Example 01
Structure missing evidence
"Two of the four role-evidence vectors are thin. Here are the open questions a consultant would normally ask before grading is safe."
Companion Intelligence
One governed advisory layer — six modes, many agents, zero autonomous decisions.
Evalio Companion Intelligence is not a scattered set of chatbot features. It is one governed advisory surface that supports operators and consultants while every material workforce decision remains with the named human owner.
AI prepares · Consultant reviews · Client decides.
This is the operating sequence for every evidence review, scenario, and decision packet the Companion prepares.
Governance narrative
Every consequential workforce decision travels one governed path.
This is the operating model that connects the Decision Room, Companion Intelligence, Trust, and Proof. Each stage has a named human owner. Companion assists; it never approves, grades, or decides.
- 01Evidence
Role evidence, decision context, and source-rights posture enter under a structured intake. Gaps and contradictions are surfaced, not hidden.
Accountable: Operator · Evidence Agent prepares - 02Evaluation
The protected methodology runs structure, grading, and scenario logic against the assembled evidence — a bounded evaluation, never a finalised decision.
Accountable: Protected methodology - 03Review
A consultant reviews the prepared evidence and applies professional judgment before anything advances.
Accountable: Consultant reviewer - 04Challenge
Assumptions are tested, alternatives weighed, risks named, and open questions recorded against the record before approval.
Accountable: Reviewer · challenger - 05Approval
A named human approver accepts the decision basis at a governed approval gate, or sends it back.
Accountable: Named approver - 06Decision
The accountable owner decides. The decision, its rationale, and its evidence basis are bound together as the governed decision of record.
Accountable: Client decision owner - 07Artifact
A bounded, representative artifact is generated — evidence-dependent and traceable, never a final automated decision.
Accountable: System · bounded output - 08Audit Trail
Evidence, reviewer, rationale, approver, and timestamp are recorded append-only and stay re-openable.
Accountable: Append-only record - 09Institutional Memory
The governed decision becomes durable precedent — traceable context the institution can reopen and learn from.
Accountable: Institution
Accountable
A named human owns every consequential decision. Companion Intelligence prepares evidence and frames options under this sequence — it does not approve, grade, set pay, or override governance.
Six modes
One Companion. Six bounded modes.
The Companion is structurally bounded. Each mode has a stated purpose, what it is allowed to do, and what it is structurally blocked from doing. The boundary is enforced — not by convention.
Mode
Guide
Orient users through workflows, required inputs, next steps, and safe actions.
Allowed
Explain workflow state, required evidence, missing steps, and how to proceed.
Blocked
Cannot decide grade, pay, promotion, commission, entitlement, legal status, or final approval.
Mode
Evidence
Structure evidence, flag gaps, detect contradictions, and prepare reviewable summaries.
Allowed
Extract, classify, compare, and summarize evidence with caveats.
Blocked
Cannot fabricate missing facts, invent market data, or treat weak evidence as conclusive.
Mode
Mentor
Teach credible Total Rewards concepts in clear language.
Allowed
Explain concepts such as compa-ratio, range position, pricing readiness, source rights, and governance status.
Blocked
Cannot reveal protected scoring logic, scoring weights, thresholds, or methodology mechanics.
Mode
Decision
Support human decision-making by summarizing evidence, caveats, and tradeoffs.
Allowed
Prepare decision considerations, evidence packs, and rationale drafts.
Blocked
Cannot make, approve, lock, or finalize decisions.
Mode
Scenario
Test assumptions and compare controlled alternatives.
Allowed
Model what-if alternatives with assumptions, caveats, and review status.
Blocked
Cannot present scenarios as approved outcomes.
Mode
Growth
Consultatively point users to relevant Evalio capabilities when evidence shows need.
Allowed
Suggest relevant modules based on observed friction, blocked workflows, or repeated evidence gaps.
Blocked
Cannot pressure, manipulate, create false urgency, or push irrelevant upsell.
Agent map
A governed family of agents — bounded scope, governed handoff.
Each Companion agent has explicit scope. None acts outside the bounded Companion Policy. None replaces consultant or client judgment. The agents compose; they do not decide.
Buyer & onboarding
Buyer Guide Agent
Helps a buyer locate the right Evalio surface for their decision.
Onboarding Agent
Walks an operator through the six governed onboarding sections.
Data Readiness Agent
Surfaces what evidence Evalio still needs before pricing or grading is safe.
Evidence & methodology
Evidence Agent
Structures role evidence, flags contradictions, prepares reviewable summaries.
Survey / Source-Rights Agent
Tracks survey provenance and source-rights posture for compensation evidence.
Job Architecture Agent
Supports role mapping into a coherent job architecture under consultant review.
Pay, market, and reward design
Market Intelligence Agent
Surfaces indicative market context — never fabricates benchmark data.
Benefits Insight Agent
Frames benefit and allowance posture for consultant review.
Reward Design Agent
Prepares reward-mix scenarios for consultant and client review.
Pay-for-Performance Agent
Frames performance-to-pay linkage scenarios without making the decision.
Offer Governance Agent
Supports offer review against role evidence and structure posture.
Scenario, workforce & decision
Scenario Agent
Models what-if alternatives with explicit assumptions and caveats.
Workforce Planning Agent
Frames workforce cost and capacity scenarios for consultant review.
Decision Agent
Prepares decision packets, rationale drafts, and review context.
Reporting, dashboards & audit
Report Generation Agent
Drafts governance-safe report sections — bounded, cited, reviewable.
Dashboard Insight Agent
Frames dashboard context with caveats; does not assert conclusions.
Audit Agent
Surfaces governance posture and override events for review.
Partner, practitioner & voice-of-customer
Consultant Copilot
Supports the consultant in evidence review, scenario framing, and decision drafting.
Partner Academy Agent
Helps partner practitioners use Evalio under the governance contract.
VOC Agent
Captures voice-of-customer feedback for governed product improvement.
Agent depth is governed by workspace activation and entitlement. Public proof does not expose agent mechanics, scoring logic, or protected methodology. Some agents are workspace-gated; the buyer-facing agents are visible here as bounded previews.
Boundaries
What the Companion will never do — enforced at architecture level.
- Cannot assign a grade autonomously.
- Cannot approve a pay or compensation decision.
- Cannot issue a promotion decision.
- Cannot determine commission, incentive, or variable-pay entitlement.
- Cannot determine benefits, severance, end-of-service, or actuarial entitlement.
- Cannot produce a legal conclusion or compliance certification.
- Cannot expose scoring weights, thresholds, or protected methodology mechanics.
- Cannot operate outside the bounded Companion Policy.
- Cannot fabricate market data, survey data, or client data.
- Cannot take a material action on a decision without a recorded human reviewer.
The Companion assists. A named human reviews. The Decision Room records.
No Companion output reaches a workforce decision without consultant review and a named human approver. The Companion is advisory only — it never assigns a grade, approves pay, or issues an approval.
Representative Advisor Portrait · Human-Reviewed Process · Not Client DataPublic-safe examples
What the Companion sounds like when it is doing its job.
Example 02
Explain pricing readiness
"This role is not ready for pricing yet — the survey source-rights posture is unconfirmed, and the role evidence has a contradiction the reviewer should resolve."
Example 03
Prepare a scenario for review
"Here are three reward-mix scenarios with their assumptions and caveats. None of these is an approved outcome — the consultant and client decide."
Example 04
Frame a decision packet
"This Decision Room record needs: the evidence packet (linked), the scenario considered, the reviewer rationale, and the named approver. The Companion does not approve."
Workflow
How a Companion interaction is governed end to end.
No Companion output reaches a workforce decision without consultant review and a named human approver. The workflow is the boundary.
01
Intake
Operator provides role, evidence, or scenario context. Companion confirms scope.
02
Structure
Companion structures the evidence and flags gaps or contradictions.
03
Frame
Companion prepares a bounded interpretation — confidence-explicit, caveat-explicit.
04
Review
Consultant reviews the prepared output and applies judgment.
05
Decide
Client decides. The decision is recorded in the Decision Room with rationale.
06
Govern
Every invocation is logged. Overrides are recorded. The audit trail is preserved.
In-module Companion
Where the Companion shows up inside Evalio.
Job evaluation
Evidence Agent + Consultant Copilot frame the role for the reviewer.
Salary structure
Reward Design Agent prepares structure scenarios for consultant review.
Merit cycle
Pay-for-Performance Agent frames cycle alternatives; the client decides.
Workforce planning
Workforce Planning Agent models scenarios with explicit assumptions.
Decision Room
Decision Agent prepares the packet; the named approver decides.
Onboarding
Onboarding Agent + Data Readiness Agent govern intake without deciding readiness.
Bounded showcase
See the Companion render with the boundary visible.
Below is a representative Companion surface mounted in a methodology context. The governance line at the bottom is non-removable.
Representative preview · not client data.
Next steps