Tu Empresa Digital OS — Business Model, Pricing & Builder Network
reference· pm · updated 2026-09-05 · source
Output of the 2026-06-13 strategy session (José + Claude brainstorm). Complements the MVP plan (formerly
apps/admin/src/plans/tedos-pm-mvp-strategy.mdx— MDX plans are retired, ADR-006; seeplan-output-format.md), resolving its §11 (revenue) and §14 (open decision #1: pricing).Status:
reference. The delivery model (SaaS) and the builder-network framework are agreed direction; the pricing numbers and split %s are a PROPOSAL pending sign-off in the dedicated pricing session. Supersedes section §4 (per-seat $29/$89/$299) oftedos-product-decisions.md, which is from the desktop/B2C era and is obsolete for the B2B MVP.
1. Delivery model — closed SaaS
Directional decision: the client does not receive the code; they receive their instance in SaaS mode.
| Open delivery (rejected) | SaaS / closed (chosen) | |
|---|---|---|
| Where it lives | Client's servers | Our servers |
| Code | Client's | Ours, closed |
| Updates | Hard | Centralized, to everyone |
| Reuse of design/philosophy | Drifts per client | A single system |
| Control / IP | Lost | Retained |
The client owns: their instance, their data (exportable), their domain. Not the code — like Salesforce. A differentiator against agency "lock-in": SaaS yes, hostage no — the data is always exportable; you pay for a living system + a partner who guides with proven technology, not for being trapped.
This implies rewriting Locked Decision #4 of the MVP plan (today it says "the client keeps the OS on their own domain" in the delivery sense) to the SaaS sense. Pending.
The three surfaces (role model — superseded, ADR-017)
This role-based split originally proposed living inside a single
apps/admin. The three-surface split that actually shipped isapps/portal(tenant/business) /apps/console(operator platform) /apps/clients/comprender(retiring into portal) — ADR-017, 2026-08-31.apps/adminis frozen. The roles below are still useful as product-intent reference.
| Surface | Who | What |
|---|---|---|
| Operator Portal | Agency (you) | Cockpit: clients, deals, builders, finance, kit |
| Builder Portal | Builder network | Bounties, earnings, training |
| Client OS | End client | The product they operate, on their domain |
Detailed screen map (6 roles, specs per screen): draft ted-os-mapeo-pantallas
(Cowork session 2026-06-13) — pending formalization as a plan/Paper artboards.
2. Pricing (PROPOSAL — pending sign-off)
Currency: MXN. Market anchors: an established MX agency charges $100–250k per project and $10–35k/month for advanced maintenance; US AI consulting charges $5–15k USD/month (premium ceiling).
Implementation · one-time payment (4–6 weeks)
| Tier | What's included | Price |
|---|---|---|
| Basic | 1 flow + panel | $120,000 |
| Standard ⭐ | 1 end-to-end flow + OS on their domain | $150,000 |
| Advanced | Multiple flows / modules | $180,000–250,000 |
Billing: 50% on close + 50% on delivery.
Monthly subscription · SaaS
- $15,000/month flat per company, all-in (hosting · agents · updates · support · reports).
- Annual −15% (~$153k/year). Internal guardrail: Claude API ≤ ~17% of MRR per client.
Dedicated option (large client)
Isolated instance that you control (never the code): monthly from $25,000, implementation from $250,000.
Year 1 example (standard client) + the sales argument
| Item | Amount |
|---|---|
| Implementation | $150,000 |
| 12 × $15,000 | $180,000 |
| Year 1 total | $330,000 |
| Following years | $180,000/year |
Savings vs. alternatives: hiring 2 employees (senior dev + AI specialist, fully loaded) ≈ $2.4M/year; a single AI specialist ≈ $1.17M/year. "For less than the cost of one employee per year, a complete, premium AI system, in 6 weeks — and it's yours."
3. Builder Network — cooperative model (PROPOSAL)
Thesis: José trains curated cohorts (5–8); platform access; capacity without headcount. Inspired by Asian cooperative models + Mondragón:
| Principle | Origin | Application |
|---|---|---|
| Those who do the work keep the majority | Amul (India) | Design + build = the bulk of the payout |
| Split by contribution, not hierarchy | Japan co-ops (patronage) | You're paid for what you contribute to THAT deal |
| Platform takes a coordination cut, not the majority | Amul Federation | Agency charges for origination + QA + brand + kit |
| Capped spread + common fund | Mondragón (~1:5; 25% to the collective) | Healthy gap; a % feeds the kit + training |
| Autonomous micro-enterprises per deal | Haier Rendanheyi | Each client = a micro-team that shares its P&L |
| Bounties per piece | Gitee / open source | Bounty board per issue/feature |
| Earned, non-transferable participation | Huawei (virtual shares) + ESOP | Earned by contributing, switched off if you leave — protects the IP |
Split — Implementation (proposal)
| Role | % | On $150,000 |
|---|---|---|
| Build (builders) | 40% | $60,000 |
| Coordination + QA + relationship (agency) | 25% | $37,500 |
| Design | 15% | $22,500 |
| Origination (whoever brings the client) | 10% | $15,000 |
| Platform + Kit + Common fund | 10% | $15,000 |
Makers (design+build) = 55%, the majority.
Split — Monthly subscription (proposal)
| Role | % | On $15,000 |
|---|---|---|
| Platform + infra + relationship (agency) | 60% | $9,000 |
| Maintenance builder | 30% | $4,500 |
| Common fund + Kit | 10% | $1,500 |
Gamified bounties + ladder
Each piece of work = a bounty (points + pesos). Leveling up unlocks better bounties and a better split.
| Level | How you get there | Unlocks | Split |
|---|---|---|---|
| Apprentice | Training + 1 accepted asset | Small bounties, internal only | Low (recovers the training cost) |
| Builder | X quality bounties | Supervised client work | Standard |
| Senior | Y bounties + reused assets | Large bounties, maintainer | Standard + bonus |
| Master | Strong sustained contribution | Co-defines the kit, dividend | Cap + dividend |
- Reuse royalty: if a builder's asset is reused for other clients, they earn a continuing royalty → rewards feeding the flywheel.
- Loyalty dividend (future phase): a recurring % of the fund, non-transferable, switched off on leaving.
- Training is José's investment and gets recovered: the Apprentice stage pays less (real internal work) + the common fund (10%) finances the next cohorts + non-transferable participation prevents anyone from walking off with the investment.
Split rules
Full transparency per deal · origination always credited · capped spread (~1:5) · the common fund is reinvested (kit + training), not anyone's profit · if José covers multiple roles, he charges for those roles (early on he covers nearly everything and gradually releases slices as he trains builders).
Segmentation (MVP plan §11): whoever operates buys TED OS; whoever builds uses the kit. Never sell the kit to a direct client.
4. Related assets
| Asset | Where | What it is |
|---|---|---|
| MVP plan | (retired MDX, ADR-006 — was tedos-pm-mvp-strategy.mdx) | Master MVP strategy (§11 revenue, §14 pricing) — historical only |
| Decks brief | ted-os-slide-decks-brief.md (repo root) | Partners + Builders decks to execute in Paper |
| Screen map | Cowork session draft | 6 roles, specs per screen — pending formalization |
5. Pending sign-off (José)
- Engagement billing structure (fixed price vs. deposit+milestones) and the final pilot band.
- Final subscription model (flat confirmed) + annual discount.
- Definitive split %s and common-fund / dividend terms.
- Finalize the SaaS-sense revenue model (§2 Pricing above) — the prior MVP-plan MDX this was tracked against is retired (ADR-006); it is not a file-rewrite, it's a pending José sign-off same as the other items here.
- Mark §4 (per-seat pricing) of
tedos-product-decisions.mdas superseded.