Beyond Efficiency: Moving AI from Admin Tasks to Thinking Tools in Public Affairs

When I run workshops on AI in public affairs, I often start with a simple question: “To what extent are you already using AI in your day-to-day work?”

The answers usually trace out a bell curve. A couple of early adopters experimenting confidently. A majority dabbling at the surface — drafting emails, checking translations. And a handful who haven’t started yet.

At first glance, this is exactly what you would expect from any new technology. But dig a little deeper, and the adoption story becomes more complicated.

The “younger = more advanced” myth

I saw this clearly when teaching an executive master’s course with a group of professionals in their mid-twenties. Out of 15 participants, only two had paid accounts for LLMs. Several didn’t use them at all — often because their companies prohibited AI use in day-to-day tasks.

The result was that many were unfamiliar with even basic prompting. My assumption that the youngest demographic would naturally be ahead of older colleagues turned out not to be the case. This wasn’t a question of ability or curiosity, but more about organizational access and policy. The adoption curve may not map neatly to age or seniority; it may simply reflect the uneven way organizations are responding to AI.

A collective blind spot

These uneven patterns of adoption made me reflect on something bigger: perhaps, as a profession, we are all facing a kind of collective Dunning–Kruger effect.

We talk a great deal about “AI” in public affairs, but very few of us have the full range of skills to understand what adoption and implementation really entail from a technological perspective. We tend to treat AI as a single thing, when in reality it is a collective term covering many different disciplines and technologies, with applications stretching from text generation and monitoring to predictive modelling and even robotics.

That gap between how we talk about AI and how we truly understand it may be one of the biggest barriers to meaningful adoption.

From efficiency gains to thinking gains

What most people do use LLMs for — when they use them — is the obvious stuff: writing emails, fixing translations, speeding up everyday text. That matters. These tasks eat up time, and efficiency gains are real.

But if we stop there, we miss the real opportunity: AI as a thinking tool.

Why call it a thinking tool? Because it doesn’t just produce text — it helps us structure, test, and extend our own reasoning. Woprk with an LLM to:

• Map how two pieces of legislation interact.

• Explore implications for a sector or stakeholder group.

• Generate alternative framings of the same issue.

• Surface gaps and assumptions in a draft strategy.

The value isn’t in the words it generates, but in the angles it exposes. That makes AI not only a thinking tool but also — when used deliberately — a search tool that widens the field of view and a sparring partner that challenges our assumptions.

Efficiency gains save time. Thinking gains generate insight. And in public affairs, insight is what drives influence.

Splits in professional practice

Among professionals, I also see another divide. Many understand public affairs, policy, and legislative processes deeply, but have limited exposure to the data, coding, and technical side of AI. a smaller group come from the opposite direction — comfortable with datasets, analytics, and coding, but less fluent in the nuances of policymaking.

Bridging that split is critical. If public affairs professionals are to adopt AI meaningfully, they need to understand both sides — not to become engineers, but to develop enough literacy to connect the technology with their practice.

And this isn’t just about individuals. This split may also play out differently across regions and political cultures. In Europe and many national contexts, public affairs has traditionally emphasized relationships, processes, and institutions. By contrast, in places like the United States, data-driven strategy — opinion polling, voter profiling, predictive targeting — has been central for decades.

If AI accelerates the global shift toward more data-driven public affairs, some cultures may need to adapt more quickly than others. That adaptation will require not only technical skills but also the ability to integrate data-driven insights into the traditional craft of advocacy and influence.

Why this matters for teams

The difference between “efficiency AI” and “thinking AI” is strategic. Efficiency AI gives you more time.

Thinking AI gives you more insight.

Public affairs teams that stay in the first category will save hours on routine work. But those that make the leap to the second will accelerate their ability to understand, frame, and act on complex policy challenges.

Final reflection

The adoption curve is shifting, but unevenly. Some teams still don’t have access. Some individuals only use AI for admin. Others are beginning to experiment with strategic applications.

The key is not age, role, or even sector. It’s mindset. If AI is just a faster email writer, you’ll capture efficiency. If it becomes a partner in structured thinking, you’ll capture something much more powerful.

So the real question is: how are you using AI today — and is it helping you think, not just type?