AI pricing hits the work graph

On 7 Sep, Asana reported Q2 FY2027 revenue of $216.4m, up 10%, with a 10% non-GAAP operating margin and raised full-year guidance. The market still marked the shares down 13.7% after the earnings announcement. Reports pointed to a disappointing Q3 outlook, margin pressure from the AI push and RBC's note that the packaging transition creates a revenue timing headwind.

The read is not that investors dislike AI. It is that the category has moved from AI demos to AI accounting. CFO Dive had already reported on 4 Sep that Asana's AI pricing shift may create near-term revenue pressure. That made the earnings reaction less a surprise than a test case: growth, better margins and AI ambition were not enough when the charging model was still moving.

Atlassian made the same issue explicit from the other side. On 6 and 7 Sep, reports covered a new usage-based pricing model for AI features in Jira's parent suite, with thresholds and excess usage charges, positioned around predictability and cost control. That is the cleaner end-state: AI becomes a metered resource, not a slogan hidden inside seat expansion.

monday.com made the week's bluntest packaging move. On 6 Sep, advanced columns were removed from the Pro tier. On 7 Sep, three tools were removed from the Basic tier and advanced columns were again recorded as removed from Pro.

That matters because it landed in the same window as Asana's packaging transition and Atlassian's AI metering. Three different vendors moved around entitlement, usage or AI pricing inside a few days. This was not a week of grand product launches. It was a week of deciding which capabilities remain included and which become paid gates.

For operators, the signal is practical. Work management vendors are no longer treating the work graph as a flat collaboration surface. Columns, tools, agents and usage are being separated into billable layers. The vendors that explain those layers cleanly will have an easier time with procurement than those that let customers discover the boundary at renewal.

Airtable's week began with ownership change. Reports on 4 Sep and 6 Sep said Bending Spoons had completed its acquisition of Airtable in an all-cash deal, with no deal value disclosed in the observed coverage.

Then the homepage started moving. Between 5 and 7 Sep, Airtable's hero rotated among three positions: AI upgrading teamwork, builder speed becoming business momentum, and all teams and workflows connected in one workspace. The supporting copy also moved between system-of-record language for workflows, data and agents, and AI-powered workflow language around unified data and collaboration.

That sequence is the kind of small public motion that reveals a larger strategic argument. Airtable is choosing between three claims: it can be the AI teamwork layer, the fast builder platform or the connected workspace for every team. All three are credible from its product history. Only one can be the first thing a buyer remembers.

Notion's week was more focused. On 6 Sep, reports said it had partnered with CJ to create an AI-driven research and development knowledge system for the Korean market. Separate coverage said Notion's CEO is making an all-in bet on AI and planning a hiring binge.

That is a different posture from the public-market pricing debate. Notion's observed move tied AI to a named enterprise knowledge use case and a regional customer context. It still fits the same category shift, but through deployment proof rather than pricing mechanics.

ClickUp was quieter. It launched a new blog page focused on its task dashboard and removed a blog post comparing Zapier and IFTTT. In a fortnight dominated by pricing transitions, acquisition messaging and earnings reactions, that restraint is itself informative. ClickUp did not surface a comparable pricing or AI commercial change in the observed period.