Business model and distribution experiment
Pricing
Pricing direction · 1 October 2026: Free — one Admin Agent, provisionally 100 tasks/month; Start — $19/month, full Admin Agent functionality; Pro — $39/month, Admin + Frontend Agent. Metering, paid quotas and top-up pricing remain open pending Agentfy analytics. This supersedes the earlier $24 hypothesis.
Commercial hypotheses; no customers, revenue or published plans. Prices below are proposed tests, not researched willingness to pay.
Charge for a maintained recurring workflow that produces a checked result, rather than for access to a chat box. The owner should know what a job can change, how much work is included and what extra usage would cost.
A testable offer
Start with one shop, assisted setup, supplier mapping, reviewed price jobs and a job history. Test willingness to pay at €79 and €149 per store per month with clearly stated, differing job-volume and support assumptions; do not present an unmeasured “unlimited” plan. Define row/job allowances from real execution costs before offering either tier. A one-off installation/mapping fee may be necessary for customized shops and should be quoted separately.
These are interview and paid-pilot price points. They are not recommendations to publish a pricing page today. Billing currency, taxes, refund terms and regional purchasing power still need decisions. Requiring the merchant to buy a separate model subscription would complicate the promise; provider-inclusive billing is the preferred hypothesis, subject to unit economics.
Economics to calculate before scale
Per-store contribution = collected subscription and setup revenue minus model, infrastructure, payment-processing, onboarding and ongoing support costs. Include failed/retried jobs and human exception handling. Measure support time explicitly: a cheap model call does not make a custom OpenCart integration cheap to maintain.
Value comparison = observed owner hours saved × the merchant's own value of time, plus separately evidenced costs avoided. Do not count hypothetical recovered revenue as measured benefit. Small shops with infrequent updates may fit a pay-per-job offer better than a subscription.
Distribution sequence
- Recruit design partners through founder-accessible OpenCart merchants and agencies; availability is unverified until they agree.
- Produce a demonstration using a consented or synthetic supplier file, showing rejected rows and recovery as well as a successful update.
- Test problem-specific landing pages and a supported-version extension listing. A listing is a channel experiment, not a guarantee of discovery.
- Consider an agency plan only after learning whether repeated installations lower acquisition and support costs. Keep each shop's credentials and approvals isolated.
The OpenCart marketplace exists; we do not need to invent an entirely new discovery channel. We still need to learn which channel reaches merchants who have this problem and authority to pay.
Market size: deliberately not invented
No credible count of reachable, active, suitably configured and willing-to-pay OpenCart stores was established here. Build a bottom-up estimate from qualified agency portfolios and actual conversion: reachable shops × eligibility rate × paid conversion × annual revenue per shop. Report assumptions and a range. Total e-commerce GMV is not our software revenue opportunity.
Continue with the pilot gates.