in equivalent mid-level marketing-ops FTE + half an estimator
Annual replacement at parity if you bought the equivalent tools and services separately. Custom-building this one feature alone would land in our Strategic build band ($400k–$900k). Source ladder + per-line citations on methodology .
Concretely, what you get
- Ads: Strategy / Copy / Creative / Audit / Optimizer.
- Content: Topic / Research / Writer / Charts / Editor / Originality / Source Verifier.
- Brand + Print: Brand / Designer Review / Print Network.
- Operations: Storm / Crew Day Pack / Pricing / Reviews / Referrals.
- Engineering: Code Vulnerability / Code Fix / Script Writer / Reviewer.
- Sentiment + Migration: Sentiment / Migration Intake / QA / Cutover.
- Per-client supervision modes: Autonomous, Supervised, Off.
- Monthly AI budget cap per client with 80% / 100% notices.
The shelf-of-tools this one offering removes
Categories, not brand names — pricing benchmarks observed from public pricing pages, agency proposals, and freelance rates current to the year. Every range is backed by a line-by-line worksheet you can audit once your trial is live.
| What you'd otherwise buy | Type | Typical price |
|---|---|---|
| A 1-FTE roofing marketing coordinator | Labor | $48k–$78k/yr fully loaded |
| A 1-FTE estimator helper | Labor | $45k–$65k/yr |
| A 1-FTE office admin | Labor | $38k–$58k/yr |
| Agency markup (typical roofing media buy) | Agency | 10–20% of ad spend |
What you'd pay elsewhere vs. what this costs you here
Vendor ranges observed from public pricing pages and agency proposals, current to the year. Full source worksheet shared with you once your trial is live.
What it adds to your business in dollars
One person with Homericly does what used to take a back office of five (cited on /why/10x-productivity). The agents aren’t a chatbot; they’re a fleet, each with a single responsibility, versioned prompts, test harnesses, and a human-supervision toggle.
Three or four moves. Then you walk away.
You: Pick the supervision mode per agent.
System: New accounts default Supervised; graduate to Autonomous after 30 clean days.
You: Set your monthly AI budget cap.
System: You get 80% + 100% notices; system-initiated storm response keeps running over cap.
The outcomes this feature feeds into
Before you ask
What if an agent makes a mistake?
Every agent has a failure log + retry-with-backoff path + escalation to staff. Code-related agents never auto-merge — every fix goes through staff Engineer review before merge.
Can I turn agents off?
Yes — per-client, per-agent: Autonomous, Supervised, or Off. Default for new clients is Supervised for 30 days, then graduate to Autonomous on a clean track record.
How are agents kept up to date?
Versioned prompts in /agents/{name}/v{n}.md. Test harnesses with 10+ golden examples per agent. When a prompt or model is updated, the test suite re-runs before deployment.
Who watches the AI budget?
Both sides. You see your monthly cap; warnings fire at 80% and 100%. We also have a platform-wide aggregate cap; if it’s hit, autonomous client-initiated work pauses while system-initiated work (storm response) keeps running.
What about prompt injection or data leakage between clients?
Every AI call carries the client’s organization_id in metadata. No prompt or response can contain another client’s data — embeddings are namespaced, queries are scoped, audit logs catch any escape attempts.
Tour the agent fleet that runs quoting, chasing, scheduling, and paperwork.
Then take your hours back.
No credit card. No strings. We stand the whole platform up on a subdomain alongside your existing site so you can compare the numbers directly — leads, bookings, ticket size. The Standard guarantee covers cancellation between day 90 and day 180 at full refund.