AI's first organisational gain is transparency, not automation
The dominant industry focus on agentic automation is misallocated for most organisations; the more significant near-term gain is making documented organisational knowledge reliably accessible through AI — automation comes later, and depends on the transparency arriving first.
Industry discourse about AI in 2026 is dominated by agentic automation: workflows that AI executes end-to-end, complex multi-step processes that run with minimal human involvement, vendor demonstrations of agents that book travel or close tickets or write code without supervision. The centre of gravity in product marketing, conference keynotes, and analyst reports is the agent.
This is a contested view, but the position taken here is that the agentic centre of gravity is misallocated for most organisations. The reliability and governance scaffolding required to run agentic systems on operationally important work does not yet exist in most environments, and the productivity gains from agentic systems outside coding contexts are likely to be modest in the near term. Organisations deploying agents on real workflows are spending significant effort on the substrate problems — observability, error handling, escalation, audit — that the agentic framing tends to underestimate.
The pattern is that the more significant near-term gain, available now and at lower cost, is organisational transparency. Most organisations already have documented procedural knowledge sitting in document stores, wikis, ticketing systems, and intranets. Most of that knowledge is not reliably findable; the people who could answer the question often don’t know that the answer is documented; the documentation that does exist is often stale or scattered. AI changes this. A retrieval layer over the existing knowledge — properly structured, with attention to source quality and confidence signalling — turns latent organisational knowledge into a daily working asset. Staff get to answers faster; new joiners onboard faster; the organisation knows things it didn’t know it knew.
The order of operations matters. Transparency is a prerequisite for the more advanced capabilities the agentic framing assumes: on-the-fly diagrams, on-demand dashboards, personalised guidance based on user context, eventually agents that act on the organisation’s behalf. None of those work reliably without a foundation of accurate, accessible knowledge. See Make the firm itself a Claude project for the operational shape this typically takes, and Useful AI is a context problem for the broader principle.
The position is honestly framed as opinion rather than settled fact. The next few years will tell. The case for putting transparency first is that it is achievable now, generates compounding capacity, and is the foundation everything else needs anyway.