Updated the milestone dates to align with DrupalCon Orlando for 1.1 though we could scope down the releases and do a 1.1 in the middle and 1.2 for Orlando... tbd Chatting with fable 5 re {width=900 height=510} With the real 1.1 window being ~6 months (late Sep → mid Mar), my earlier "keep 1.1 tiny" advice no longer applies — the slide's ambition level is reasonable for the timeline. Here's my reassessment. First, one slide bug: the 1.1 column says **"March 2026"** — should be **March 2027**. ## The main gap: the slide and the actual milestone don't overlap The six issues actually sitting in the 1.1 milestone today (performance prefiltering, selection diagnostics, access separation, consumer follow-ups, temporary subscriptions, the RAG plan) are almost all trust/hardening work — and *none* of it appears on the slide, while the slide's nine bullets are all new capabilities. For a status video that's fine rhetorically, but I'd add one bullet that represents the hardening theme, both because it's what early adopters most need to hear and because it's the work most likely to actually ship: - **"Performance & Scale Hardening"** (covers [#3586347](https://git.drupalcode.org/project/ai_context/-/work_items/3586347), [#3586167](https://git.drupalcode.org/project/ai_context/-/work_items/3586167)) — sites with thousands of context items is the 1.1 credibility story. ## Bullet-by-bullet verdict | Slide bullet | Backlog anchor | Verdict | |---|---|---| | Starter Templates | #3586255, #3586340 | **Keep** — demo-able, high adoption value, realistic | | Selection Tuning | #3577451, #3586349 | **Keep, or merge with Smart Selection** — to an audience these read as one thing | | Smart Selection | #3568679, #3586188 | **Keep but hedge** — LLM-powered ranking is the most research-y item; most likely to slip to 1.2 | | Context Editor AI Assistant | #3586254 | **Keep** — flashiest demo of the set | | Context Connector Plugins | #3586216 | **Keep if** the Connected Context discussion has converged enough to say out loud; otherwise it's the vaguest bullet | | Observability Integration | #3586242, #3567803, #3586165 | **Keep** — and it's actually scheduled, so it's a safe promise | | Tool API Support | #3574926 | **Merge with MCP** into "MCP & Tool API Integration" — same story (CCC context reaching external AI ecosystems), and #3574926 is still just an evaluation | | MCP Support | #3574416 | Merge as above | | Enhanced RAG Support | #3586352, #3547045 | **Keep** — "Enhanced" is doing good work here since the committed deliverable is a plan plus integration docs, not a RAG engine | That merge frees a slot for the performance bullet without growing the list past nine. ## Two adds worth considering - **Drupal 12 readiness** ([#3586269](https://git.drupalcode.org/project/ai_context/-/work_items/3586269)) — if D12 lands during the 1.1 window, this is a bullet audiences actively look for, and it's cheap to claim since the work is largely mechanical. - **Governance & Permissions** ([#3586357](https://git.drupalcode.org/project/ai_context/-/work_items/3586357), [#3586187](https://git.drupalcode.org/project/ai_context/-/work_items/3586187)) — only if you want to court the enterprise/compliance audience in this video; otherwise it folds into the hardening bullet. ## How I'd sequence the six months (for your planning, not the slide) - **Alpha1 (early):** the trust work — performance prefiltering, diagnostics, access separation — plus the scope-API refactor issue we drafted, since everything else builds on it. - **Betas (middle):** the feature bets — starter templates, editor assistant, MCP/Tool API, selection tuning. These are parallelizable and individually droppable to 1.2 without breaking anything else. - **RCs (late):** RAG/connector groundwork lands as *plans, docs, and extension seams* rather than engines — matching the "Enhanced" framing. The structural point: the slide currently promises nine outcomes with zero slack. Sequencing hardening first means that even in the worst case — every feature bet slips — 1.1 still ships a story ("faster, observable, governable at scale"), and each feature that does land is upside. That's also why I'd resist trimming the aspirational bullets further: with the "some may move to 1.2" caveat you're already planning to give verbally, the slide's job is to show direction, and this list points the right way.