The Engagement Model

We embed engineers.
You own the AI.

Navedas isn't a SaaS subscription. Two field engineers embed in your operation, build the CX or back-office function around your actual policies, and leave you operating it on infrastructure you already own. Five stages. Three commercial milestones. One engagement.

Five-stage Co-Build

From your first CSV
to your team running the engine.

Onshore-only delivery available for HIPAA, FedRAMP, and EU-sovereignty workloads.

01 · Diagnose
48 hours · $500 fixed fee
1,000 decisions mapped to your policies
Send a CSV of recent decisions. We run them against your actual Context Graph and surface every off-policy line item with dollar exposure quantified. Productized scope. Try the engagement before you commit.
02 · Discover
Week 1 · Two FDEs on the ground
Your policies, your edge cases, your decision graph
Two field engineers spend a week with your operators. They read every policy document, sit with the supervisors, watch the queue, and codify the real decision rules, including the ones nobody wrote down.
03 · Design
Weeks 2–3 · Architecture + Reasoning Ledger
The governance layer your operation actually needs
Decision Gate, Reasoning Ledger, Operator Console: wired to your Context Graph and your existing tooling. We design for your stack (data residency, SSO, audit trails), not a generic SaaS pattern.
04 · Deploy
Weeks 3–4 · Live in one policy area
Production traffic, with a fast rollback path
We deploy inside your VPC. First week is shadow-mode, second week is live in a single workflow with kill-switch. Your operators see decisions, citations, and verdicts in the Operator Console the whole time.
05 · Operate
Ongoing · You own the engine
Your team runs it. We tune, ship audit packs, handle incidents.
The Operator Console, the Context Graph, the Reasoning Ledger. Yours. We stay embedded as the ongoing engineering layer: monthly audit packs, model tuning, incident response, scope expansion to the next workflow.
Three commercial milestones

Diagnose → Build → Operate

Each tier is a checkpoint with explicit scope and an outcome you can verify before the next.

Tier 1 · Diagnose
$500 Audit
48 hours · fixed fee · money-back
1,000 of your decisions analyzed against your policies. Revenue Recovery Report shows where the leak is. $25K minimum finding or your money back. Productized scope. Try the engagement before you commit.
Start your $500 audit →
Tier 2 · Build
$10K–$20K / mo
30 days to live · two FDEs embedded
Field engineers embed for one policy area. We build that part of your operation around your actual policies; you run it. Reasoning Ledger on every decision, Operator Console for your team, weekly exposure-reduction report.
Start the engagement →
Tier 3 · Operate
$120K+ / yr
Continuous engineering · every workflow
Engineers stay embedded for ongoing tune and incident response. Every policy area governed. Context Graph and Reasoning Ledger at scale. Executive Dashboard, Operator Console, SSO, SLAs. You own the platform.
Talk to the inventors →
What you own at the end

Stop the engagement
and you still own the AI.

The engine is yours. The policies are yours. The precedents we trained on your data are yours. The engagement is engineering; the platform is your asset.

Policy Graph

The encoded version of your actual policy library. Every rule, threshold, exception. Yours to extend, audit, or hand to the next vendor.

Reasoning Ledger

Every decision the engine made, with citation to the policy clause that drove it. Audit-ready by construction. Stays in your VPC.

Operator Console

The interface your team uses to review, override, and approve decisions. Branded for your operation, runs on your infrastructure.

Trained precedents

Every verified resolution becomes a precedent the engine learned. After 1,000 tickets, it knows your playbook better than your best supervisor. That training data is yours.

Who this is for

Eligibility, in plain commercial terms.

→ Eligible operations: any company running AI in customer-facing or back-office decisions where dollar exposure, policy compliance, or audit trails matter. CX, fulfillment, billing, claims, compliance.
→ Data residency: deployable in your VPC. Onshore-only delivery available for HIPAA, FedRAMP, and EU-sovereignty workloads. No training on your data.
→ Compliance posture: SOC 2 Type II, ISO 27001, GDPR, HIPAA Ready. NDA before any data transfer. Full audit pack on request.
→ Commercial structure: Diagnose is fixed-fee productized. Build and Operate are scoped engagements with explicit deliverables. No subscription auto-renewal. Outcome-bound.
Engagement FAQs

The questions that actually come up.

What happens if we stop the engagement?
You own the AI we built. The Context Graph, the Reasoning Ledger, the Operator Console, all the precedents we trained on your data. Yours. The engagement is engineering; the engine is your asset.
Why not just sell us a subscription?
A subscription is wrong-shaped for what we do. Configurable SaaS assumes you'll adapt to the product. We build the product around your actual policies. That requires engineers on the ground, not docs.
How long until we see value?
48 hours for the Diagnose audit (dollar figure, policy citations, recovery path). 30 days for live production traffic in one policy area at Build tier. Operate tier is continuous from there.
Where do the engineers come from?
Internal. Two field engineers from the team that built CSAT.AI (in production since 2021). No SDRs. No partner-referred contractors. The people writing the code are the people you'll be talking to.
Can we expand to another workflow later?
Yes. Build tier covers one policy area; expansion to additional workflows is scoped per area. Operate tier covers every policy area you run. Engineers stay embedded for the lifetime of the engagement.
Start here

See the leak before you commit.

Send 1,000 tickets. 48 hours. $25K minimum finding or your money back.

Start your $500 audit → Talk to the inventors →