Three independent research outputs published between January and March 2026 describe, from different angles, the same underlying problem.
Forrester's research for SAP identifies fragmented strategies, siloed data, and poor communication of vision as the core barriers preventing organisations from achieving lasting transformation — and finds that only 6% of organisations are true leaders in change. Two Harvard Business Review studies on AI in organisational settings find that AI systems use markedly different question mixes than human leaders, underemphasising the productive and subjective questions that surface what's unsaid, and that large language models exhibit a consistent bias toward recommending strategies that align with managerial buzzwords rather than context-specific logic.
Taken separately, each finding is significant. Together, they describe an organisation that is fragmented in its processes, over-reliant on AI tools that flatten rather than sharpen collective thinking, and systematically blind to the subjective, experiential, and contrarian perspectives that genuine strategic insight requires.
The problem is not that organisations lack thinking tools. It is that the tools they are adopting optimise for what is easy to produce rather than what is most important to understand.
What the research actually shows
The Forrester diagnosis is that transformation fails at the organisational level because the people and functions that need to think together are not equipped to do so. Vision is communicated downward but not built collectively. Strategy is produced by small groups and consumed by everyone else as finished conclusions. The 6% who succeed are those who have made shared understanding structurally repeatable rather than dependent on exceptional facilitation.
The first HBR study adds a precise diagnostic: AI systems that facilitate organisational discussion systematically underweight the questions that matter most. Interpretive analysis is overemphasised; productive questions — those that surface implications and generate options — and subjective questions — those that engage values, stakes, and what people actually believe — are consistently underrepresented. The result is discussions that feel efficient and produce outputs that lack the epistemic depth that consequential decisions require.
The second HBR study is sharper still: on strategy, an LLM behaves more like a junior consultant parroting what's popular than a colleague who stress-tests assumptions. The term the researchers use is "trendslop" — strategies that align with managerial buzzwords rather than context-specific strategic logic.
Why collective intelligence infrastructure is the answer
The organisations that will think well about complex change are not those that delegate sensemaking to AI summaries or strategy documents. They are those that build the structural conditions for genuinely plural, epistemically diverse collective intelligence — and retain human judgement at the centre of it.
Where AI systems produce trendslop, SparkMaps preserve the dissenting reasoning, peripheral perspectives, and productive anomaly that no LLM is designed to hold. Where Forrester calls for cross-functional structures that drive alignment, Hunome provides the epistemic architecture to make that structural rather than aspirational. Where HBR warns that AI underweights subjective and productive questions, Hunome's Ignite characterisation system gives tacit, experiential, and speculative ways of knowing equal structural weight to analytical evidence — making the invisible visible rather than averaging it away.
The research describes the problem clearly. The infrastructure to address it exists.
---
Sources: - Forrester Research for SAP, January 2026 — Organisational Transformation Readiness - Harvard Business Review, February 2026 — AI Question Bias in Organisational Discussion - Harvard Business Review, March 2026 — LLM Strategy Recommendations and Trendslop