The process, step by step

1. Pick a mode. Either you're growing a business you already have, or starting from zero with just an idea or a goal number. Growing an existing business is the default path — the agents go where your buyers already are instead of building audience from scratch.

2. State one goal, once. Not a backlog, not a sprint of tickets — one line, like "increase MRR by $10K" or "add 1M views a month." That single goal is what the whole team works toward continuously, which is the part that broke my mental model the most: no re-prompting, no task queue you manage yourself.

3. Pick a speed tier, not a feature tier. This is the detail I didn't expect: the two plans ($50/month "Starter" and $300/month "Turbo") aren't different capability sets. Turbo just runs at 6x the usage — same team, same quality checks, same everything else, just faster throughput toward the goal. I've seen plenty of products split tiers by locking features behind a paywall; splitting purely by throughput is a cleaner story to reason about.

4. A refundable deposit holds your spot. 20% of the first month ($10 or $60) to get on the waitlist, refunded in full if you don't get off it, and only converted to a real charge once you're off the waitlist and have confirmed the goal. Low-stakes way to test a long waitlist without committing capital up front.

5. Connect accounts, then a Boss agent takes over. Once the team has access to wherever the actual work happens, one Boss agent owns the goal, sets its own deadlines in hours, and coordinates the content/growth/sales/support agents under it. You're not assigning individual tasks at any point past this.

6. Your ongoing job shrinks to short, specific moments. A phone code, a payment confirmation, a yes-or-no call — the stuff that genuinely needs a human, nothing more. You can watch a live control room or just message the Boss directly if you want a status check, but neither is required.

Why this is worth noting as a pattern

The thing I'd flag to anyone building or evaluating agent products: the unit of work here isn't "a task" or "an agent," it's "a goal," with a coordinating layer (the Boss) translating that goal into whatever the team needs to do next, scored before it ships. That's a meaningfully different architecture from a chain of single-purpose bots, and it changes what "hiring" even means — you're not provisioning functions, you're assigning ownership of an outcome.

Full step-by-step with links to each piece: here.

The four signals

Consistency matters because AI answers are built by cross-referencing multiple sources. A business with three different descriptions across its own site and its listings is a harder recommendation to make than one with a single clear story told the same way everywhere.

Third-party mentions matter more than most teams expect, because the AI isn't just reading your site — it's weighing what other people say about you. Being genuinely useful in the threads and comparison content your prospective customers already read is one of the more reliable ways to end up in the sources an AI pulls from.

Structured data removes ambiguity. Instead of guessing your category, hours, or FAQs from unstructured paragraphs, the model gets clean, labeled data to work with.

Useful public content beats marketing copy because vague language doesn't quote well, and a direct, plainly-written answer to a real question does.

Why this is harder than it sounds

None of this is a setup-once-and-forget situation. New threads start, old mentions age out of relevance, and your own content needs to stay current against whatever people are actually asking this month, not last year. That's genuinely hard to keep up with by hand across every platform where your business gets mentioned — and it's a different kind of work than running a one-off SEO audit, because there's no dashboard that tells you whether you're winning. The only way to check is to actually ask the model the question yourself, repeatedly.

If you want the fuller version of the four-signal breakdown and the full FAQ: here's the guide.

Would be curious whether anyone's actually automating the "watch mentions and keep the FAQ current" part yet, or if everyone's still doing it by hand once a quarter and hoping.

The three models

Custom-built agent. A dev shop builds you one agent from scratch for a specific workflow. Typical range: $25,000-$300,000+ one-time, plus ongoing maintenance after that. Scope and quality depend entirely on the build, which means the quote is really a proxy for "how much custom engineering are we signing up for," not a fixed price for "an AI agent."

Per-seat sales agent. One agent, usually scoped as a point tool rather than a full team, billed per seat: roughly $50-$500/seat/month depending on the vendor and feature set.

Per-resolution support agent. Billed by usage instead of seats — about $0.50-$2.00 per resolution for most self-serve tiers, with enterprise tiers moving to quote-based pricing that can start around $50,000/year.

The common thread across all three: before you've gotten any value out of it, you're either staring down a five-figure-plus build quote, or you're on a subscription/usage meter that scales with every seat, action, or resolution you add. That's a reasonable tradeoff for a narrowly-scoped point tool. It's a much harder one to swallow if what you actually want is something that behaves more like a team working a goal than a single function call.

Where flat pricing changes the math

I ended up looking closely at one product that doesn't fit any of those three buckets, because it's not pricing a single agent at all — it's pricing a full team (a "Boss" plus the agents under it) working one goal, at a flat monthly rate regardless of seats or resolutions: $50/month on a steady "Starter" tier, or $300/month on a 6x-usage "Turbo" tier. Usage speed is the only variable between the two plans; the team, the features, and the quality gate on everything that ships are identical on both.

What made me actually trust the number was the deposit structure: you put down 20% of your first month ($10 on Starter, $60 on Turbo) to hold your spot on a waitlist, and it's refunded in full, immediately, if you don't get off it. You're only billed for the rest of month one after you're off the waitlist and have confirmed your goal. That's a meaningfully different risk profile than a $25K build quote or a per-resolution meter you can't fully predict in advance.

I'm not saying flat pricing is always the right model — a narrowly-scoped point tool billed per seat makes sense if that's genuinely all you need. But if you're comparing quotes and one of them is "$50 or $300 flat, refundable deposit to start," it's worth checking what that actually includes against what the five-figure build or the per-seat tool gets you for the same money.

Full pricing breakdown and the comparison table I used: here.

Here's the distinction that actually held up once I dug into it: SEO optimizes for a ranking algorithm that returns a list of links. AEO optimizes for an AI assembling one synthesized answer from many sources at once. Those are different outputs, judged by different mechanisms, and winning one doesn't guarantee the other.

Why it's a different problem, not a rebrand

With classical SEO, you control your own page and you're playing to a known set of levers — backlinks, keywords, page structure, site authority. You want to be in the top handful of results because people scan and click.

With AEO, the thing doing the judging isn't ranking your page against ten others — it's pulling together an answer from your site, sure, but also forums, review aggregators, comparison articles, and anywhere else your business gets mentioned. A few things fall out of that:

• Ranking #1 in search and getting named in an AI answer are genuinely different wins. You can have one without the other, in either direction.
• Third-party mentions start mattering more than on-page optimization. An AI answer is pulling from what other sources say about you, not just what your own site says.
• Clean, consistent, factual information beats keyword density, because an AI is more likely to cite something it can confidently parse and summarize.
• There's no dashboard for AEO the way there is for search rankings. You can check your SEO position directly; checking whether ChatGPT or an AI Overview names you means actually asking it the question yourself, more than once.

Neither is pay-to-play. Both come down to actually being the clearest, most trustworthy answer to the question someone's asking — SEO just answers it with a ranked list, AEO answers it directly.

Why I'm writing about this now

A growing share of product and service research doesn't start with a search box anymore — it starts with "what's a good CRM for a 5-person agency" typed straight into a chat interface. If your business isn't one of the one or two names that comes back, you don't get a chance to compete for that click. You're just not in the conversation.

Most teams I've talked to still only have an SEO motion, and treat "ranks well" and "gets named by AI" as the same KPI. They're correlated, not identical, and the timelines diverge too: a backlink or title-tag fix can move a search ranking in weeks; an AI's answer updates whenever its sources get re-crawled, which can be faster for a single fact but slower for a reputation that depends on a lot of accumulated third-party mentions.

If you want the fuller breakdown — including a side-by-side table of what each one actually optimizes for — the full guide is here. Curious whether other teams are tracking AEO as a separate line item yet, or still lumping it into "SEO" on the roadmap — let me know what you're seeing.