# Scirium: Vision, Positioning, and Roadmap

Status: PLANNED (design-only, no code yet)
Brand: Scirium (formerly codenamed Wall-O, retired 2026-08-16)
Domain: scirium.com
First client: Wall Orthodontics (pilot)

This document is the single source of truth for the Scirium product. Every team (design, marketing, documentation, engineering) reads from and references this file. If a team changes how the product is described, this file changes first and the change propagates out.

## 1. One-liner

Ask your company anything. Get a cited answer from your own documents, scoped to the exact department you are talking to.

## 2. What Scirium Is

Scirium is a self-hosted, white-label knowledge assistant for internal staff. A company points Scirium at its own document libraries (SharePoint, OneDrive, employee handbooks, policies, procedures). Staff ask questions in a chat window. Scirium answers using only that company's documents, and every answer carries inline citations pointing to the source.

The name comes from the Latin "scire" (to know) plus the -ium element suffix. The element of knowing.

## 3. The Vision

Every company already has its knowledge. It lives in the handbook, the policy PDF, the billing procedure, the onboarding checklist. What the company does not have is a way for staff to get a correct, cited answer in ten seconds without interrupting a human.

Scirium is that layer. It sits on top of the documents a company already owns, scoped so that each channel answers one domain, grounded so it never guesses, and self-hosted so the data never leaves the company. The vision is not "another chatbot." It is the trustworthy layer between staff and the institutional knowledge they already have but cannot reach.

## 4. Why This Beats a Regular Chat

A regular chatbot (ChatGPT, Claude, Copilot in a browser) is one undifferentiated model with no grounding in your documents, no audit trail, and no scoping. Scirium is the opposite on every axis that matters to a business.

| Axis | Regular chat | Scirium |
|---|---|---|
| Grounding | May hallucinate; no source | Every answer cites its source documents inline with a Sources footer |
| Scoping | One model answers everything | One channel = one knowledge domain + one agent + one scoped source. Billing answers billing. HR answers HR. Never cross-contaminated |
| Data residency | Sent to a vendor's cloud | Self-hosted, white-label, data stays in the company's M365 and your infra |
| Audit trail | None | Every question and answer logged with citations, model, tokens, latency |
| Identity | Anonymous | Each domain is a named agent persona (e.g. "Scirium Billing") |
| Failure mode | Confident wrong answers | If the answer is not in the scope, it says so instead of guessing |
| Tenancy | N/A | One deployment serves many isolated tenants (one Rocket.Chat workspace per tenant, row-level security, vector namespace isolation) |

The core invariant, stated plainly: a channel has exactly one agent and exactly one knowledge scope. If a company needs a second domain, it creates a second channel with a second agent and a second scope. This is what makes Scirium trustworthy where a general chatbot is not.

## 5. Concrete Examples of What It Does

- HR: "How do I submit a PTO request?" → answer with a citation to the PTO policy section of the employee handbook.
- Billing: "What is our refund policy on a disputed charge?" → answer citing the billing policy PDF, paragraph and page.
- IT Help: "What is the WiFi password for the conference room?" → answer from the internal IT runbook.
- Onboarding: "Walk me through the new hire checklist." → step-by-step answer citing the onboarding checklist document.
- Policy: "Can I expense a client lunch over $100?" → answer citing the expense policy, with the exact threshold.

The pattern in every case: the answer is correct, it is scoped to the right domain, and the user can click through to the source document to verify it themselves. That last part is the product. Trust is the feature.

## 6. Roadmap: Where It Is Now vs v3/v4

| Version | Capability | Autonomy |
|---|---|---|
| v1 (implementation target) | Grounded, cited Q&A only. Read-only over scoped M365 document libraries. No actions taken. | Zero writes. Answer-only. |
| v2 | Risk-tiered actions. Reporting first, then low-risk writes that follow propose → approve → narrow audited write. | Conditional, human-approved |
| v3 | Broader autonomous actions by risk tier, deeper connectors, granular per-channel permissions. | Tiered autonomy |
| v4 | Multi-connector breadth, analytics, admin self-service at scale. | Mature, governed |

The gap between v1 and v3/v4 is autonomy, not accuracy. v1 is already fully grounded and cited; what the later versions add is the ability to act, and only ever within an approved risk tier. Scirium never becomes a free-roaming agent. The propose-approve-audit gate is permanent.

## 7. Current State

Design-only. No code, no orchestrator, no containers. The business proposal and critical review are deployed (under the legacy /wall-o/ URL, pending redeploy to /scirium/). The pilot is Wall Orthodontics, an orthodontic practice using Microsoft 365, for internal staff knowledge only.

Hard scope, non-negotiable: internal staff documentation. No patient records, no PHI, no HIPAA scope, no medical or vision AI. X-ray and clinical image review is a separate, FDA-regulated product and is permanently outside Scirium.

## 8. Positioning Keywords (for design and marketing)

- Grounded, cited, scoped, trustworthy, self-hosted, white-label, multi-tenant, enterprise-grade.
- The product is a technology product. The visual language should say "cutting edge, industry-leading corporate, these people are good."
- Tone: confident, precise, understated authority. No hype filler. The trust is the message.

## 9. Brand and Domain

- Name: Scirium (sigh-ree-um). Latin "scire" = to know.
- Primary domain: scirium.com.
- Defensive TLDs available and recommended: .io, .ai, .co, .app.
- Cloudflare at-cost pricing: .com $10.44/yr, .io $50/yr, .ai $70/yr (min 2-year term), .co $15 first/$30 renewal, .app $14.20/yr.
- Footer branding: "A product of IT Pro Partner."

## 10. Audiences

- Technical: the business proposal, the product model, the deployment doc, this roadmap.
- Business-savvy (advisory team, non-technical): sections 1-5 and 8 above. They need to understand the difference from a regular chatbot, see real examples, and understand the v1-to-v4 autonomy curve without implementation detail.