SmartStaff.io vs ServiceNow AI Agents

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

sn_aia_execution_plan is 'the top-level record for a single agentic workflow run', but the documented tooling around it is post-execution forensic analysis.

Can a person read that plan before the work starts?

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

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

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?

MARKETING CLAIM ONLY. The product page asserts agents 'monitor results, incorporate feedback, and adapt their approach… continuous learning', but no mechanism is documented anywhere. Treat as unsubstantiated.

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

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

Can the customer run it in infrastructure they control?

Is the product interface available in Czech?

Czech is reported among roughly 24 language packs, but this could not be confirmed against the primary language-support page.

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

Can an individual tool call be gated mid-execution?

Per-tool Execution mode: Supervised ('will ask user for permission') or Autonomous. ServiceNow's own guidance: autonomous for reads, supervised for writes.

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

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

A 'Version control for AI agents and agentic workflows' page exists in the Australia docs but its rollback semantics could not be rendered.

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?

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)?

ISO 27001 since 2012, plus 27017/27018/27701/9001/22301 and SSAE 18 SOC 2 Type 2 — and notably ISO/IEC 42001 (AI management systems), the only vendor in this matrix confirmed to hold it.

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?

Can an outside system start a run over HTTP?

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

Inbound Email is a supported trigger, but 'The AI Agents team has deprecated support for creating AIA triggers directly on the sys_email table. Instead, use Flow Designer with Inbound Email Triggers' plus the 'Trigger AIA use Case from email' action.

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

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

SPP EU (Protected Platform for the European Union), opt-in via contract amendment: EU data centres, EU-based support, IP geofencing, formal Break Glass procedure requiring customer approval.

Can two versions be compared side by side?

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

Can work be started from a chat platform?

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

The SPP EU white paper never states where LLM inference occurs. Now LLM is proprietary; third-party models arrive via Gen AI Controller (Azure OpenAI / Gemini / AWS Claude, varying by product and region); with BYOLLM 'SPP EU does not by itself control where the BYO model runs'.