The Intelligence Paradox: Human Strategic Judgment Matters More than Ever in AI age

The C-suite and boards today are grappling with a pressing question: what exact value a strategic advisor brings to the table if artificial intelligence can track, aggregate, synthesise, and analyse geopolitical developments, regulatory dynamics and political economy trends at scale?

Sounds about fair. And this question warrants a coherent response rather than a defensive one.

In short supply: Clarity, specificity, confidence & accountability

Let’s be clear, AI has fundamentally transformed the intelligence ecosystem. One can now effectively institutionalise – at speed and at scale – aggregation of publicly available signals across geopolitics, geoeconomics, public policy, regulations, and macro trends – tasks that used to take teams of analysts significant time and effort.

Be it enterprises, small and medium businesses, or start-ups, any organisation with access to AI tools and the orientation to adopt them can generate a fairly comprehensive picture of various moving parts, as applicable to their respective business context.

What this clearly implies is that synthesising available data and information is no longer a differentiator – it’s table stakes.

What remains unchanged is the nature of the decisions that strategic advisory, driven by actionable intelligence, is intended to facilitate. Senior executives and directors have more access, than ever, to briefing documents and dashboards than they could go through.

What they need today are three things. First, clarity and specificity on what a given information set means for their particular business situation. Second, confidence in a recommended course of action that they deem credible. And third, someone who can take accountability for the quality of the analysis behind the recommendation.

On all these three counts, AI, no matter how sophisticated it is and will evolve to be, falls short, and leaves CEOs, COOs, CFOs and boards high and dry.

Five constraints AI can’t engineer away

Some of you might push back on this and argue that companies can design, architect and tailor AI set-ups around their proprietary data and business processes, generating bespoke intelligence that is distinct from generic analysis.

Yes, that indeed is possible and feasible. However, it is also inadequate, for the following structural constraints relating to AI:

Data asymmetry is a feature, not a bug: Your proprietary data, spanning your organisation’s operations, risk exposure across the value chain and historical decisions, is rich on the inside. Yet, it is virtually completely silent on the outside. The intelligence gap in geopolitics and political economy is structurally external. While AI increasingly can and is getting better at understanding the firm's internal context by virtue of being provided proprietary business information, it still suffers from the blind spot externally. AI can’t provide you visibility into “the shadows” – what’s actually moving in diplomatic and political corridors, policy channels, or the informal conversations leading up to regulatory measures. No configuration can solve for this constraint.

Actionable intelligence lies not in datasets, but relationships & interpretation: What a ministry, department, political entity or regulatory agency is plausibly likely to do – as opposed to their official positions? What is the thinking and direction of travel in informal diplomatic and political channels? How are senior policy makers looking at a given issue, both semantically and substantively, before they formally unveil a policy?

The answers to these questions don’t lie in either your in-house enterprise training or historical data, or in the public domain. It lies in the ability to qualitatively interpret signals emanating from conversations and interactions driven by relationships and institutional access.

In short, AI-enabled datasets and dashboards won’t solve your decision problem, in terms of delivering either clarity, specificity and accountability, or confidence behind a specific recommendation.

That’s where human network intelligence and insights, preceding the public domain by months, still matter. No system configuration or AI frameworks can replace it.

AI can’t substitute domain expertise: To configure AI tools for credible geopolitical, policy and political economy analysis, users must ask the right questions relevant to the given business context, embed the appropriate filters and frameworks, define – at the design stage – the set of specific material signals, and opt for variables that carry analytical weight.

All of this is precise domain knowledge, or subject matter expertise. Without it, businesses, big and small, run the risk of providing ill-specified inputs, and in turn, producing sophisticated-looking outputs. While these polished, slick AI-generated insights and reports “feel” credible, the underlying logic and structural rigour may not be quite sound.

Automation biases compound geopolitical risks amid volatility: The more sophisticated the internal AI analytical tool is, the less visible and manifest its failure modes tend to be. Outputs that seem structured, exhaustive and analytically rigorous lead to user confidence that the underlying analysis may not justify. This cognitive trap is particularly risky in the context of geopolitics, geoeconomics and political economy, where the landscape is ever fluid, the questions inherently complex, and the signals fundamentally ambiguous. Furthermore, the situations where AI is most vulnerable –unanticipated, shocking political and policy turns, abrupt regulatory shifts, and other such paradigm breaks – are precisely the situations where human strategic judgement can and does make a big difference.

AI, by design, can’t take a position: Ok, so let’s say, one can configure a robust system that can ingest and process both in-house, proprietary data and external signals. Still, it will only deliver scenarios, probability distributions and risk matrices. It will continue to hedge its bets and shy away from delivering a definite, specific recommendation, and be willing to stand behind it. This is a structural, systematic feature of AI.

Strategic business decisions need human strategic judgement that can condense and distil uncertainty into a judgment, and to be accountable for it.

Addressing the constraints for better decision making

We, at Geosynthesis, unambiguously believe that strategic advisory, driven by differentiated human judgement, can help senior business leaders address these structural constraints of their internal AI analytical function.

In fact, as enterprise AI systems get more evolved and mature, C-suite and boards will increasingly require an external perspective that can validate, challenge, and augment the former.

Strategic advisory should not, and cannot, replace internal capability. What it can do is to make your internal capabilities safer to rely on, by producing intelligence that does not reside in public datasets, creating analytical frameworks detached from your organisation's frame of reference, and delivering unbiased, independent judgment.

Delivering advisory value

So what does all of this translate into eventually? Well, practically speaking this is not a positioning argument, rather a mandate to deliver differently.

Business leaders need to go beyond relying on largely informational AI-driven, commoditised intelligence – comprehensive evaluation of macro trends, trend analyses synthesising publicly available information, and scenario planning matrices capturing multiple potential outcomes.

C-suite does not need yet another briefing document. Strategic advisory’s value proposition needs to be fundamentally different in its DNA: crisper, sharper, opinionated, and directly linked to a specific decision a specific client faces.

The deliverable is not a briefing report, but a clearer strategic choice – underpinned by domain expertise and relational intelligence – with a trusted counterpart who can take accountability for the recommendation being delivered.

The rise of AI, in our view, does not render strategic human judgment in the geopolitical advisory landscape less significant. On the contrary, it highlights exactly where the value of strategic intelligence lies, and raises the bar for delivering it credibly.