Axceed Stewards

Accountability that's measured, not guessed.

Axceed Stewards is the Human Layer of BOS โ€” it turns every signal into a named owner, every owner into a measurable Steward Score, and every score into evidence instead of opinion. The Three-Tier Evidence Model combines live system signals, second-party-verified evidence, and structured multi-rater assessment into one continuous score โ€” no consultant scorecards, no spreadsheet quarterly reviews, and no self-graded outcomes anywhere in the chain.

Sign in to your dashboard

Built on a hierarchy and an evidence chain, not a form

๐Ÿงญ
Universal Evaluation Hierarchy
Evaluators resolve automatically from your real reporting line, with panel-size enforcement built in โ€” no one hand-picks their own review panel, and every role's Data Owner is derived from the org structure itself, not typed in by hand.
๐Ÿ“ถ
Tier 1 โ€” live system signals
KPIs auto-generate from connected system signals via Axceed Nerve and module integrations โ€” credibility locked at 1.0, no human entry, no stale quarterly numbers.
๐Ÿ“Ž
Tier 2 โ€” verified evidence
A Steward submits evidence โ€” a milestone, a report, a certification โ€” and a second party verifies it. A Steward can never verify their own evidence. Ever.
๐Ÿง‘โ€โš–๏ธ
Tier 3 โ€” structured multi-rater assessment
Behaviour and leadership are scored through a structured rubric across multiple raters, not one manager's opinion โ€” the smaller, most-governed slice of the score, exactly where bias risk is highest.
๐Ÿšช
Exit handover tracking
Reusable per-role handover templates, per-exit checklists, and an org-wide stalled-handover monitor โ€” with the same never-self-verify guard as evidence: the person handing over can't sign off their own handover.
๐Ÿ“ˆ
Performance Improvement Plans
Goals, check-ins, and outcomes grounded in the same evidence spine โ€” a PIP subject can't initiate their own plan, conduct their own check-in, or decide their own outcome. Enforced at the database level, not just the UI.

Frequently asked questions

How is a Steward Score actually calculated?

It's a weighted blend of all three tiers โ€” system signals, verified evidence, and structured assessment โ€” each carrying its own credibility weight, normalised to 0โ€“100. A Floor Rule caps the score at 60 if any Critical signal is breached and unacknowledged, regardless of how strong the other two tiers look.

Can someone verify their own evidence, approve their own PIP outcome, or sign off their own handover?

No, anywhere in the platform. Evidence requires a second-party verifier, PIP check-ins and outcome decisions block the subject from acting on their own plan, and handover items require an assignee-not-verifier split โ€” these are enforced guards, not policy asked nicely.

Is bias detection fully automated today?

Single-rater risk flagging is live and writing real flags โ€” the first of the Seven Fairness Principles checks shipped. Manager inflation/deflation, rater-clustering, and score-signal-divergence detection are designed and specified, rolling out next; we'd rather tell you exactly what's live than round up.

Do I need to be on the full BOS platform to use Stewards?

No โ€” Stewards runs standalone, with signals entered manually or via API from your existing systems, or integrated with Nerve and other Axceed modules if you're already on BOS. Same platform, either deployment mode.

Give every signal a named owner.

Sign in to your dashboard