Champion Enablement Kit

1. Executive summary

Aevesa is the independent Prove layer for production AI agents: runtime enforcement plus offline-verifiable evidence that Legal, Compliance, and auditors can trust without logging into a vendor dashboard.

The problem

Agents with tool access can change systems of record (payments, records, PHI, orders).
Gateways and guardrails govern routing and model boundaries, but leave a Prove gap: no portable proof of what executed, under which policy, with which human authority.
Most enterprises have agent policies on paper; few have full security sign-off to deploy agents in production.

What Aevesa adds

Route → Classify → Prove: keep your gateway and AI-SPM; add Aevesa as the Prove layer on the architecture diagram.
Intercept: block or pause high-risk tool paths before side effects.
Decide: structured human-in-the-loop when policy requires it.
Prove: tamper-evident receipts on the Aevesa Ledger, verifiable at verify.aevesa.com with no Aevesa login.

Business case (for economic buyer)

Risk without Prove layer	Outcome with Aevesa
Agent production blocked; AI ROI stalled in pilot	Legal and Compliance sign-off with auditor-verifiable artifacts
Post-incident forensics on vendor-scoped logs	Receipt-before-action + offline integrity check in under 90 seconds
EU AI Act Art. 12/14 retrofit after architecture freeze	Automatic logging + human oversight packs mapped to evidence patterns
Single-threaded Slack approvals with no chain of custody	HITL bound to cryptographic receipt and approval identity

Recommended land motion

Start with a 90-second Evidence Gap demo and a scoped attest or MCP proof-of-value. Expand to Art. 12/14 export packs and Assurance Autopilot once the security engineer validates offline verify.

2. Stakeholder talking points

Use the card that matches who your champion is briefing. Lead with Prove, not feature lists.

CISO / Security engineering

Pain: Agent blast radius, MCP supply chain, tool abuse at runtime.

Say: "If the ledger cannot commit, the tool does not run. Receipt-before-action is the invariant. Your team can verify a gateway deny receipt in 90 seconds without trusting our UI."

Demo: Evidence Gap + Developer verifier guide.

Compliance / Legal / Internal audit

Pain: Art. 12 logging and Art. 14 human oversight evidence; vendor-scoped JSON exports.

Say: "Your GRC tool holds the workflow. Aevesa holds the proof. Auditors paste a receipt at verify.aevesa.com and recompute integrity offline."

Attach: Trust Bundle, Enterprise verification guide.

CFO / Economic buyer

Pain: AI investment blocked; compliance retrofit cost; incident liability.

Say: "We are not buying another AI-SPM platform. We are unblocking production agent ROI with evidence Legal already asked for. Land is a scoped POC; expand when attest is live."

Frame: Cost of delayed agent deployment vs. cost of one unprovable high-risk action.

AI platform / Engineering lead (your champion)

Pain: Risk committee wants an architecture diagram with governance; MCP agents outside gateway path.

Say: "Keep Unity AI Gateway, Portkey, or Cyera for route and classify. Aevesa is the Prove layer: intercept, HITL, ledger receipts. Complement, not rip-and-replace."

Next step: Design partner / Fortress briefing.

Procurement / Vendor risk

Pain: Security questionnaire, DPA, proof of independent verification.

Say: "Open evidence: liability-receipt schema, public verify portal, Trust Bundle with OWASP ASI and EU AI Act prEN mappings. Verification does not require exporting raw prompts to a third-party analytics vendor."

3. Comparison sheet

Share when the buying committee asks "why not just use what we have?"

Capability	Status quo (logs + policy PDF)	Gateway + guardrails only	+ Aevesa Prove layer
Pre-execution block on tool calls	✗	Partial	✓
HITL bound to evidence chain	✗	✗	✓
Offline auditor verification	✗	✗	✓
Tamper-evident receipt chain	✗	✗	✓
EU AI Act Art. 12/14 export patterns	Manual	Partial	✓
Complements existing gateway / Cyera	N/A	✓	✓

Kill line for platform logs: "Platform logs are vendor-scoped autopsy reports. Aevesa receipts verify offline with no Cyera login, no Databricks access, no trust in our dashboard."

4. Objection handling

Pre-answer the objections that stall deals in Legal, Security, and Procurement.

We already have an AI gateway (Portkey, Palo Alto, Unity AI Gateway).

Good. Keep it for route and observability. Gateways do not bind human approval to tool execution or produce offline-verifiable receipts. Aevesa cosigns gateway decisions and closes the Prove gap.

We have prompt guardrails (Lakera, cloud safety SKUs).

Guardrails govern text at the model boundary. Permitted but misaligned actions still pass through. Aevesa governs actions at runtime and proves outcomes on the ledger.

We want one AI-SPM vendor, not another tool.

Position Aevesa as the Prove add-on SKU, not a replacement AI-SPM. Revisit when audit asks for offline verification or EU AI Act Art. 12 evidence is on the critical path.

Microsoft released an Agent Governance Toolkit. Why not build in-house?

The toolkit supplies enforcement primitives you operate yourself. Aevesa supplies managed HITL workflow, production export packs, hosted verify infrastructure, and auditor-ready evidence product. Building ledger + open verify + Art. 14 packs in-house is typically 12 to 18 months; Art. 12 architecture should be designed in now.

Our Cyera / UAG demo was enough for security.

For discovery and block, yes. Ask internal audit: can they verify a high-risk agent action offline, without logging into Cyera or Databricks, six months after the event?

We do not have production agents yet.

Nurture with the Evidence Gap link. Re-engage at the milestone when agents get write access to systems of record or the board adds AI risk to the scorecard.

EU AI Act is too early for us.

Art. 12 automatic logging architecture must be designed before production freeze. Retrofit is expensive. Start with a scoped attest POC and one exportable receipt.

How is this different from LangSmith / Langfuse traces?

Traces help engineers debug in dev tooling. Receipts prove what was permitted under policy, tamper-evident and auditor-grade, with independent verify at verify.aevesa.com.

Security review will take quarters.

Lead with Trust Bundle, Open Evidence API, and a 90-second offline verify demo. Pre-load the security questionnaire answers and verification guide before Procurement opens a formal assessment.

What if we only use Copilot with read-only access?

Likely disqualifier today. Aevesa fits when agents take write actions on systems of record or MCP tools run outside a single gateway path.

5. Internal email template

Your champion can copy, personalize names, and send to their CISO, Head of Compliance, or platform steering committee.

Subject: Proposal: Add the Prove layer for production AI agents

Hi [Name],

Our agent pilots are stalling at the same gate: Security and Compliance will not sign off on production write access without evidence they can verify independently of vendor dashboards.

I recommend we evaluate Aevesa as the Prove layer on our agent stack (Route → Classify → Prove). We would keep [gateway / Cyera / existing tools] for routing and classification. Aevesa would:

1. Intercept high-risk tool paths before side effects
2. Route sensitive actions through structured human approval when policy requires it
3. Emit tamper-evident receipts verifiable offline at verify.aevesa.com (90-second demo, no login)

Why now: [EU AI Act Art. 12 timeline / board AI risk review / upcoming audit / agent write-access milestone]

Suggested next steps:
• Security engineering runs the Evidence Gap demo: https://verify.aevesa.com?demo=gateway
• Compliance reviews the Trust Bundle and verification guide
• Scoped 2-week POC on [gateway attest / MCP path] with success criteria: one offline-verifiable receipt in auditor workflow

Happy to walk through this in 30 minutes. I am not asking to replace our current stack, only to close the Prove gap our architecture diagram is missing.

[Champion name]

6. Mutual action plan

Track progress with your Aevesa contact. Adjust dates to your procurement calendar.

Week 1Champion shares this kit + Evidence Gap demo link with Security and Compliance stakeholders.
Week 1Security engineering validates offline verify (developer guide).
Week 230-minute briefing with platform + security + compliance (Aevesa SE optional).
Week 2Compliance receives Trust Bundle + sample Art. 12/14 export structure review.
Week 3Scoped POC kickoff: gateway attest webhook or MCP intercept path defined.
Week 4POC exit: one production-representative receipt verified offline by internal audit or security.
Week 5Commercial: security questionnaire, DPA, and economic buyer ROI review.
Week 6+Expansion decision: Vanguard / Compliance-in-a-Box tier, GRC integration, Fortress if VPC required.

POC success criteria (agree in writing)

At least one high-risk path intercepted or approved with HITL.
Receipt verifies at verify.aevesa.com without Aevesa dashboard access.
Compliance stakeholder confirms export fits their auditor workflow.

7. Conformance badge embed

Paste into slides, Notion, or partner microsites. 58 gateway checks (36 webhook + 22 OTLP) with Big Four vendor coverage.

PowerPoint / Keynote / PDF (static)

Insert picture from URL:

https://aevesa.com/assets/badges/conformance-gateway-58.svg

Live slide footer (script)

<div data-aevesa-conformance-badge data-layout="deck"></div>
<script src="https://aevesa.com/js/aevesa-conformance-badge-embed.js" async></script>

Open copy-paste page with live previews · Guide: docs/sales/CONFORMANCE_BADGE_EMBED.md