AI Inside: Democratising Technology with Microsoft

In the rapidly evolving landscape of artificial intelligence, few companies have the reach and influence of Microsoft. As the tech giant celebrates its 50th anniversary this year, it continues to shape how businesses and individuals interact with technology. We recently had the pleasure of hosting Menatallah, a Senior Azure Cloud Solution Architect at Microsoft, on the Project Flux podcast. Her insights into Microsoft's AI strategy, particularly around Copilot, offer a fascinating glimpse into how AI is being democratised across industries.

The Future of Work in the AI Era

When asked what's been on her mind lately, Menatallah didn't hesitate: "How can the youth, especially those who are in the university or are joining university, build career pathways for this upcoming AI era?" It's a question that resonates deeply, especially for those with children approaching university age, as they prepare for a world where traditional career paths are being disrupted.

The challenge is particularly acute because, as many experts have noted, our education system was largely built for the industrial era. "It's not related to a certain industry," Menatallah emphasised. "I see that this is across any industry. It's how can you build the new careers that we will be expecting in this new era."

Her conversations with university students reveal a hunger for guidance about what's possible in this new landscape. Whilst some roles will inevitably disappear—a reality Menatallah acknowledges we "can never deny"—the key is to "go with the flow" and focus on what comes next. "It's really about going with this AI flow and not saying, I'm not related to this. It's not going to touch my job," she advised. "This is kind of being under the rock because at the end of the day, it will impact every single industry, wherever you are."

Menatallah's AI Journey: From Data Mining to Azure

Menatallah's own journey with AI began in 2009 during her Master's in Software Engineering. "What really interested me was a subject called data mining," she recalled. Under the guidance of the late Professor Mustafa Huneim at Nile University in Egypt, she discovered what she calls "a brand new world."

This was well before terms like "big data" and "data science" had entered the mainstream lexicon. "It felt for me that this is the next best thing," she shared. "And ever since then, I've been working on developing my knowledge there in whatever way that is available."

What's striking is how much more accessible AI education has become since those early days. "The knowledge available and the possibility of learning is much easier than it was like 10 or 11 or 12 years ago," Menatallah noted. Microsoft itself has played a significant role in this democratisation through platforms like Microsoft Learn, which offers clear pathways to certification that can be shared on LinkedIn—a point our co-host Yoshi highlighted as particularly valuable.

Microsoft's Copilot Strategy: Beyond the Hype

Microsoft's approach to AI, particularly with Copilot, is grounded in its longstanding mission: "Empowering every person and every organisation on the planet to achieve more." As Menatallah explained, this mission continues to evolve alongside technology.

Interestingly, Microsoft has moved away from the phrase "democratising AI" that was popular a few years ago. "We're using it less because right now AI is democratised everywhere. Everyone can use it," Menatallah explained. "And that's ultimately the goal.

That's ultimately where we're heading. We are making every single tool AI enabled."

This strategy reflects a profound insight: rather than requiring users to learn entirely new tools or technologies, Microsoft is integrating AI capabilities into familiar applications. "Whatever you're doing and whatever tool you're using and you're used to, this is where we are heading towards getting Copilot inside that tool," Menatallah said. The message is clear: "You don't worry about, I need to learn another technology. Oh, I will miss that train because it's not relevant to me."

This approach aligns perfectly with our mission at Project Flux—making AI accessible to everyone in the project profession, not just tech specialists. Yet as James noted from his recent experience at a conference in Athens, there's still a gap between access and adoption. Many project managers have Copilot but aren't using it effectively, often believing they need extensive training first.

Practical Applications for Project Managers

For project managers specifically, Menatallah highlighted several game-changing applications within Copilot for Project, which is part of Dynamics 365's project operations:

1.Work breakdown structures: "This can help project managers build a work breakdown structure from just a description in natural language," Menatallah explained. The AI creates the structure, and the project manager refines it.

2.Risk mitigation planning: Copilot can help identify and prioritise project risks, focusing attention on what truly needs action.

3.Report generation: A critical part of any project manager's job, now streamlined through AI assistance.

These capabilities free project managers from mundane tasks, allowing them to focus on what truly requires their expertise: "It's more about emptying your schedule into more something that is using your actual skill as a PM at the end of the day," Menatallah noted.

However, she was also candid about the current limitations. Copilot provides frameworks but doesn't handle the narrative aspects of reports. It also doesn't yet manage dependencies between milestones or assign responsibilities. "Let's agree that it's still at the beginning," she acknowledged, emphasising that there remains "a big part that the PM needs to be doing."

This human-in-the-loop approach resonates with what James described as the "bookending approach"—human at the beginning to set direction, AI in the middle to handle the grunt work, and human at the end to verify and refine. It's a model that provides comfort to those concerned about AI replacing human judgement.

Security and Data Governance

One of the most significant barriers to AI adoption is security concerns. Menatallah addressed this head-on, explaining that Microsoft's approach is to build on existing security infrastructure rather than creating entirely new systems.

"The main point of importance here is how we're handling data," she emphasised. "We take our time to onboard different products... because one of the things that we need to be ensuring is that it's abiding by our security boundaries."

This focus on responsible AI means ensuring that tools are safe and accessible for everyone whilst preventing harm. Importantly, Copilot inherits the security permissions and labels that already exist in an organisation's systems. As Menatallah put it: "Copilot only inherits the security or the labels that you already have."

This led to an illuminating exchange about a common concern: could someone use Copilot to access sensitive information like salary data? Menatallah's answer was clarifying: if someone could access that information through existing systems, they could access it through Copilot. If they couldn't access it before, Copilot wouldn't change that. "If there's a vulnerability within your SharePoint, it was there already," as James summarised.

Data as an AI Enabler

The conversation took an interesting turn when discussing Microsoft Fabric and its data mirroring capabilities. Whilst not directly an AI feature, mirroring serves as a crucial enabler by making data more accessible.

"The concept behind it is that we're trying to replicate data in a near real-time manner without the need of what we call the ETL, or the Extract and Transform and Load," Menatallah explained. This eliminates a significant burden for data engineers and makes it easier to bring legacy data into AI-ready environments.

For organisations with data scattered across various systems—a common scenario in project-heavy industries like construction—mirroring offers a way to make that data accessible without the cost and complexity of moving it. "Whatever updates takes place in the source, this is reflected," Menatallah noted, adding that Microsoft is continuously expanding the range of data sources that support mirroring.

Once mirrored into Microsoft's "one lake," the data becomes easily accessible to various AI and analytics tools. This approach addresses one of the fundamental challenges in AI adoption: making existing data usable without extensive restructuring.

Implementation Blind Spots

When asked about common blind spots in AI implementation, Menatallah highlighted a critical issue: "Organisations might have deployed Copilot, but no one is really using it." This represents a significant missed opportunity—investing in technology without realising its value.

The solution, in her view, involves education but goes beyond simple training. "People really need to understand, you know, how can they use those tools?" she explained. "If people are not being empowered to understand, okay, how can they use it? When can they use it? What are their boundaries? And get this kind of flow going... they will have no clue."

She shared an anecdote about delivering a session to professionals who had Copilot available but weren't using it. "After the session I did, they're like, we'll start using it today. It's already there, but we never used it. But after we learned the value, we're super enthusiastic to start using it." This enthusiasm, she believes, is essential for successful AI adoption.

The Cultural Shift

Menatallah proposed a novel approach to driving adoption: "I would designate what I would call an AI champ within each kind of department." These champions would speak the language of their colleagues and demonstrate AI's value in context-specific ways.

"Each department has its language and this is the jargon they prefer hearing," she noted. Having someone who can show how AI applies specifically to finance, marketing, or other functions makes the technology more relatable and its benefits more apparent.

This departmental approach represents an evolution in thinking about technology adoption. Rather than relying solely on central IT or innovation teams, it recognises the importance of domain expertise in driving meaningful change.

The parallel with Excel's adoption in the late 1980s and early 1990s is instructive. As James pointed out, Excel wasn't immediately adopted across organisations—it took years for it to become ubiquitous. Similarly, AI adoption will follow a curve, though perhaps accelerated by our greater familiarity with technological change.

The Future of AI in Organisations

Looking ahead, Menatallah sees AI fundamentally changing how we interact with information. "AI is going to change the way we consume information, the way we interact with information," as Jugal noted during the conversation. Menatallah agreed, suggesting that "we are going to be living in a chatbot world, even in projects."

This shift extends beyond just how we access information to how we work with it. The integration of agent capabilities—AI systems that can perform complex tasks autonomously—represents the next frontier. As these capabilities are incorporated into tools like Copilot, they'll further transform how work gets done.

Throughout all of this change, continuous learning remains essential. As Menatallah emphasised, the AI landscape is evolving rapidly, and staying current requires ongoing effort and curiosity.

Conclusion: Starting Your AI Journey

The conversation with Menatallah offers several key takeaways for anyone navigating the AI landscape:

1.AI will impact every industry, making it essential to embrace rather than resist the change.

2.The most effective approach is to start with familiar tools that now have AI capabilities built in.

3.Security concerns are valid but should be addressed through existing governance frameworks.

4.Adoption requires more than deployment—it needs champions who can demonstrate value in context.

5.The future will bring even more integration between AI and everyday work tools.

Microsoft's vision, as articulated by Menatallah, is one where AI becomes an integrated companion rather than a separate technology—where it's simply part of how we work rather than something additional we need to learn. It's a vision that aligns perfectly with what we're trying to achieve at Project Flux: making AI accessible and valuable for everyone in the project profession.

Want to stay updated on the latest in AI and project management? Subscribe to the Project Flux newsletter for weekly insights, tools, and expert perspectives. And explore Microsoft Learn for resources on getting started with Copilot and other AI tools.