Donely
    Linewize by Qoria
    Proposal · v1 · 29 May 2026

    A white-label AI employee
    for every Linewize IT admin.

    Donely proposes to build, operate, and white-label an AI employee that handles the day-to-day work an IT admin does inside Linewize - so the admin never has to log in to the content filter again. Qoria sells it. Donely runs it.

    Prepared for
    Qoria · Linewize
    Prepared by
    Donely, Inc.
    Harsha Abegunasekara
    Chamaru Amasara
    00 · Executive summary

    School IT admins spend 20–40% of their day inside the Linewize content filter - triaging teacher requests, tuning policies, verifying safeguarding alerts, and chasing bypass attempts.

    Donely will deliver a white-label AI employee that does this work for them. It lives in the messaging tool the school already uses (Teams, Slack, Telegram, WhatsApp), drives the Linewize UI on the admin's behalf, and only pings the admin for a one-tap approval when a human decision is required.

    Phase 1 ships in weeks using front-end automation over the existing Linewize UI - no Qoria engineering required. Phase 2 swaps the browser layer for direct MCP / API calls as Qoria ships them. Same product, same UX, lower latency.

    We are asking for: NDA, a Linewize QA tenant, and 1–2 pilot districts.

    01 · The persona

    The school IT admin spends 20–40% of their day in the filter.

    One person, often one per campus, sitting between teachers, students, parents, principals, and Linewize. Their day is a steady stream of "unblock this for my class," "why is this getting through," and "the test starts in 10 minutes."

    They also own the rest of the school network. Every minute they spend triaging filter requests is a minute not spent on the things only a human IT admin can do.

    North star
    "The IT admin should never have to log in to the content filter again."
    - Hansa Wijayasundara, Qoria · 29 May 2026
    IT admin scenario map
    15 recurring tasks · from Hansa's master sheet
    AUTOMATABLE
    01Daily triageLegit educational site blockedDaily · 5–20m
    02Daily triageTemporary allow / unblock requestDaily · 2–10m
    03Daily triageIdentity / group sync issueDaily · 5–15m
    04Risk & safetyHarmful site not blocked (teacher report)Weekly · 10–45m
    05Risk & safetyStudent bypass attempt (VPN, proxy)Daily · 15–60m
    06Risk & safetySafeguarding red-flag verificationDaily · 10–30m
    07Risk & safetySelf-harm / suicidal-content alertWeekly · escalate
    08Reactive supportParent / principal complaintWeekly · 15–60m
    09Reactive supportTeacher 'test starts in 10 minutes' requestDaily · ad-hoc
    10Policy & controlCohort policy change (e.g. block YouTube 8–2)Weekly · 15–60m
    11Policy & controlTime-bound temporary policy + auto-expiryDaily · 5–20m
    12Policy & controlTrend review: top blocked apps / sitesWeekly · 10–30m
    13Policy & controlAlert tuning / noise reductionMonthly · 30–90m
    14GovernanceHTTPS / certificate issueMonthly · 30–120m
    15GovernanceCompliance / breach reportingQuarterly · 30–180m
    02 · The solution

    An AI employee, not another dashboard.

    The admin keeps working where they already work. The agent does the filter work in the background and surfaces a clean approve/decline only when policy or judgement requires it.

    Lives where the admin works

    Microsoft Teams, Slack, Telegram, WhatsApp. The admin gets a DM, taps a button, the work is done in Linewize.

    Drives the Linewize UI

    Day 1: computer-use over the existing Linewize web app - login, navigate, create policy, set expiry, log out.

    Policy-aware, not just reactive

    If the school has a hard rule (no YouTube during class), the agent auto-declines and tells the teacher why. Admin never touched.

    Closes the loop end-to-end

    Temporary unblock → time-bound policy → reminder before expiry → reply to teacher when done. One thread, zero context-switching.

    Knows the school's history

    Past requests, prior IT responses, classroom calendars, and risk signals - used to suggest the right action, not just any action.

    Built to upgrade to MCP

    When Qoria opens backend APIs, Donely swaps the browser layer for direct calls. Same persona. Lower latency. No re-training the admin.

    03 · How a request flows

    A teacher asks. The admin taps once. Done.

    Worked example - a teacher requests an unblock 10 minutes before a test.

    1. 1
      Teacher posts in Teams

      “Need khanacademy.org unblocked for Yr 9 Maths, 11:30.”

    2. 2
      Donely classifies & checks policy

      Reads the request, checks the school's rules and history, plans the Linewize action.

    3. 3
      1-tap approval to admin

      Telegram DM: “Allow Khan Academy for Yr 9, 11:30–12:30?” [Approve] [Decline]

    4. 4
      Agent drives Linewize UI

      Logs in as the admin, creates a time-bound policy, sets auto-expiry, signs out.

    5. 5
      Closes the loop

      Replies to the teacher in Teams. Notifies admin: done. Reminds before expiry.

    Interactive walkthrough

    See it run end-to-end.

    Three scenarios, click-through, across the real Linewize UI, Microsoft Teams, and Telegram. Built so Hansa, the Head of Engineering, and the CPO can each walk it themselves in under two minutes.

    • YouTube unblock - auto-decline by hard policy
    • Gambling-help site - 1-tap Telegram approval, time-bound
    • VPN bypass - daily triage + remediation
    04 · White-label, Qoria-fronted

    Qoria sells it. Donely runs it. Schools see Linewize.

    Shipped as a managed service under the Linewize brand. Qoria sales enables it per customer in one click - no engineering involvement to onboard a district.

    Customer onboarding flow
    1. 1. Qoria sales hands a school an email-gated link
      Branded Linewize / Smoothwall onboarding URL.
    2. 2. School registers - chooses the IT-admin persona
      “Next → Next → Next”. Enters tenant URL + admin credentials, stored encrypted.
    3. 3. Connects the messaging tool the admin already uses
      Teams, Slack, Telegram, or WhatsApp. One OAuth click.
    4. 4. Agent goes live
      First alert lands in the admin's DM within minutes. No login, no training, no dashboard to learn.
    05 · Phasing

    Ship value now. Re-platform onto APIs when Qoria is ready.

    We are explicit with Qoria leadership: no backend APIs exist today, and we are not asking for any. Phase 1 proves the product on the front end. Phase 2 swaps the integration layer underneath, with zero change to the IT admin's experience.

    Phase 1 · Now
    SHIPS IN WEEKS

    Front-end automation over Linewize

    • Computer-use agent drives the existing Linewize web UI
    • Map top 5 IT-admin scenarios end-to-end with full QA harness
    • Messaging layer: Teams / Slack / Telegram / WhatsApp
    • Per-school knowledge base of past requests & responses
    • Pilot in 1–2 districts; measure minutes saved / admin / day
    Phase 2 · 6–24 mo
    AS QORIA SHIPS APIS

    MCP / direct API integration

    • Donely converts Qoria APIs into an MCP / CLI backend for the agent
    • Same agent, same DM UX - browser layer retired
    • Sub-second policy actions instead of UI roundtrips
    • Login-with-Qoria SSO for onboarding
    • All 15+ IT-admin scenarios in production
    06 · The ask

    Three things from Qoria to move forward.

    Nothing requires Qoria engineering effort. We are asking for access, not work.

    01
    Sign a mutual NDA
    So Hansa can share the IT-admin scenario sheet, sample requests, and a Linewize QA tenant.
    02
    Provision a Linewize QA tenant
    Read/write access for one Donely service account, scoped to a non-production environment.
    03
    Identify 1–2 pilot districts
    Friendly customers we can co-pilot for 6–8 weeks, measuring minutes saved per IT admin per day.
    Decision needed by
    Mid-June 2026

    So we can sign NDA, scope, and have a working pilot in front of a real district before the new school year.

    Reply with approval to proceed
    07 · Team & contact

    Who builds this.

    Donely is a venture-backed team based in San Francisco, building the AI-employee layer for the tools schools and businesses already pay for. We've shipped computer-use agents into production - on real workloads, not demos - and we're bringing that to Linewize.

    San Francisco · VC-backed
    Harsha Abegunasekara
    Harsha Abegunasekara
    Co-founder & CEO, Donely.ai
    [email protected]
    Chamaru Amasara
    Chamaru Amasara
    Co-founder & CTO · Computer-use lead
    [email protected]
    Donely
    ×Linewize by Qoria
    © 2026 Donely.aiConfidential - for Qoria internal review