From first conversation
to live agent.
The path is straightforward, but it has steps — and we ask things of you at each one. Here is what it actually looks like to work with us, written for a working business owner who has limited time and zero patience for jargon.
Six stages.
Roughly six weeks.
Timelines vary by industry and complexity. A simple booking chatbot for a hair salon moves faster than a multilingual client intake agent for a tax office. We give you our best honest estimate after the discovery call.
-
First conversation.
You email us, fill in our intake form, or chat with the try-it-free assistant. We learn what your business does, what problem you are trying to solve, and what languages your customers speak. There is no pitch on this call — we are figuring out whether we can help you, and whether we are the right people for what you need.
-
Discovery call.
Thirty to forty-five minutes on a video call or by phone. We walk through how your business currently handles the problem you want to solve — who answers the phone, what questions come up most, where the gaps are. We tell you honestly whether an AI agent is the right fit, or whether what you actually need is a better phone tree or a part-time hire. About a third of the conversations we have end here, and that is fine.
-
Build.
We draft your agent — its personality, its rules, its language for both English and Spanish, the questions it asks, the things it will not say. You see samples before anything goes live. You can request changes. The build is not finished until you say it is.
-
Trial.
Before the agent talks to your real customers, it talks to you. We run scripted conversations through it — both ordinary scenarios and the ones that go wrong. You see what it says when a customer is angry, when someone asks for a refund, when someone tries to get pricing information it shouldn't give out. The agent does not go live until these conversations pass.
-
Go-live.
The agent starts handling real conversations on whatever channel we agreed on — your website chat, Instagram DMs, WhatsApp, voice calls. From day one, every conversation is monitored by a second AI agent whose only job is to flag things that look unusual, so we can review them quickly. You get a dashboard. We are on call.
-
Ongoing improvement.
Agents get better the longer they run. We watch the actual conversations, find the moments where the agent said the wrong thing or missed an opportunity, and tune it accordingly. Your monthly fee covers this. You do not pay for improvements — they are the product.
Three things from you.
We are doing the work of building, testing, monitoring, and improving your agent. We need three things in return — and when we don't get them, the agent suffers.
-
Time on the discovery call and during build.
An agent is only as good as what we know about your business. We need the real questions your customers ask, the real edge cases, the real reasons people get frustrated with your current setup. Generic answers produce generic agents.
-
Sample review before go-live.
We will send you a transcript of how the agent handles conversations. You read it. You tell us what feels right and what doesn't. This is not optional — it is the difference between an agent you trust and one you don't.
-
Honest feedback once it's live.
When the agent does something embarrassing, tell us right away. When a customer says something you wish the agent had caught, send us the screenshot. The library of cases we build with you is what makes the agent quietly excellent six months in.
The boring parts
we will not skip.
California has specific rules about AI in businesses — about recording disclosure, about data privacy, about automated decision-making. Most are not complicated. Some have real penalties. Here is what we handle, and here is where we tell you to call your own attorney.
Recording disclosure (CIPA).
If your AI agent uses voice and records calls, California law requires that callers be told. We handle the disclosure language and the workflow that ensures it happens before any recording begins. The script is documented and reviewable.
Customer data (CCPA / CPRA).
Customer data your agent collects belongs to you and your customers — not to us. Our written agreement spells out storage, retention, deletion on request, and what happens at the end of our engagement. Specific terms live in the Terms of Service.
Automated decision-making (ADMT).
California is finalizing rules around AI making decisions that affect consumers. Our agents are built to support and route — not to make final decisions on pricing, eligibility, or service refusal. The hard calls stay with you.
Industry-specific rules.
If you are an accountant (IRS 4557), an insurance professional (California Insurance Code), a tax preparer, or in another regulated trade, additional rules apply. We have read them, and we configure your agent accordingly. We do not, however, replace your industry counsel.
What we don't do.
We are not your attorney, accountant, or insurance broker. If your situation involves real legal exposure, a tax question, or coverage decisions — we will tell you, and we will tell you to call the right professional. Pretending otherwise is how small businesses get hurt.
Written, on file, reviewable.
All of the above is in your written agreement. Our agreement is currently being finalized by our attorney — when it is ready, every client gets a copy before they sign, and our pricing page links to it publicly.
We're onboarding our first clients now.
If you run a small business in Southern California and you'd like to be part of the founding cohort, send us a note. We will reply personally. Founding clients keep their rates for the duration of our agreement — that's the promise.
Email us