Faster execution, wider strategic drift

Most executive teams don't lack a strategy. They lack proof that the work happening beneath them is connected to one. For an organisation adopting artificial intelligence at blazing speed, the drift is more serious now than it ever was: it burns tokens and resources, it slows momentum, and it widens the gap between what you intend and what you actually build.

Clarity is not the constraint

Four in five leaders say their strategic priorities are clear. Fewer than one in ten can show that most of their portfolio work is explicitly linked to those priorities. Atlassian's 2026 research, Leading the human + artificial intelligence enterprise, built on interviews with 308 senior leaders, calls this the clarity trap: we keep getting clearer about direction and expecting execution to follow. It doesn't.

Clarity tells you what you intend. Connectivity is what the organisation actually does with that intent. The machinery of the last twenty years made you better at reporting on intent. It did not make you better at executing it.

Changing the priority is the easy part

Asha Thurthi makes this point sharper in her demo of Strategy Collection at the Atlassian Teams'26 conference. Priorities change all the time, she explains, but changing the priority is only the first decision. The real difficulty sits downstream: does the investment change with it? Does the money and capacity follow the new direction, and how long does the whole realignment take? The practical test for leadership is not whether priorities shift, but whether investment and execution follow them, and how quickly. This is why Atlassian's Strategy Collection is worth a look; the session demo lays out the shift plainly.

The question every Chief Executive Officer should face: if a priority changed this morning, what work, funding, and capacity would move, and would you even know it moved this quarter? For most enterprises the honest answer requires a fresh round of status requests, spreadsheet reconciliation, and a dashboard assembled from scratch. By the time the picture is complete, it is stale.

More planning is not the answer

Atlassian's finding is blunt: how often an organisation re-ranks its priorities has essentially zero predictive relationship with decision speed or artificial-intelligence effectiveness. A revised strategy is not a pivot unless it reaches the execution layer.

The knowledge layer is where it gets served

As the mix of people and agents shifts, strategy has to reach more than people. A changed priority reaches a team through a standup, a meeting, or a channel; an agent reaches context through a layer, and no meeting updates that layer. If the context agents act on stays stale, the drift compounds silently. The artificial-intelligence context layer has to be kept current.

The real bottleneck is becoming the knowledge layer: the connected, current, owned record of what an enterprise actually means and wants, the layer underneath every tool, meeting, and decision. The shift is not toward more tooling, but toward a different layer: knowledge that is machine-readable, versioned, and connected to the work.

Then: reporting on intent

• Status requests
• Spreadsheet reconciliation
• Dashboards built from scratch
• Decking around an answer

Now: executing on intent

• strategy.md, owned intent
• Knowledge bundle, a structured, versioned set of markdown files
• Connected work data

A knowledge bundle is a directory of typed, cross-linked markdown files, owned by someone accountable on the intent side, not the reporting side. It is how a company stops reassembling its own strategy every quarter and starts executing it.

2027 is the year of knowledge

The tooling arms race was built for a slower era. We are entering the year where context carries more weight than compute in deciding who executes on intent and who just reports on it. Gartner now tracks artificial-intelligence context platforms as a standalone market, estimated at $28 billion in 2026, and its planning assumption is that by 2027 organisations that prioritise semantics in their artificial-intelligence-ready data will gain up to eighty percent in agentic accuracy and cut costs by up to sixty percent.

I foresee 2027 as the year of knowledge. The question for every enterprise is whether its knowledge layer is ready, or whether it will keep drifting, faster, in the wrong direction.

Follow for Part 2: turning a knowledge layer into a working operating recipe, the practical path for an enterprise to take this from essay to adoption.

Sources

1. Atlassian, Leading the Human + AI Enterprise: A New Playbook for Strategic Portfolio Management (2026), primary research with 308 senior enterprise leaders covering the clarity-trap finding (80% clear / 11% linked), the null finding on re-ranking frequency, and domain connectivity as the strongest predictor of AI effectiveness. atlassian.com/collections/strategy/resources/state-of-strategy-report
2. Atlassian blog, Shift happens: Finding and fixing the gap between strategic intent and actual work (28 September 2026), the companion write-up to the report and the basis for the drift framing here. atlassian.com/blog/leadership/gap-between-intent-and-work
3. Gartner, How to Navigate the Emerging AI Context Platform Market, Market Overview for AI Context Platforms (1 September 2026), covering the $28 billion 2026 figure and the 2027 semantic-first accuracy/cost planning assumption. gartner.com/en/documents/8335617