Straight answers, including to the awkward ones.
Money, risk, who does the work, and what happens if it goes badly. If a question you have isn't here, ask it — we'd rather deal with it now than have it kill the project later.
About working with us.
Money, risk, who does the work, and what happens if it goes badly.
Why should we trust you with this?
Don't, yet — make us prove it, and notice that we've built the engagement so you can. The audit is free and you owe nothing for it. The written plan is yours to keep whether you hire us or not. Our pricing is published on this site instead of negotiated behind a discovery call. You own the code, the repo, and the domain from day one, so you can fire us on a Tuesday and keep everything, working. And if the audit doesn't turn up enough value to justify the work, we'll say so rather than sell you something anyway. Every one of those is a piece of leverage most firms keep for themselves. We gave them up on purpose, because we'd rather be judged on the work than on a pitch.
What does the free AI audit actually include? What's the catch?
You get a written document: where AI would save you time or make you money, ranked by value, with a dollar figure attached to each one and a recommendation of what to build first. It also includes something almost nobody checks — whether AI crawlers like GPTBot and ClaudeBot can even read your website, because for most modern sites the answer is no. The catch is that there isn't one, and the reason is selfish: our audits are good enough that a meaningful share of people who read one hire us. That's a better use of our sales time than cold-calling you.
How much does this cost?
Every service on this site has a published starting price, which is more than most agencies will tell you before a discovery call. Most first projects land between $3,000 and $12,000 depending on scope. You'll have a fixed price and a fixed date before you commit to anything — we don't do open-ended hourly billing, because it punishes you for our inefficiency.
How long until we see something working?
First working version in about two weeks for most builds. Not a slide deck, not a wireframe — a thing that runs, that you can click, that your team can try. We start small deliberately: prove one piece works, then expand what's working. A six-month project with nothing to show for the first four is how agencies hide.
Who actually does the work?
The engineer building your project — that's who you talk to. No account managers, no offshore hand-off, no junior learning on your budget while a senior name sits on the invoice. You email the person writing the code and you get an answer from the person writing the code. We take a deliberately limited number of projects at a time to keep it that way, which is also why we'll tell you early if we're not the right fit for yours.
Will the AI say something wrong to my customer?
That's the correct thing to worry about, and it's the difference between an AI build that works and one that becomes a liability. Everything we ship is grounded — it can only answer from your documents, it cites its source, and when it doesn't know it says so and emails the question straight to you instead of inventing something. We try to break it before you launch and we show you what we found.
Is our data used to train someone's AI model?
No. Your documents are used to answer your customers' questions and nothing else. We use enterprise API endpoints with training disabled, and we'll put it in writing in the contract.
We already have a marketing agency. Does this replace them?
No — it's a different job, and the two sit alongside each other without stepping on anything. Agencies do campaigns and creative; we build systems that keep running when nobody is tending them. If you want a quick read on whether your agency is thinking about any of this, ask them one question: what does GPTBot receive when it requests our homepage? Their answer will tell you a lot.
What if we don't like the work?
You own everything from day one — the code, the repo, the domain, the hosting, all in your name. If we're not working out, you take the work and go, and you're not trapped in a platform you can't export from. We've deliberately made it easy to fire us, because the alternative — customers who stay because they can't leave — is a business we don't want to run.
Do you work with businesses outside California?
Yes. We work with businesses across the United States, remotely. Almost none of this work requires anyone to be in the room, and we'd rather put the budget into the build than into airfare. If you'd genuinely rather meet in person, ask and we'll sort it out.
The specific ones.
Every service has its own set of hard questions. Here they all are in one place.
AI Support Assistant
An assistant trained on your manuals and FAQs that answers customers 24/7 — in your voice, not a generic bot's.
Full details →What stops it from making something up?
It's retrieval-grounded: it can only answer from documents you gave it, and it cites the source on every answer so you can check it. If the answer isn't in your material, it says it doesn't know and emails the question straight to you rather than guessing. We test this adversarially before launch — we actively try to make it hallucinate, and we show you the results.
What if it gives a customer the wrong answer anyway?
You'll see it, because every conversation is logged and reviewable. You can correct a bad answer at the source — fix the document, and the assistant is fixed. We also let you hard-pin answers for anything legally sensitive, like warranty or safety, so those are never paraphrased.
Will it replace my support staff?
Not the goal, and we won't pitch it that way. This doesn't replace anyone — it answers the questions that were interrupting them. The repetitive tier-one questions get absorbed so your people can handle the ones that actually need a person, which is the work you hired them for in the first place.
Our documentation is a mess. Is that a dealbreaker?
No — it's the normal starting condition, and cleaning it up is part of the work. We'll tell you honestly during the audit if the gaps are big enough that we'd need to write documentation first.
AI Product & Build Advisor
A guided advisor that asks a few questions and recommends the exact right product — then explains why, and handles the objection.
Full details →How is this different from the filters we already have?
Filters make the customer do the work — they assume you already know what you need. An advisor does the work for them: it asks about the outcome they want, not the spec they think they want, and it translates. That difference is the entire conversion gain.
What if it recommends the wrong product and we eat a return?
The advisor recommends from rules you approve, and we build in the same guardrails your best rep uses. When it isn't confident, it says so and emails the customer's question straight to you instead of forcing a bad recommendation. And it asks the fitment questions your product page never asks — which is the mechanism that stops a wrong-part order before it becomes a return.
Does this work if we don't sell online?
Yes. If you sell through dealers or by quote, the advisor becomes a qualified-lead engine — it produces the exact configuration the customer needs and routes it to you or the nearest dealer as a warm lead.
AI Search Visibility & Schema
Be the brand AI assistants name when someone asks for what you sell. Most of your competitors haven't even checked.
Full details →Is 'getting recommended by ChatGPT' a real thing or marketing hype?
It's real, and it's measurable — you can ask the assistants what they recommend in your category today and see whether you're named. What nobody can honestly promise is a guaranteed placement: these are recommendation systems, not ad slots, and anyone guaranteeing you a spot is lying. What we can do is remove every technical reason you're excluded, which most companies have never done.
We already pay an SEO agency. Is this the same thing?
Ask them one question: what does GPTBot receive when it requests our homepage? Most SEO retainers are still optimizing for a ten-blue-links world and have never tested this. It's a fair question and a good agency will have a good answer.
How do we know it worked?
We take a baseline before we start — the exact prompts a buyer would use, and who gets named — and we re-run it at 90 days. You see the before and after. If nothing moved, you'll know, because we'll show you.
Reviews & Reputation Engine
Automated review requests that actually get answered, plus monitoring that catches an angry customer while you can still fix it.
Full details →Can you filter out the bad reviews before they get posted?
No — and you should walk away from anyone who offers to. Review gating is against Google's policy and, since 2024, the FTC's rule on fake reviews carries real civil penalties. What we do is legitimate and works better anyway: ask everyone, catch problems early, and fix the thing that caused the complaint.
Will you write reviews for us?
Never. Fabricated reviews are illegal, they're detectable, and getting caught is far more expensive than a mediocre rating. Every review we help you get comes from a real customer who had a real experience.
How many more reviews should we realistically expect?
It depends entirely on your volume and how good the experience actually is, so we won't quote you a multiple we can't stand behind. What we can say is that the gap between businesses that ask systematically and businesses that don't is enormous — because most businesses never ask at all.
Content & Video Engine
Turn the expertise you already have — in your head, your emails, your manuals — into content that ranks, teaches, and sells.
Full details →Isn't AI content penalized by Google?
Unhelpful content is penalized, regardless of who or what wrote it. Google's own guidance is explicit that AI assistance is fine — low-value, unoriginal content is not. Since everything we produce is grounded in your real expertise and reviewed by you before it goes out, it's the opposite of the thin content the penalties target.
Will it sound like a robot wrote it?
Not if it's built right, and you're the check on that — nothing publishes without your approval. If a draft doesn't sound like you, it doesn't go out. That's the whole reason the workflow has you in it.
How much of my time does this take?
Review and approval, realistically an hour or two a week once it's running. We designed the pipeline around the assumption that your time is the scarce resource — because it is.
Modern Websites
Rebuilt for speed, accessibility, and conversion — and readable by the AI systems your customers now ask for recommendations.
Full details →We just paid for a website two years ago. Why would we redo it?
Maybe you shouldn't — and if the audit says the honest answer is 'fix three things, don't rebuild,' that's what we'll tell you. But run one test first: view your homepage's source code with JavaScript disabled. If the body is empty, search engines and AI assistants are receiving nothing, no matter how good it looks to you.
Will we be able to update it ourselves?
Yes. We wire up whatever editing workflow fits how you actually work, and we hand over documentation. You should never need to pay us to change a phone number.
What happens if we stop working with you?
You keep everything — the code, the repository, the domain, the hosting account, all in your name from day one. We don't hold anything hostage. It's a bad way to run a business and a worse way to sleep at night.
Question not here? Just ask.
You'll get a real answer from the person who'd do the work — not a form response, and not a calendar link.
Written plan in 3 business days · No sales call · No obligation