Process Ownership

Most of what the AI Strategist does starts with watching and advising. Process ownership goes further: you hand it one process, and it runs that process’s queue for you. It watches every open run, checks in personally with whoever an overdue step is waiting on, follows up when nothing moves, and escalates a genuinely stuck run to you with a clear ask. Your team keeps doing the work; the Strategist makes sure the work keeps moving.

What the Strategist does as an owner

Once it owns a process, the Strategist runs a simple, relentless routine several times a day:

It watches every open run. A step past its due date, or one sitting far longer than that step normally takes, gets its attention. A day of grace comes first: people usually clear their own queues.
It checks in with the person the work is waiting on. A short, personal email: which step, which run, how long it has been waiting, and a link straight to the task. If something is blocking the work, the person can just say so - in a reply on the process or in a strategy session.
It follows up once. If the step still has not moved a few days later, it checks in a second time, and says plainly what happens next: it will raise the run for a decision so nothing quietly slips. It never nags beyond that.
It asks the person doing the work what’s actually wrong. When a run has to be raised for a decision, the Strategist first asks the assignee one direct question: what is this step actually waiting on? The person gets an email with a link that opens straight into a conversation with the Strategist about that exact item - no permissions needed, and if they leave and come back, the same conversation picks up where it left off. Their answer, quoted in their own words, shapes what happens next; silence for a few days means the Strategist proceeds on its stated assumption.
It escalates with the right decision teed up. The stuck run comes to you as a single clear item: what is stuck, who it is waiting on, what the Strategist already did, and what the assignee said. Based on the cause, it recommends one specific action - reassign the step to the person who has done it before with the lightest queue, extend the deadline when the blocker is outside your walls or time-bound, or cancel the run when the assignee says the work is no longer needed (nothing is deleted; a canceled run can be restarted). One approval and the Strategist carries it out itself; decline and it brings you the next-best option. When the run moves again, it closes the item out itself.
It reports its results, measured. The moment you hand a process over, the Strategist records how that process performed before the handover: the typical run time over the previous 90 days. That baseline never changes, and its owner’s report inside your AI Daily Briefing compares runs completed on its watch against it - honestly in both directions, and saying “too early to measure” until there is enough data to be sure. Runs in flight, what it chased, what completed, and anything waiting on your call appear alongside.

It runs campaigns against the numbers

Keeping runs moving is table stakes; an owner also answers for the trend. When the process’s typical run time drifts well past the baseline it recorded at handover, the Strategist opens an improvement campaign and says so plainly: “this number is mine until it’s back.”

A campaign is not a report - it is a sequence of concrete changes, pulled one at a time, each measured on its own before-and-after:

Give the bottleneck step a deadline. It finds the step where runs actually sit, and proposes a due-date policy based on that step’s own demonstrated pace - so slow runs start landing on its chase list automatically.
Rebalance the bottleneck step. When the slow step’s work all lands on one person by default, it proposes pointing new tasks at the person who has done that step before and has the lightest queue today. Tasks already in flight keep their assignee.

Each lever is a one-approval proposal; approve it and the Strategist makes the change itself, then measures for two weeks before judging. A lever that works gets held; one that does not gets replaced by the next. When the number returns to baseline, the campaign closes with the story of what did it - and if every lever is exhausted and the number still has not moved, the Strategist says exactly that and asks you to direct a structural change rather than pretending otherwise. Every campaign event is journaled, and progress appears in its owner’s report in your AI Daily Briefing.

When it has earned your trust

Every reassignment you approve builds a streak. Once you have approved five of the Strategist’s moves in a row, and at least three of the moved steps were genuinely completed soon after the move, it makes you a standing offer: let it route stuck steps on this process itself. The offer waits in your Decision Queue and changes nothing until you accept it.

There is no clock on the streak, only consecutiveness: a process that runs a handful of times a year earns trust at its own pace, and declining a recommendation simply resets the streak to zero rather than counting against the Strategist forever. Decline one early on while you are getting a feel for it, then approve the next five, and it has earned the offer.

Accept, and the routine becomes fully hands-off for the routine cases: when a step stays stuck after two check-ins, the Strategist moves it to the qualified person with the lightest queue on its own, records the move in its journal, and tells you about it in the Daily Briefing. Three guardrails never relax: it moves any given step at most once on its own (a step that gets stuck again always comes back to you), any run you are already deciding on stays yours, and a move you declined stays declined - if that run gets stuck again, it comes back to you rather than being moved on its own. Decline the offer and every move keeps coming to you for approval; you can also flip the setting back at any time.

What it will never do

Ownership has hard edges, and they are yours to hold:

It never changes how the process works on its own. Redesigning steps, reassigning defaults, or automating anything still arrives as a proposal for your approval, exactly like everywhere else.
Every action is on the record. Each check-in, follow-up, and escalation is written to the Strategist’s decision journal, so you can always see exactly what it did and when.
You can take the process back anytime. Pause or disable the AI Process Owner focus area (ask the Strategist in chat, or use the Focus Areas page under Administration → AI Strategist) and the handover ends immediately.
The account-wide autonomy switch applies. If your organization has turned off autonomous action for the Strategist, it stops sending check-ins too.

Hand a process to the Strategist

There are three ways ownership starts, and every one of them is your explicit decision:

Accept its offer. When the Strategist sees a process with real work in flight and steps regularly going overdue, it may offer to take ownership - one standing proposal that names the process and the evidence. Approve it and the handover is made; decline it and the Strategist sets the idea aside.
Ask for it in chat. In any strategy session, tell the Strategist to take ownership of a process by name. It plays the handover back to you - what it will do, how often it checks, how to take the process back - and proceeds only after your explicit yes.
Add it by hand. Administrators can add the AI Process Owner focus area from the Focus Areas page and point it at the process.

What your team experiences

People assigned to steps in an owned process notice exactly one change: when a step of theirs sits too long, the AI Strategist checks in with them directly - politely, with the task linked, and at most twice. Assignment notifications, task screens, and everything else about working the process stay exactly as they are.