SmartStaff.io vs Google Gemini Enterprise

Does each execution produce a plan artefact before the first unit of work runs?

Can a person read that plan before the work starts?

'Transparent thinking' (GA 2026-07-27) 'shares its real-time reasoning and planning in the user interface before calling tools' — display only. PlanReActPlanner streams PLANNING and ACTION in one continuous turn.

Can a person rewrite, reorder or drop steps and then approve?

Is there a plan/run confidence score derived from checkable structure rather than asked of the model?

Does a correction on a finished run alter behaviour on subsequent runs without a human editing a prompt?

Feedback service (Preview, 2026-07-29) is to 'collect, analyze, and manage end-user qualitative feedback… troubleshoot your agents'. No statement that later runs change.

Can a person list what the system has learned, see where it was applied, and retire it?

Can Claude/ChatGPT approve or reject pending work through the MCP surface?

Does the platform mint an address that starts a run, as opposed to polling a mailbox you own?

Can the customer run it in infrastructure they control?

Is there a documented EU location for the customer's stored data?

Is the product interface available in Czech?

No Czech — the Google Cloud docs locale switcher itself omits Czech. Czech is supported at model level only.

Can an individual tool call be gated mid-execution?

Is there version history with a way back to a previous version?

Split by tier: Agent Designer has no documented versioning and editing is destructive; Agent Runtime has 'immutable revisions of deployed agents, and split traffic'.

Is there a promotion path between environments, or git-backed source control?

Can you adopt a ready-made agent persona, as opposed to a workflow template?

Can a run start on a schedule?

'You must manually refresh your schedules every 14 days', maximum 5-minute delay, 'only supported in multi-region locations'.

Can work be started from a chat platform?

Is there a machine credential, so server-to-server callers need not hold a user session?

Is there a SOC 2 Type II report?

Is the vendor ISO 27001 certified (not merely 'aligned' or engaged)?

Is there a durable process definition distinct from an individual run of it?

Out of the box, does a human have to approve something before side effects happen?

Strictest default in the field: 'By default, all calls to any action require confirmation from the user, because Gemini Enterprise assumes that any operation is potentially destructive and can mutate data.'

Is there a review gate on the deliverable before it is sent?

Is a team of differently-roled agents a first-class product concept?

Can knowledge be attached to an individual agent and retrieved at run time?

Can it consume external MCP servers as tools?

Does it expose itself over MCP to external AI clients?

Is there a per-run record of what happened, step by step?

Is there a documented way to constrain where the prompt is processed, or to bring your own model?

Real EU machine-learning processing exists, but Gemini 3.1 Pro has neither residency nor MLP in the EU, Gemini 3.6 Flash is global-only, and Search grounding is global-only. Only Gemini 3.5 Flash is fully EU-supported, and Google's docs recommend global unless you have a regulatory reason.

Can two versions be compared side by side?

Is that MCP server protected by OAuth 2.1 with spec discovery, rather than a static bearer key?

Can an outside system start a run over HTTP?