1
Briefing
1–2 weeksWe scope the document job, deployment environment, data boundary, and review requirements. You confirm fit before any commitment.
Customer artifactFit memo and workflow boundary
Pilot Engagement
A fixed-scope evaluation for one sensitive-document workflow. Bring the job your reviewers already struggle with; we ship hardware, install adapterOS locally, and measure whether it is safer, faster, and easier to defend than cloud AI — without a token meter running in the background.
Start here
The pilot is for teams that want AI help on documents they cannot casually upload. Bring one workflow, the boundary it must respect, and the review process it has to fit.
Pilot phases
2
Scope & sources
2–4 weeksDefine the documents, workspace behavior, reviewer route, and acceptance criteria for useful source-backed outputs.
Customer artifactApproved source set and acceptance criteria
3
Install
2–4 weeksWe ship hardware pre-loaded with adapterOS, install it in your environment, and configure the workspace for the selected document job.
Customer artifactLocal adapterOS workspace
4
Validate
1–2 weeksYour team runs the workflow. We verify that answers cite the right materials, outputs are useful, and reviewers receive the context they need.
Customer artifactSourced answers and reviewer notes
5
Operate
OngoingStrong pilots transition to managed deployment. Weak pilots end with a plain recommendation about what did and did not work.
Customer artifactExpansion or stop recommendation
What you provide
- Your deployment environment
- Your workflow and use case
- Your review or compliance path
- Your reviewers and acceptance criteria
- Network and facility constraints
What we provide
- Hardware pre-loaded with adapterOS
- On-site or remote installation
- Specialist workspace configuration
- Source and review-path validation
- Direct support during pilot
What you receive
- A working local document workspace
- Source-backed answers from your materials
- Reviewer notes and control results
- Update and dependency manifest
- Pilot recommendation and next-step plan
Pricing
Commercial engagements begin with a fixed-scope pilot — not a per-token subscription. Price depends on the workflow, facility constraints, support needs, and hardware profile, and it is agreed before deployment.
Hardware is included. No separate procurement cycle just to try the workflow. Local operation also avoids the hidden cost of shipping the same documents to a cloud model again and again.
Tell us early if you need air-gap, offline operation, PHI boundaries, or other compliance constraints so we can confirm fit before proposing scope.
What a pilot leaves behind
A pilot should end with a useful workflow and a reviewer-readable trail, not just a demo transcript. This is the public shape of what we validate with the customer.
Workspace result
- Selected workflow, source set, and reviewer route
- Answers, summaries, comparisons, or draft review notes
- Citations tied to the source materials used
- Control result and local retention plan
Decision packet
- Usefulness findings from real reviewer sessions
- Source quality and failure-mode notes
- Deployment burden, support needs, and update posture
- Recommendation to expand, tune, or stop
FAQ
What runs locally?
The configured workflow, local model path, policy checks, source review, and record generation run on hardware in your environment.
What leaves the environment?
The pilot is designed around no routine data egress for sensitive source material. Support boundaries and any diagnostic export are scoped during briefing.
What evidence is produced?
A successful workflow produces sourced answers, reviewer notes, control results, and a deployment configuration record your reviewers can inspect.
What does pilot success mean?
Success means one real workflow can run inside the agreed boundary, produce useful source-backed work, and fit the customer’s review process.
Do you charge per token or per message?
No. We scope pilots and deployments around the workflow, not a meter that ticks up every time someone asks a follow-up. That makes cost more predictable for regulated teams doing real document review.
What is MLNavigator not selling?
Not a cloud chat subscription, model marketplace, or open-ended consulting program. The first commercial motion is a controlled deployment around one workflow and one review path.
What we are not selling
Not a cloud service or SaaS subscription. adapterOS runs on local hardware with no external dependencies.
Not a consulting engagement. We ship a product with managed support, not billable hours.
Not a model marketplace. You choose the models. We provide the governed workspace and review layer.
Not a broad platform pitch. The current commercial story is a narrow, controlled pilot—one workflow, one environment, one review path.
Request pilot scope
We respond within 2 business days.