The AI employee a person stands behind.

Supervision is the product.

Gartner forecasts that 40% of agentic-AI projects will be scrapped by 2027 — not because models fail, but because organisations can't govern them. Meanwhile a senior executive assistant costs $120–162K CAD loaded and there aren't enough of them. Donna closes both gaps at once: an AI agent that drafts at machine speed, inside an operating model where a trained person reads everything before it ships under a customer's name.

We are not building a better model. We are productising the thing Fibernetics has actually done for fifteen years: run business-critical operations that other companies bet their day on, with a person accountable when it matters. Two engines drive the business:

• The Desk (revenue now): supervised executive seats at $1,500 CAD/month — Donna drafts, a NEWT-staffed reviewer signs, the customer approves only where they set the line. Live in beta today with a published price and a 5-seat capacity cap.
• The Platform (scale): the reviewer dashboard itself, sold as software to the world's ~2M executive assistants — each EA becomes the accountable human for their own fleet of agents and carries 3× the executives. Free for EAs during beta; converts to a SaaS seat.

The billion-dollar shape (illustrative): ~$100M ARR at a software multiple — roughly 4,000 supervised seats plus 25,000 EA platform seats — reached by expanding from the EA agent into a roster (Finance, Support, Voice) on the same supervised-desk substrate, distributed through Fibernetics' existing carrier and VAR channels.

NEWT's first AI employee.

Mission: make agentic AI usable by real businesses by attaching a named, accountable human to every outcome.

What it is: Donna is an AI executive assistant personified on Dawn Gamble — a real senior EA of fifteen years at Fibernetics — and operated under her review. She runs calendars, inboxes, travel, meeting prep and follow-ups. Nothing she produces reaches another human until a person has read it. Donna ships under NEWT, the business division of Fibernetics, a Canadian carrier running business voice and network since 2011.

Stage: live early beta. Two marketing sites are deployed with a working application pipeline; pricing is published; the first five supervised seats are open; the reviewer dashboard is spec'd from the operating desk Dawn already runs.

An assistant they can't hire. An agent they can't trust.

The insight competitors keep missing: supervision is not overhead to eliminate — it's the feature customers are actually buying. The EA-facing version inverts the same product: the customer's own assistant becomes the reviewer, and Donna makes her the most capable person in the building instead of the most threatened.

One agent, one desk, three authority tiers.

The customer draws the line once on an onboarding call; it never moves without them.

The Desk dashboard — the real product

What Dawn uses today becomes what we sell tomorrow: a single pane of glass where one reviewer runs many principals. Per-person buckets (email, calendar, tasks, notes, documents), a review queue ordered by risk tier, an insights page that surfaces patterns (missed sales calls, recurring requests) as proactive suggestions, and one-tap escalation to the principal. Channels meet people where they are: email and calendar natively, iMessage/SMS domestically, WhatsApp for international scale, Teams and call transcription next.

Architecture

• Claude (Anthropic) + Mosaic — frontier reasoning plus instant deploy of servers, sites and apps; the same stack shipped both beta sites and the live application pipeline in days, not months.
• Model failover — automatic fallback to a second provider or a local model, because an assistant that goes down with its vendor isn't an assistant.
• Data residency — customer content stays in the Fibernetics environment in Canada; scoped to connected mailboxes; not used to train general models; exported and removed on exit.
• No new portal for customers — Donna works inside the email, calendar, phone and finance tools the customer already runs. The dashboard is for the reviewer.

A labour market priced for scarcity, meeting a software market priced for scale.

All figures below are illustrative planning assumptions from published 2026 ranges, stated so they can be checked — not performance guarantees.

• TAM — the work itself: executive/administrative support labour in North America alone runs on the order of $150B/yr (≈2M roles × ~$75K average loaded cost). Donna monetises a slice of that work at a fraction of its labour price.
• SAM — supervised seats + EA software: SMB and mid-market executives without an EA (the NEWT customer profile) plus the ~2M EA population as software seats. Illustratively: 500K reachable executive seats × $18K/yr + 2M EA seats × ~$1.2K/yr ≈ a $11–12B serviceable pool.
• SOM — the plan's horizon: ~4,000 supervised seats and ~25,000 platform seats ≈ $100M ARR — under 1% of SAM.
• Why now: three-quarters of enterprises say they've adopted agentic AI; almost none have anything genuinely live. The distance between the two is supervision — a job that has to be staffed, trained and productised, which is an operations problem before it is a software problem. Carriers do operations.

Who buys, who uses, who spreads it.

Everyone else sells the model. We sell the signature.

The moat, honestly

• The reviewer network + correction loop. Corrections don't go into a ticket queue — the reviewer is the designer, so a correction changes the product that week. Every supervised seat compounds a proprietary dataset of "what good judgment looks like" that model vendors don't have.
• The EA flywheel. Every other vendor threatens EAs; we arm them. The free assistant seat recruits the exact population that (a) buys software seats, (b) staffs our desk as we scale, and (c) sits between every vendor and every executive inbox as gatekeeper.
• Carrier trust + channel. Fifteen years of business-critical uptime, data in Canada, one accountable throat — plus an installed base and VAR network no agent startup can rent.
• Switching cost. Connected mailboxes, calendars, tone models and authority tables make seat N+1 cheaper for us and leaving costlier for them.

Concentric circles, each one funding the next.

• Circle 0 — the 5-seat beta (now). Local, friendly businesses within ~100 km for in-person support; warm leads first. Free during beta, published $1,500 price, capacity cap stated plainly. The intake pipeline is live and tags applicants by seat.
• Circle 1 — VARs as customers (Q1'27). Resellers of NEWT phone systems get seats before they sell them — the demo is their own desk. Their book of SMBs becomes the first paid cohort.
• Circle 2 — the EA community (rolling). donna-for-eas is the acquisition engine: free assistant seats, a written "pitch your exec" note in the product, and the only vendor message in the category that EAs can forward without fearing for their job. EAs pull Donna into companies; execs don't have to be sold.
• Circle 3 — carrier bundle + international. Donna joins the NEWT bundle (voice + internet + receptionist + AI employee, one bill, one throat). WhatsApp unlocks non-NA markets where iMessage/SMS numbers fragment.

Two engines, one substrate.

Unit economics — supervised seat (illustrative)

The strategic point: the supervised seat is priced at roughly a seventh of a loaded EA and below a part-time offshore VA — while the platform seat costs less than an EA's monthly parking. Neither engine asks the customer to believe anything about AI; both ask them to believe a person they've met.

The desk is the factory.

• Review SLA: outbound checked in under 30 business-minutes (2 hrs off-hours); when we'll miss it, Donna hands the customer the finished draft to send themselves — the customer is never blocked on us.
• Desk scaling: Dawn is reviewer #1 and the training standard. Each new reviewer is an experienced EA, trained on the authority-tier playbook, shadowing live seats before holding their own. Target ratio: 8–10 supervised seats per reviewer at maturity; customers are told by name when anyone new touches their account.
• Onboarding: Day 0 thirty-minute call and authority table; days 1–3 shadow mode (Donna reads, touches nothing); days 4–10 Tier 1 live; days 11–30 the customer moves the line.
• Reliability: operated on carrier discipline — monitored, multi-model failover, incident paths that already exist for voice and network customers.
• Security roadmap: data-processing terms before anything connects; scoped mailbox access, revoked on disconnect; SOC 2 Type I in the funded plan. Until certified, we say so plainly — the honesty is load-bearing.

The person, the builder, the carrier.

From five seats to $100M ARR — the arithmetic, not the poetry.

Every figure is an illustrative planning assumption for decision-making, not a forecast.

• Cost structure: reviewers (the dominant variable cost, ~50% of supervised-seat revenue at maturity), AI compute (~5–10%), platform engineering (fixed, small — the stack is deliberately lean), sales via existing channel (low CAC by design).
• Break-even (illustrative): ~35–40 supervised seats carries the early team; the platform seat's 85% margin pulls blended margin up as mix shifts toward software — which is the whole strategic design.
• Funding need: $1.5M CAD for 18 months (internal venture allocation or seed): Desk dashboard build, 6 reviewers hired and trained, SOC 2 Type I, WhatsApp + voice channels, VAR partner program, and marketing that stays in the honesty register.
• Use of funds: ~45% engineering, ~30% desk staffing & training, ~15% compliance/security, ~10% GTM.

Three horizons.

What kills this, and what we've done about each.

Green-light the venture. The beta is already live.

• Decision: constitute Donna as a NEWT product line with $1.5M CAD / 18-month allocation and Dawn formally allocated to it.
• Next 30 days: fill the 5 beta seats from live intake (first target: the warm Delta Air lead); start the Desk dashboard build; begin reviewer #2 recruitment from the EA community funnel; kick off SOC 2 scoping.
• Success gate for the next tranche: 5 seats live with retention through day 30, review SLA held, and a signed VAR pilot — evidence, not projections.

Market sizes, financial projections, margins and seat counts in this document are illustrative planning assumptions drawn from published 2026 ranges and internal estimates; they are not forecasts or performance guarantees. Product claims describe the live beta as deployed. Prepared with the founders' knowledge for internal decision-making.