Tability vs ZenTao

Tability is an OKR and strategy platform that spent 2026 turning a plan hierarchy into a queryable graph, then pointing agents at it. This batch is consolidation rather than new ground: workspace admins can now switch Agents off from MCP settings while preserving existing schedules, relationships and history; AI Mode gains write access to KPIs and KPI Views with explicit confirmation on destructive actions; scheduled prompts can read across plans, objectives, key results, initiatives, check-ins, members, agents and KPIs to produce formatted executive emails. Setup for delegated agents now hands you install instructions for Codex, ChatGPT or Claude Cowork.

The agent track and the graph track have merged, and the work has shifted to the unglamorous half: permissions, confirmations, admin kill switches, and telling a human which agent needs attention and why. AI Mode is steadily converting modules from readable to writable - it took general workspace actions in May, the KPI module shipped in July, and it now writes KPIs too - so the pattern is that each new object type gets inspected first and edited a release or two later. Notifications now distinguish updates authored by an agent from those authored by a person, which is the kind of detail that only matters once agents are actually writing to the workspace.

Expect initiatives and check-ins to be the next objects AI Mode can write rather than only read, following the same inspect-then-edit pattern the KPI module just completed. The admin enable/disable switch also suggests per-role or per-team agent permissions rather than a single workspace-wide toggle.