Avanade has positioned itself as the world’s leading Microsoft expert. After more than two decades, large enterprises have come to be intimately familiar with it. But now, Avanade is also aiming to embrace the mid-market, allow them to lead in an ever-shifting technology landscape and accompany them in the AI revolution. What does that entail? We discuss it with Avanade’s leadership.
“It’s very challenging for customers to stay ahead of technological developments,” CEO Rodrigo Caserta tells us. He has led Avanade since September 2024, having been with the company since 2015. He also gained extensive experience at Accenture, Avanade’s co-founder along with Microsoft. The two companies formed Avanade all the way back in 2000.
Intense competition
Avanade supports companies in their adoption of Microsoft products. To do this in the best way possible, it requires specialists that are very much sought after by countless competitors. About four to five years ago, Avanade realized it needed to think differently about talent acquisition, Caserta explains. The company now hires employees based on potential. They must have the skills to always anticipate change and work with industry experts to translate innovations to the customer. Why is that? The CEO explains that one’s knowledge base at any one point in time may not be quite so relevant a few years later. In fact, in his time at Avanade, he has long ago observed that technological shifts are far too fast to hire only domain experts at every turn.
To satisfy customers, Avanade faces high expectations. Late last year, the company published its findings on the mid-market—a diverse collection of organizations that typically have small IT teams and are generally less structured than the world’s largest companies. Despite their smaller size, these organizations have ambitious goals. According to Avanade’s research, mid-market companies expect AI to deliver a 4X annualized return (!) on investment, a trend that emerged in early 2023 when AI became the biggest promise (or hype) in boardrooms. However, this market segment cannot and will not make large upfront investments. Caserta notes that CFOs of these companies are extremely critical: grand AI promises aren’t readily accepted, and visible results are essential.
Avanade defines mid-market organizations as companies with revenues between $500 million and $5 billion. This includes many household names—tech companies like HubSpot, Atlassian, and Squarespace, for example. It’s a diverse market spanning virtually every sector, giving Avanade plenty of potential customers to address. Caserta indicates that approximately three-quarters of Avanade’s revenue comes from work with large enterprises. The growth opportunity clearly lies in the remaining 25 percent. How does the company plan to leverage that? And what does it have to offer mid-market customers as Microsoft specialists?
Distillation
Avanade CTO Aaron Reich first describes how Microsoft is currently innovating with AI. This is the starting point for both Avanade and potential customers; new products in 2025 cannot be separated from their new “AI-driven” features, regardless of what that actually entails. Reich identifies three ways Microsoft is approaching this. The first is the simplest: the tech giant has embedded AI in existing products like 365 and Power BI. Then there are standalone AI solutions, such as Microsoft Copilot as a dedicated application. The third and most complex approach involves building an AI platform—creating a whole that is greater than the sum of its parts. This platform allows organizations to build tailored solutions with targeted AI tooling and innovated work processes.
All of this is manageable for enterprise organizations. They have the resources and sizeable IT teams needed to leverage Microsoft’s innovations alongside Avanade. But the mid-market is too diverse and too focused on immediate results to invest heavily before seeing returns. The knowledge gained from Avanade’s work with larger organizations must be distilled to be acceptable to mid-market clients.
The first mid-market concern complicating AI adoption is financial. However, Caserta also argues that concerns about transparency and AI bias are largely addressed when organizations are clear about their approach and see adoption in action. For transparency, Avanade has developed the AI Control Framework, which explains how AI should be used ethically and authentically in alignment with corporate values. For example, AI-generated emails can include clear watermarks to identify them as such—a practice Avanade also uses internally, according to CTO Reich.
Microsoft is also helping drive AI adoption. It recently began offering Copilot functionality by default within its Microsoft 365 suite. Google took a similar approach with Workspace. Now everyone can use AI within familiar applications, albeit with some limitations. Caserta suggests this approach is necessary to truly drive adoption. Evidently, adoption was too slow when Copilot options were available for $20 per month per user.
Avanade primarily wants to offer customers the low-hanging AI fruit first. Once organizations experience a “quick win” in productivity through AI-driven summaries in Teams or AI suggestions in PowerPoint, they may recognize the broader potential of the technology. Yet Avanade is positioning itself quite differently in this area than we’ve heard before.
AI isn’t just a productivity play
CTO Aaron Reich doesn’t consider himself a strong believer in AI as a productivity driver. The power of the technology isn’t readily quantifiable at this point, he says. The measurement criteria for AI’s success will change, but how? “We don’t yet know what that measurement point will be. That is why we are now looking at the process side. In that area, there is a value chain, where you can generate business value.” Avanade is examining which personas and domains can transform their processes, both internally and for customers. The coveted 4X ROI won’t come simply from becoming more productive with existing work processes, according to this thinking. “That’s simply difficult,” Reich notes.
What Reich does believe in is agentic AI. Time will tell how successful hyperscalers are in practical implementation, he says. In this context, hyperscalers include not only Microsoft, Google, and AWS but also OpenAI, Anthropic, and potentially others. Currently, organizations are primarily concerned with tasks where AI can assist. Each job consists of a knowable number of tasks, where AI technology helps in different ways with varying effectiveness. But Reich identifies the next steps: goals and processes. AI agents are designed to achieve specific goals, typically using different tools. However, we’re still in uncharted territory regarding processes.
An example is order-to-cash (O2C), the process from payment processing to payment to the seller. Reich describes how this might consist of 15 different tasks, some of which can be automated. But once you’ve completed that automation, you actually need to question what all those processes are for. In an AI-driven environment, some steps might be unnecessary, or you might want to harness the data from this process differently. The vision is a complete revolution in working methods.
Workarounds
In 2025, organizations often have substantial groundwork to complete before they can boldly reengineer their processes. Caserta mentions the delaying effect of regulations, particularly in Europe. He notes that conversations about AI technology differ dramatically between regions—while Asia and North America display mostly optimism, Europeans are cautious about compliance risks and navigating GDPR legislation.
Guido van Beuningen, Country Manager Netherlands at Avanade, adds that recent economic challenges have also impacted the pace of European innovation. But also, “many customers in Europe have a frontrunner syndrome,” he says. While European organizations adhere to high security standards, enjoy good connectivity, and have long adopted the Microsoft suite in the cloud, they lag in AI adoption. Van Beuningen notes that customers are especially adept at establishing data platforms, implementing data governance, and consolidation. He believes the adoption wave will come, possibly with less friction than elsewhere due to Europe’s infrastructural maturity.
This doesn’t mean Europe can rest on its laurels. Effective AI use requires understanding how to leverage your own data. “You then naturally end up with ethical issues,” says Van Beuningen. “Banks still have seventeen layers of traditional ETLs (extract, transform, load) before they can present a report.” In other words, the data journey must be transparent. Avanade wants to design architectures that enable organizations to adopt AI constructively.
There are legitimate concerns with hasty AI adoption. Reich provides a historical example: seat belts. In the 1960s and 1970s, these were designed exclusively for men, resulting in higher injury rates for women. These design flaws were corrected far too late, when they could have been prevented from the beginning. AI adoption must avoid similar biases and unintended consequences. Therefore, Avanade is already carefully considering the future, particularly by anticipating Microsoft’s upcoming plans.
Being ahead of the future
Reich explains that AI investments aren’t new to Avanade. When Microsoft launched Cognitive Services around 2015 (now Azure AI Services), Avanade embraced it. The company must continually consider how to provide solutions that build upon Microsoft’s foundation. For instance, Microsoft’s GenAI-focused offerings include approximately 3,000 LLMs. Some models are domain-specific, while others are flexible and general-purpose.
For customers, this represents a shift from the classic build-versus-buy approach, according to the Avanade CTO. They can purchase Microsoft’s offering and then build upon it themselves. Current use cases frequently include enhancing customer experience, partially automating office work, and supporting software engineering. Avanade was surprised to discover that it’s not developers but software testers who are currently benefiting most from GenAI.
CEO Rodrigo Caserta emphasizes that it’s not just about the present. Microsoft is making investments now in technologies that may not reach the market until around 2030. Avanade needs to prepare for this future. Led by CTO Reich, Avanade’s advanced research branch is already anticipating the direction of Microsoft’s offerings. For example, Copilot Agents and the standard 365 applications will remain, but embodied AI—essentially smart robotics—will eventually be added. Avanade needs to determine now how these developments will fit into customers’ learning and development processes.
Conclusion: authentic AI
Avanade cannot afford to be reactive, Caserta states. The company must prepare for developments over the next three to five years. Its staff has already begun this work and will continue to reinforce it with the philosophy they’ve defined. Customers have become more demanding, partly because Avanade now addresses the mid-market with its smaller budgets and compact IT teams. While AI promises much, organizations aren’t nearly as prepared as they believe they are.
Avanade’s priority is ensuring these organizations succeed with their implementations from the start. Initially, this may manifest in ways that don’t fully utilize AI’s potential. After all, productivity growth isn’t necessarily the best metric for measuring the technology’s value. Many companies have already implemented machine learning through computer vision or predictive maintenance. With LLMs becoming commonplace, avoiding the technology seems impossible. However, before reaching that point, organizations should question what processes truly define their business. Many companies haven’t yet clarified this fundamental question. According to Avanade, it’s time to formulate a meaningful answer.