The 2AM problem.

You’re building a world where nobody has to call. Neither site says so.

The bet is a good one, and it is the same bet Breesy Voice already makes from the contractor’s side. Someone describes the damage once, the routing happens without a call center in the middle, and nobody waits on hold at 2am. That is what makes this different from every other restoration lead form.

We audited both properties end to end. Connect came back with 33 findings, breesy.ai with 65 across 47 pages. Different problems, one shared cause: both sites are better than they look and neither one explains itself.

Two companies. One of them is invisible.

“breesy.ai” appears zero times on this site. “breesy-connect” appears zero times on breesy.ai. No parent company in the code, no shared branding, domain four months old behind private registration. There is no path, for a person, for Google, or for ChatGPT, from this site to the real company that owns it.

You’re already fixing this from the other end.

Restoration’s weak point is the phone call. More than half of emergency calls come after hours, even strong firms answer only about 87% of them, and industry figures put each missed call at $2,500–$10,000, because the homeowner simply dials the next company while it rings out.

That is one strategy attacking the same broken thing from both sides, and it is a good one. Your site says none of it. Connect reads as an orphan lead form, not as half of a company thesis.

Four of your five buttons point at a JSON file.

The hero offers: View Services, AI Agent API Guide, View AI Agent JSON, Read llms.txt, View OpenAPI JSON. The form’s own subhead reads “AI agents should submit JSON to POST /api/requests.” That sentence sits above the box where a frightened person types what happened to their house.

37.6% of your homepage is developer documentation. The homeowner gets 107 words before hitting an eight-row API endpoint table.

The robot gets a better error message than the human.

Both submit the same incomplete request. Here is what each one gets back.

They filled out ten fields on a phone, one-handed, in an emergency. Now they start over with no idea what was wrong, and because there is no phone number, there is no other way in. For a phone-free product this is not a conversion leak, it is the service failing shut. You have no analytics installed, so nobody knows how often it happens.

One page is two and a half times wider than the screen.

Fifty pages that say the same thing.

You have a landing page for all fifty states. Compare Alabama to California and 253 of 260 words are identical, the only differences are the state name and its two-letter code. Those fifty pages are 81% of all the words on your site.

You’re collecting home addresses with no privacy policy.

The form asks for a phone number and a street address, both required, then hands them to a third-party contractor who will call. There is no consent line, no privacy page, no terms. Breesy’s own policy covers breesy.ai only.

I am not your lawyer and this needs one. I’m flagging it because it is live right now, and it is the only thing on this list whose downside doesn’t depend on traffic.

breesy.ai has never sent a single advertising signal.

Across all 47 pages there is one tag: basic traffic reporting, configured bare. No Google Ads tag, no Meta pixel, no LinkedIn Insight, nothing. Which means no campaign can bid to conversions, and no retargeting audience has ever started filling.

Advertising needs roughly 30 days of tracking history before it can optimize. That clock has not started, which makes this the first thing to fix regardless of what else we do.

One image file. Forty-seven pages.

This is why the creative shoot sits in phase one rather than waiting for a content calendar. It is the longest lead time on the whole engagement and the least compressible, a screen recording can happen next week, a job-site shoot cannot. The single highest-value asset is a 60–90 second demo using real Breesy Voice call audio, because a buyer wants to hear the AI answer a call before they book anything.

Both sites are better than they look.

Neither of these is a rebuild, and that is the reason a quarter is realistic. breesy.ai in particular was built by someone who knew what they were doing, the failures are failures of finishing, not of craft.

The first quarter.

Sequenced around what unlocks what, not around the calendar. Your developer ships the code; we write it, spec it and check the work.

What I would not do.

The job is to make both sites say out loud what the company already believes: the phone call is the broken part of restoration, and Breesy is removing it from both ends. Voice answers the ones that still happen. Connect makes sure they don’t have to.

joanna@bednightstories.consulting