An Interview with Alex Lieder & Daniel Zinner
Artificial Intelligence has moved from experimentation to everyday application in many organisations. Yet one fundamental question keeps coming up in PMA Academy trainings, client projects, and leadership discussions:
Which AI system should organisations actually work with — and how do they enable teams to use it effectively?
To explore this, we spoke with Alex Lieder, AI expert and advisor, and Daniel Zinner, PMA Co-Founder, about AI Prompting as a system capability, the realities of Google Gemini and Microsoft Copilot, and what organisations consistently underestimate when rolling out AI at scale.
Not really. Framing this as a winner-takes-all debate misses the point. Both systems are powerful, but they are embedded in very different ecosystems. The real question is not which tool is better, but which system fits the organisation’s structure, data landscape, and ways of working.
Exactly. AI Prompting is not a feature — it’s a capability. And capabilities don’t live in tools alone. They live in people, workflows, governance, and leadership behaviour.
Microsoft Copilot is deeply integrated into the Microsoft 365 environment — Outlook, Teams, Word, Excel, PowerPoint. For organisations already standardised on Microsoft, Copilot feels like a natural extension of daily work.
Google Gemini, on the other hand, is embedded in Google Workspace and benefits from Google’s strengths in search, information synthesis, and multi-modal AI. It often performs very strongly in exploratory, creative, and research-heavy tasks.
From a prompting perspective, both systems are capable — but they behave differently depending on context, permissions, and data access.
Three differences show up repeatedly in practice. First, system integration. Copilot fits seamlessly into Microsoft-centric workflows, while Gemini offers more flexibility for exploration and ideation.
Second, data context. Copilot can leverage internal documents and collaboration data extremely well — if governance is set up properly. Gemini often excels at synthesising external knowledge.
Third, prompt behaviour. Gemini responds well to open, creative prompts. Copilot performs particularly well with structured, task-oriented prompts tied to documents and workflows.
These differences shape adoption — and frustration — if teams aren’t enabled properly.
The same patterns appear across industries:
- unclear expectations about what AI can and cannot do
- fear of data leakage or compliance violations
- inconsistent prompting skills
- over-reliance on outputs without critical thinking
- frustration when results vary in quality
Many teams assume these are tool problems. In reality, they’re enablement problems.
“Organisations need to think beyond the prompt.” - Alex Lieder
AI Enablement never belongs to one function alone. The strongest setups always involve close collaboration between IT and HR.
IT ensures, system access, security and governance, and technical integration.
HR and Learning ensure upskilling, change management, and cultural adoption.
“Leadership involvement is necessary and required as role model.” - Daniel Zinner
Without HR-driven learning, tools stay underused. Without IT governance, risks increase.
We consistently see three levers that make the difference.
First: education before acceleration. Teams need a shared baseline — how LLMs work, what prompting really is, and where the limits are.
Second: use-case-driven enablement. Generic AI training doesn’t stick. Prompting must be anchored in real roles, real tasks, and real workflows.
Third: clear guardrails, not restrictions. Well-defined do’s and don’ts build confidence instead of fear.
“AI needs a playground without too much pressure, so employees have a safe space to explore the tools.” – Alex Lieder
Very quickly. We see:
- deeper integration into daily tools
- more context-aware responses
- stronger multi-modal capabilities
- improved enterprise governance
The gap between AI as a tool and AI as a system capability is closing fast. Organisations that wait for perfection will fall behind those that enable responsibly.
Yes, and it’s usually pragmatic. Mid-sized companies often lean toward Google Gemini or lighter Copilot setups due to flexibility and cost considerations. Large corporations typically adopt Microsoft Copilot, driven by existing Microsoft contracts, security frameworks, and internal document ecosystems.
There is no universal best choice — only a context-appropriate one.
Absolutely — but only then.
Organisations report faster document creation and analysis. They also describe improved consistency in communication and reduced manual effort. Additionally, enablement results in better cross-functional collaboration.
But these benefits only materialise when people know how to prompt, when to use AI, and when not to. The goals is not to create prompt engineers but rather enable teams to leverage AI to design their work more efficiently.
Check out our PMA Perspectives on AI Enablement for Corporations
We don’t treat AI as an IT rollout. We treat it as organisational capability building.
That means:
- real business use cases
- safe experimentation
- leadership involvement
- learning beyond prompts
“Enable rather than teach” is not a slogan — it’s a design principle. Participants don’t just listen. They apply, test, reflect, and integrate AI into their work.
From Insights to Practice
The insights from this PMA Chat are based on real client projects, PMA Academy trainings, and applied use cases across corporate functions. Selected frameworks and visuals from our work can be reused for internal enablement initiatives.
Learn more about structured AI Enablement in the PMA Academy AI Enablement Courses here:
Authors:
Daniel Zinner is an international HR expert, entrepreneur, and communications consultant. His expertise lies in HR, strategy, digitalisation, and transformation strategy.
Alex Lieder is an experienced founder who built an AI-driven HR tech product for skill assessment as early as 2020. He now builds a startup accelerator and teaches founders how to apply large language models such as ChatGPT, Claude, and Gemini in real-world business contexts.
