01
%
84
say their AI investment is to replace employees to save costs
The Challenge
Avanade discovered a tension at the heart of mid-market organizations: how decision-makers view AI’s value. While their top priority is integrating AI with their processes to create new revenue streams, 84% believe their organization is investing in AI to replace employees and cut costs.
A laser focus on headcount and productivity could cause organizations to overlook new paths for growth, says Florin Rotar, Avanade’s chief AI officer. “We think that it’s wrong to look at [AI’s value] only from that narrow dimension and that most growth will come from process reinvention.”
“One of the first steps in a long-term AI strategy is to put your people first,” says Rotar, which means addressing the individual needs of organizational units, functions and roles to rethink current processes. For instance, one large retailer improved their call center experience by using AI to provide a one-line summary of customers and their issues, allowing agents to get up to speed more quickly and provide a better customer experience.
Consider also how AI can create entirely new revenue streams. For example, retailers can use AI-powered website search analysis to identify emerging customer trends, such as interest in specific clothing styles. Generative AI can then help create digital product designs that reflect these trends, accelerating manufacturing and bringing products to market more quickly.
The Strategy
Redefine Productivity
Craft A Long-Term AI Value Vision
“Instead of measuring productivity with time saved or tasks completed, we should quantify AI’s contributions by focusing on high-impact outcomes,” says Rotar.
When Avanade measured how its teams saved time by offloading manual tasks to Microsoft Copilot, for example, it also tracked the uplift in collaboration and satisfaction across teams.
“If we can use gen AI tools, copilots or agents to carry out high-volume or repetitive tasks—and even entire workflows—this frees human employees for more strategic and creative tasks,” says Reich, and leaders will need to find new ways to recognize and reward these contributions.
Are You Missing
AI Growth Opportunities?
Leaders should create a framework that addresses ethical considerations, legal implications, data privacy concerns and the need for transparent decision-making processes when using AI.
Every AI use case will have unique considerations. Avanade has clients who use AI to guide utility workers in the field, where health and safety are paramount concerns. Other clients use AI to drive manufacturing design ideas, with security and intellectual property as the primary focus.
“Fostering a culture of responsible AI requires the same kind of attention and rigor you need to instill values like safety, inclusivity or quality throughout your organization,” says Reich. He advises leaders not to treat responsible AI as a tax but to use it as an opportunity for continuous learning and change.
Use AI Responsibly
“We believe [this] is an imperative for the whole organization, not just a handful of departments, roles or experience levels,” says Rotar. He suggests approaches that include role-specific workshops, expert-led training sessions, and creating a shared library of guidance and successful prompts. “We've found that the most effective training includes a combination of AI fundamentals, best practices for getting the most out of your AI tools and reinforcement of responsible AI principles.”
He says that most employees will use AI in predictable ways, so it’s easier to anticipate upskilling needs. However, team members or departments “pushing the envelope of what’s possible with AI” may need targeted guidance and support.
Cultivate AI Fluency
The Strategy
AI urgency is real; 85% fear they’ll be less competitive if they don’t speed up AI adoption. However, a haphazard or piecemeal approach without the right people and data foundations can expose companies to risks.
Decision-makers appear keen to address these fundamentals; 79% plan to increase investment in training to help employees adopt emerging tech tools. Meanwhile, data and analytics platforms are the top investment priorities to ensure that data is secure and fit for AI.
The Challenge
worry their organization will lose its competitive edge if it doesn’t accelerate AI adoption
%
85
Is Your Organization Built
For AI Acceleration?
02
“You can’t build quality AI programs if your data is inaccessible, poorly governed or unsecure,” says Reich.
He urges organizations to review their data estate, together with their security approach, and create a roadmap to ensure data is AI-ready. This process includes determining strong data governance processes using a tool like Microsoft Purview and looking at how to manage both the structured and unstructured data through modern platforms. “A data platform like Microsoft Fabric that delivers real-time intelligence can help to give access to the right data to the right people at the right time—and the ability to get to a first minimum viable product in weeks versus months.”
Ask If Data Is Fit For Purpose
You once needed extensive coding experience to create an app; now, AI allows anyone to create one using natural language commands. These apps can be infused with agentic AI—which can autonomously handle complex tasks, build memories, solve problems and communicate with other agents in multi-step processes—making them even more powerful.
With these new capabilities unlocked for everyone, Reich urges leaders to rethink how apps are created within their organizations. “This may entail eliminating silos between professional developers and app-building non-developers, enabling them with the data they need, and determining who is best to build a given app.”
Embrace New Forms Of Collaboration
As generative AI is embraced in the workplace, leaders must consider and prepare for a kaleidoscope of new challenges. For example, Avanade’s research found that many professionals are concerned that their leadership cannot differentiate between AI-generated and human-generated work. Reich cites employee appraisals as another hurdle leaders must navigate: "Will employees that use AI face a different performance review and achievement expectation than those that don’t have access or don’t use AI? How will organizations set performance criteria for autonomous AI agents?”
Rotar says leaders should establish a vision, purpose and principles for AI's role in the organization. “If that vision isn't established upfront, it's hard to gauge whether your efforts and investments are paying off.”
Manage The Human-Machine Dynamic
The Strategy
AI is now part of many teams’ everyday workflows, and the reviews are in: more than half of senior managers give their AI ‘colleagues’ an A-grade. By treating AI agents as coworkers rather than mere tools, teams can use them to reinvent processes to ignite greater creativity, collaboration and productivity.
However, with that comes the need to consider how to manage this new workplace environment. Humans desire a sense of belonging, engagement and collaboration at work. Yet, only 40% feel very confident in their organization’s ability to manage the emotional and social dynamics of employees working alongside AI.
The Challenge
Have You Made
AI Part Of The Team?
03
Here’s how senior managers grade AI’s performance:
Respondents asked to grade A–F
%
A-Grade
B-Grade
35
C-Grade
%
Leaders must consider how much influence and authority to give AI. “A system with poor accuracy or quality might be fine for creating images for a slide deck,” says Reich, “but it would be unacceptable for a riskier task like translating pharmaceutical instructions.”
Organizations must also quantify the potential loss or harm if the system performs erroneously to establish a risk threshold for AI. Then, considering the system's expected benefits, they must calculate how frequently the organization can handle such loss.
Gauge Your Appetite For Risk
Avanade found that 96% of organizations are exploring ways for AI models to “forget” knowledge by the end of 2025, which means selectively removing specific information from their memory or training data. Organizations can achieve this by improving data governance practices and automating model retraining and fine-tuning on a set schedule.
“Controlled ‘forgetting’ can help organizations address the complexity of governing collective knowledge,” says Rotar. Think of hotel safes designed to reset passcodes between guests, ensuring no previous codes remain accessible. This principle of selective, purpose-driven memory must now be applied to AI within organizations.
Teach AI To Forget
The Strategy
Teams need confidence in AI’s output. However, while 96% of those surveyed say they use AI for decision-making, only 26% fully trust the results.
Concern over data quality may be one possible cause, and it’s cited as the top reason for limiting employee usage or access to generative AI tools. AI is trained on large datasets, but sometimes, these contain information they shouldn’t. That information might range from someone else’s intellectual property to spyware or malicious code that can be later used to steal user data.
The Challenge
Do You Trust What AI Says?
04
of respondents say they rely on AI for decision-making
%
96
While
fully trust the results of AI tools, a decline over the past 12 months
Only
%
26
Download Avanade’s Full
AI Value Report 2025
*4,100 IT decision-makers and senior business decision-makers outside of IT were interviewed for two separate research projects in August and September 2024 across Australia, Brazil, France, Germany, Italy, Japan, Netherlands, Spain, UK, and the U.S. Respondents worked for organizations with $500M to $5bn global annual revenue in banking, consumer goods, energy, government, healthcare, life sciences, manufacturing, nonprofit and retail.
53
%
9