D3 Morpheus › Platform › Per-Client AI Opt-In
D3 Morpheus
Per-Client AI Opt-In. Run AI Where Policy Allows.
Autonomous for the clients who are ready. Deterministic for the clients who aren’t. One platform, one audit format, switched on per client when their governance catches up. When Morpheus is uncertain, it defers to a human. And the AI is in the price, not on a meter.
Clients running AIClients on Level 1, no AI
What is per-client AI opt-in?
Per-client AI opt-in means the decision to run AI is made per client tenant, not once for the whole platform. Morpheus sets its autonomy mode on each tenant independently, and the lowest mode, Level 1 Deterministic, runs rule-based playbooks with no AI in the chain at all. A service provider can run autonomous investigation for the clients whose governance allows it and deterministic playbooks for the clients whose contracts prohibit AI, on one platform that produces the same audit format for every tenant.
Why does a service provider need this?
Because a client base is never homogeneous, and under an all-or-nothing platform the most conservative client sets the pace for every other one. The usual workarounds are worse than the problem: run two platforms and reconcile two audit formats, or delay the whole deployment until the slowest client's board signs off. Per-client opt-in removes the dependency. Adoption proceeds on the clients who are ready, and the clients who are not keep running on the same platform until they are.
What does a client who hasn't opted into AI actually get?
The full deterministic product. Event intake, correlation, deduplication, rule-based playbooks, ticket creation, notifications, SLA tracking, case management, and the 800+ self-healing integration catalogue, with the same audit trail format every other tenant produces. Level 1 is not a trial tier or a degraded mode. A client can run on it indefinitely and still be a full customer of the platform.
Can a client start without AI and turn it on later?
Yes, and that is the point of running both postures on one engine. Moving a tenant from Level 1 to an AI mode is a configuration change on that tenant. The integrations stay connected, the case history stays intact, the playbooks already built keep running, and the audit trail format does not change, so nothing in the client's existing evidence record has to be re-explained to an auditor. Every mode boundary is independently reversible, per tenant, so a client that advances can also drop back.
What does the audit trail look like across a mixed book?
Identical. The four modes are configurations of the same reasoning engine and the same deterministic playbook engine, not separate products, so the evidence record is one format whether Level 1 or Level 4 handled the incident. A governance team reviewing a mixed client base reads one document structure across every tenant rather than reconciling two architectures.
Client postures at a glance
| Client posture | Mode | What AI does | What the client gets | Audit format |
|---|---|---|---|---|
| Contract or regulator prohibits AI | Level 1 Deterministic | Nothing. No reasoning step, no recommendations. | Full deterministic playbook engine, case management, 800+ self-healing integrations. | One trail per incident |
| Governance allows AI to advise only | Level 2 AI-Assisted | Investigates and recommends. Analyst approves every state-changing action. | L2-depth investigation on every alert before an analyst opens the case. | Same format |
| Governance allows AI to draft response | Level 3 AI-Led | Drafts the response plan. High-risk steps pause for approval. | Analyst reviews plans instead of building them. | Same format |
| Board has signed off on autonomous handling | Level 4 Autonomous | Runs triage to closure within the gates you configure. | Commodity alert handling off the analyst queue. | Same format |
Who this matters most to
Advisory and professional services firms, where a client's confidentiality terms may rule out AI processing of that client's data while the rest of the book has no such restriction. Public sector and regulated operators, where authority to proceed arrives on its own schedule and a deployment should not have to wait for it. Gaming and multi-jurisdiction operators, where one estate can span regulators that disagree with each other and a single platform posture cannot satisfy all of them at once.
The common sentence across all three: adoption does not have to wait for the slowest approval in the portfolio.
Common questions
Can Morpheus run AI for some client tenants and not others?
Yes. Autonomy is set per tenant, and Level 1 Deterministic runs with no AI in the chain at all. A service provider can run autonomous handling for the clients whose governance allows it and rule-based playbooks for the clients whose contracts prohibit AI, on one platform with one audit format across every tenant.
Is the non-AI mode a reduced version of the platform?
No. Level 1 Deterministic is the full deterministic product: playbook engine, case management, correlation, SLA tracking, and the 800+ self-healing integration catalogue, producing the same audit trail format as every other mode. A client can run Level 1 indefinitely.
What happens when a client's governance changes?
The tenant moves to a different autonomy mode by configuration. Integrations, case history, and playbooks stay in place, and the audit trail format does not change. The move is reversible per tenant.
Can different clients run different modes at the same time?
Yes. A provider running many client tenants can sit at Level 4 for some, Level 3 or Level 2 for others, and Level 1 for the clients whose contracts prohibit AI, and produce one audit format a governance team reads identically across all of them.
Does per-client opt-in make us compliant with the EU AI Act or NIS2?
No platform setting makes an organization compliant on its own. What per-client opt-in gives a governance team is the ability to run the posture its own oversight process has approved for each client, and a single evidence record format across every tenant to show for it.
How is the AI priced if only some clients use it?
The AI is included in the subscription rather than metered, so a provider is not billed differently for the tenants running AI modes and the tenants running Level 1. Pricing is sized by alert volume envelope. See the pricing page.
Related
See it run on a mixed client base.
A 30-minute walkthrough: one alert ingress through a Level 1 tenant and a Level 4 tenant, against the same audit format. Bring the client contract that has been holding your rollout up.
Book the demo