Pricing

Both paths start with a conversation.

WorkReef is in private beta. We are taking a small number of design partners now and a larger enterprise cohort once the beta closes. Pricing depends on the size of your AI program, the connectors you need, and the compliance bar on your side. Talk to us.

Private beta · open now

Design Partner

Talk to us

Five to seven companies, sweet spot 500–10,000 employees, Microsoft 365 stack, with an active AI program and a named accountable buyer. Direct line to the founders. Your feedback shapes what we ship.

What you get
  • The full platform: Discover, Understand, Transform, Drive
  • All 28+ connectors, with a custom REST escape hatch for what you built yourself
  • Multi-model quorum (Claude, GPT, Gemini) on every consequential call
  • Tamper-evident audit log, exportable, verifiable on your hardware
  • Bring your own LLM provider: Azure OpenAI, Bedrock, on-prem, or ours
  • Per-action approval gates and per-customer monthly spend cap
  • Persona-scoped UI for all seven personas, including the dept-lead and CISO surfaces
What's different from the GA tier
  • A direct Slack or email line to Bo (founder) and the engineering team
  • Weekly check-ins for the first eight weeks, monthly after
  • Feature requests get prioritized inside the existing roadmap, not added to the end
What we ask back
  • One two-hour scoping call up front
  • Permission to name you publicly once you're comfortable
  • Honest feedback when something is broken
Request access
GA tier · after beta closes

Enterprise

Custom

For organizations with named compliance requirements, multi-region deployment needs, or an active AI program above the design-partner band. Opens once the beta cohort has shipped its first production rollouts.

Adds on top of Design Partner
  • Customer-managed encryption keys with KMS wrapping in production
  • Data residency in a dedicated region
  • SOC 2 Type II report (target: 2027)
  • HIPAA Business Associate Agreement
  • Named technical account manager
  • SLA on incident response
  • Quarterly security review cadence
Get on the GA list
From "request access" to running agents

The sample timeline.

  1. Day 0–2

    Reply.

    We read every request the same day. Bo or someone on the founding team replies within one to two business days, usually with three time slots for a 30-minute intro call.

  2. Week 1

    30-minute intro.

    What you're running, what's stuck, your compliance bar, your timeline. We listen first. If WorkReef is the wrong fit, we say so on the call.

  3. Week 2–3

    Two-hour scoping.

    Your AI Change Leader, your CISO counterpart, and one of us. We map your stack against the connector catalog. You see real screens, not slides.

  4. Week 3–4

    Beta workspace.

    You connect three integrations. The Cartographer maps your org. You see what AI is already running inside your company. Most beta customers find at least one tool they didn't know they were paying for.

  5. Week 5+

    First transformation.

    Approve a takeover candidate. The platform provisions the agent. Shadow runs start. Maria, your AI Supervisor, watches. The promotion gate refuses to advance until thirty runs at eighty-five-percent agreement. Production lands when it lands.

FAQ

Common questions before the call.

Is there a self-serve tier?

Not yet. WorkReef governs decisions that affect people's roles and a company's AI spend. The work to bootstrap a new tenant benefits from a real conversation. We'll open self-serve once the playbook is mature enough that it doesn't.

What's the commitment?

The design partner agreement is a six-month term. We expect honest feedback and the occasional reference call. You can terminate at any point with thirty days' notice, and tenant data export is one click in admin settings.

What does implementation look like?

Connect three integrations. Run the Cartographer pass. Approve the persona assignments in bulk. Review the first wave of transformation proposals. From signup to "the platform is making real recommendations" is typically days, not weeks, assuming your stack is on the catalog.

Can we use our own LLM provider?

Yes. Per-customer inference routing is in the platform from day zero: Azure OpenAI, Amazon Bedrock, on-prem, or a mix. Model allowlist enforced on the server. The rest of the platform doesn't change when you pick a different backend.

What about HIPAA / SOC 2?

Our current security posture is on the Security page: what's done, what's mid-stride, what's deferred. Formal attestations are on the Enterprise tier roadmap. For HIPAA-relevant deployments (Movemedical is customer one) we work the BAA path as part of onboarding.

Do we own our data?

Yes. Each customer's data lives in its own isolated database, separate from every other customer at the storage layer. Bring your own key with Azure Key Vault, AWS KMS, or GCP KMS ships with the Enterprise tier. Export and right-to-be-forgotten are one-click admin flows.

What if we want to leave?

Export drops a zip with every table from your tenant database, credentials redacted. No proprietary file format. No special lock-in. The connectors keep working pointed at whatever you switch to.

What if the agents are wrong?

The drive layer is reversible. Promotion forward requires thirty shadow runs at eighty-five-percent agreement. Backward moves (autonomous to assist to off) are always allowed. Per-action approval gates block sensitive moves until a named approver releases them. The audit log records who approved what, when, and what fired afterward.

Ready to talk?

Tell us about your AI rollout. We'll get back to you with a calendar invite within two business days.

Request access Platform overview