation. In markets with tight margins, such as the pharmaceutical industry, a small mistake can turn into a loss at the end of the month.

And it was precisely this scenario that spurred a turnaround within the Gmill Group.

In this article, we summarize the main takeaways from episode #14 of InsightCast, with guest Felipe Deboni, IT director, sharing how the combination of structured data, automation, and artificial intelligence has completely changed the way the company makes pricing decisions.

The context of the operation and the challenge of pricing.

The Gmill Group operates in the distribution of medicines, handling thousands of SKUs and employing a commercial dynamic that varies according to the channel, region, and conditions negotiated with the industry.

In this scenario, pricing ceases to be a linear activity. An identical product, with the same milligram count and number of tablets in the package, can have different prices depending on factors such as geographic location, time of year, regional demand, tax burden, and existing trade agreements.

Before the current structure, this process was conducted mostly manually, which increased the risk of errors and made it difficult to predict the results.

As Deboni himself summarizes:

"The pricing was entirely human-based... and I don't even need to tell you how often they were wrong."

The turning point: business-oriented systems

The change began with a simple but difficult decision to implement: to stop treating pricing as an operational task and start treating it as business intelligence. From this, the Gmill Group developed two core systems:

System of trade agreements

The first step in this transformation was to structure the entire logic of commercial negotiation within a centralized system. Previously, information such as discount rules, margin limits, specific conditions per client, and negotiation history could be scattered across spreadsheets, isolated systems, or even the tacit knowledge of employees.

With the creation of this system, these variables become organized in a structured way, with well-defined rules, traceability, and real-time access. This reduces inconsistencies, avoids manual errors, and ensures that all areas operate based on the same criteria.

In practice, the impact is direct on business management; the company stops analyzing results only at the end of the month and starts monitoring profitability during the execution of operations. This allows for faster adjustments, greater control over margins, and safer decisions throughout the process, not just after the result has already occurred.

Automated pricing engine

The second pillar of the transformation was the creation of an automated, data-driven pricing engine capable of analyzing multiple variables simultaneously. Unlike traditional models, which rely on manual decisions or fixed rules, this system considers a combination of internal and external factors to determine the most appropriate price in each context.

The data analyzed includes cost, shelf life, and inventory turnover, as well as external information such as market behavior, regional demand, and competitive positioning. The system also incorporates commercial variables, such as specific agreements, promotional campaigns, and incentive policies.

This level of analysis allows for much more precise and dynamic pricing; the company can preserve margins when there is room, be more competitive when necessary, and respond quickly to market changes. Furthermore, it reduces reliance on guesswork-based decisions, bringing more consistency and predictability to results.

See also: How can technology be transformed into a competitive advantage in customer experience?

The role of data: without structure, there is no AI.

One of the strongest points of the episode is the clarity regarding the role of the database.

Before discussing AI, the Gmill Group needed to solve a basic problem: organizing its information. This step, often neglected by companies eager to adopt technology, is what determines whether an artificial intelligence initiative will generate real results or become just another forgotten project.

The numbers make the size of the problem clear. Gartner Research It points out that, on average, 60% of data professionals' time is spent cleaning and organizing information before any analysis. This means that more than half of the company's technical effort goes towards fixing a problem that could have been avoided with a well-planned data architecture from the start.

A McKinsey This diagnosis is reinforced in its annual report on the state of AI. The consultancy points out that the lack of integrated and high-quality data is the main obstacle cited by executives for scaling AI initiatives within organizations. It's not a technological limitation; the challenge lies at the foundation: inconsistent, disorganized, or ungoverned data directly compromises the ability to generate value with AI.

This involved structuring product data, standardizing historical databases, clearly documenting tables and fields, and evolving to a real-time data model. It's not glamorous, but it's fundamental. Companies that skip this step end up trying to build on sand. This database is what allows AI to truly function. Without it, any initiative becomes mere trial and error.

AI in practice: faster and safer decisions.

With the data organized, Artificial Intelligence acts as an accelerator. It doesn't replace strategy, but it expands analytical capabilities.

Today, the company is able to:

  • Assess demand by region.
  • Identify opportunities for price adjustments.
  • Predicting sales behavior
  • Making decisions based on multiple simultaneous variables.

When applied correctly, AI does not eliminate the human role; it enhances it, allowing decisions to be made more quickly, based on data and with greater predictability.

"AI first" mindset

One of the most interesting concepts in the episode is "AI first." The idea, at its core, is to ask yourself before creating any solution: "How can AI help here?"This completely changes the approach. Instead of starting by choosing tools or defining an already structured problem, the approach now considers from the outset how artificial intelligence can contribute to generating efficiency, scale, and better decisions.

In practice, this means that AI becomes a starting point; teams learn to look at each process asking where automation can reduce effort, where data analysis can replace guesswork-based decision-making, and where processing speed can generate a real competitive advantage in the market. As highlighted in the episode, this change doesn't necessarily depend on large initial investments, but on a different attitude towards technology.

For the Gmill Group, adopting this approach allowed the transformation to go beyond pricing and spread to other areas of the operation, creating a culture where technology and strategy go hand in hand.

Automation beyond pricing.

The same logic was applied to other aspects of the operation, with the automation of repetitive processes via RPA (Robotic Process AutomationIntelligent assistants for customer service and operations, and AI for reading unstructured orders, audio, images, and PDFs.

Previously, a team had to interpret requests submitted without standardization. Today, AI performs the initial reading, learns from human corrections, and drastically reduces operational effort over time.

The transformation in operations begins with aligning technology.

The biggest lesson from the episode isn't about the technology itself, but about how it's used today. Deboni reinforces a point that is still neglected in many companies: "IT can't just be about support. It needs to work alongside the business."

In practice, this means that the IT department needs to deeply understand how the business works, critically evaluate solutions before implementing anything, and measure the impact of what was done in terms of results.

This understanding doesn't happen in isolation; it requires close collaboration with other areas of the company, such as finance, sales, operations, and customer service, allowing IT to understand which systems are essential, which data is critical, and what impacts a failure could have.

Without this alignment, technology initiatives tend to lose connection with what truly underpins the operation, generating projects that do not solve problems or deliver value.

What can other companies learn from this case?

Even in different contexts, some lessons are universal. Pricing requires a strategic approach, and structured data is a prerequisite for any automation or artificial intelligence initiative. Without this database, the technology doesn't deliver what it promises. With it, automation frees up time and energy for better decisions, and AI empowers those who already understand the business.

A Technology alone doesn't transform anything; it's the combination of culture, well-defined processes, and the right tools that generates sustainable results. You don't need to start big, but you do need to start right, with clarity about where you want to go and a willingness to build the foundations that will support your progress over time.

Want to watch the full episode?

Listen on Spotify – InsightCast – How AI and data transformed pricing at the Gmill group #14