AI for pressure washing companies: photo quotes, rain days and job sheets.

A pressure washing or soft washing company can quote a job before anyone stands in the driveway. The customer sends pictures of the siding, the roof, the patio or the fence, and the owner prices from them. Then the job happens on a real street, next to real flower beds and a real drain inlet, and the company is answerable for what its rinse water carries away. A quote that is right on price but silent about the drain is only half the paperwork this trade needs.

Supervised AI assistants can carry three parts of that paperwork: line-item quote drafts from the photos and the owner's own price sheet, proposed new dates when the weather turns, and a pre-job sheet for every address. The method and the mix, whether a surface can take a wash, how a pre-1978 painted wall is handled, where the wash water goes and every message to a homeowner remain the owner's to settle.

A photo inquiry brings an address, a list of surfaces and a handful of phone pictures. From those an estimator in this trade looks for the size and number of stories of each elevation, the cladding (vinyl, painted wood, stucco or render, brick), what is growing on it, whether painted surfaces are chalking or oxidized, the roof material, and the state of any deck or concrete. The year the house was built goes on the record too, because a house built before 1978 raises a lead question that has to be answered before any water touches it.

Photos settle some of this and not the rest. Square footage and stories can usually be read or estimated; the grip of the growth, the softness of old paint and the slope behind the house often cannot. The table shows the split the owner can write into the estimating Bot's instructions.

In Grok Bot's documentation a Bot is a durable teammate: it has a name, a job and a conversation of its own. One user's Bots all work from a single persistent cloud computer equipped with a browser, storage for files and a terminal. Where a service offers a connector they use it, and many sites without one can still be worked through the browser, with passwords, verification codes, CAPTCHAs and payments handed back to the owner, who takes over the computer for that step. Cursor's help pages add a Vault, where saved logins and cards let a Bot sign in and check out without seeing them, and no setting asks for approval each time a saved card is used, so this design saves no card there. A routine runs a Bot's job on a schedule, with at least five minutes between scheduled runs. Nothing in the documentation covers telephone calls or texts, so the phone stays with the owner.

The documentation also advises starting with the smallest roster that does the job. For an exterior cleaning company the order follows the work: an estimating Bot, because every job starts as a set of photos; a scheduling Bot once quotes are being accepted faster than the owner can slot them; a job-sheet Bot that prepares the pre-job record for each address; and a marketing Bot for before-and-after posts and replies to reviews. The four jobs come from Aevrion's pattern for this trade; the product supplies Bots, and the owner sets up and describes each of these.

All four share one computer and its folders. That makes handoffs easy, and it also means a login saved for one is usable by all of them: the documentation says separate Bots are not a security boundary, so anything that must stay apart, such as the business bank, belongs under a different user account.

Based on the Grok Bot documentation at docs.x.ai as checked on 15 October 2026. Its statements on Bot rules, Bot-to-Bot messaging, routine triggers, connected apps, approval expiry and payments were rechecked against docs.x.ai and Cursor's help pages on 26 October 2026. The trade examples are Aevrion's illustrations, not documented Grok Bot use cases. Aevrion Ops is an independent consultancy. It is not affiliated with, endorsed by or a partner of xAI, SpaceXSI or Cursor. Grok Bot is named here only to describe it.

The customer's pictures do two jobs. The estimating Bot reads them for size and surfaces; the job-sheet Bot needs the same pictures to spot the drain inlet at the end of the driveway, the beds against the foundation and the car parked downwind. The owner attaches them to the estimating Bot's conversation, or a custom integration with the inquiry form saves them to a folder that a scheduled routine on the Bot reads, tested first and subject to what the form provider allows.

The pictures never travel through a group chat. The documentation says a Bot's handoff into a group is currently limited to text, and that an image belongs in a direct message to whichever teammate has to inspect it. So the estimating Bot sends the elevation photos straight to the job-sheet Bot, and the group sees only a line saying they were sent. Cursor's help pages say no switch turns off messaging between Bots; a Bot follows instructions about it.

The choice of method never travels with the photos. Whether a wall gets a low-pressure soft wash or a pressure wand, and what goes in the tank, is written on the quote only after the owner has decided it.

Consider, as an illustration, a request for a single-story brick ranch with painted wood trim, a stained concrete driveway and a cedar fence along one side. The estimating Bot drafts four lines from the owner's price sheet: the house wash by elevation, the trim noted separately because paint condition is unclear, the driveway by area, and the fence by length. It adds the lines the owner has made standard for this kind of job, such as plant protection, and a wash-water recovery line where the job sheet shows the local rule will need one. Where the photos leave a question, the draft says so rather than guessing.

Prices are looked up in the price sheet file on each draft, since the documentation says facts that change belong in their source system, not in a Bot's recollection. Reading a price list or customer record held in quoting or customer-management software is a custom integration or supervised browser use where the vendor permits it, subject to testing.

The quote waits as a draft in the estimating Bot's conversation. The documentation describes editable email drafts that the account holder reviews, then sends or discards, so the owner checks the price, the method line and the wording before anything goes out.

Rain dilutes a soft-wash mix and wind carries it onto cars, glass and the neighbor's hedge, so the forecast reorders the calendar all season. A scheduling Bot can run a routine that compares the forecast with the booked jobs and drafts a new order, which the owner approves before any homeowner hears of a change; the spring-rush article follows that routine, and the owner's sign-off, through one operator's April.

Wash water is regulated locally, and the rules change from one city to the next. One California city, Lathrop, bars pressure washing wastewater from storm drains, gutters and streets and lets collected water go to a sanitary sewer drain only after asking the property owner, never to a septic system (as checked on 16 October 2026). Another city can draw those lines elsewhere, which is why the job-sheet Bot researches the rule for each address and the owner confirms it.

The sheet itself lists the storm inlets near the work, the planned route for the water, the plants that need protecting, the recovery equipment the job will need and the year the house was built. The companion checklist article for this trade goes through that sheet line by line, with three cities' rules side by side. A house built before 1978 is only flagged on the sheet: the decision on lead-safe work belongs to the owner and a renovator who holds EPA's renovator credential, and the painting article on RRP records explains what that rule asks for.

Two kinds of qualification get mixed up in this trade's marketing. The Power Washers of North America describes itself as a non-profit trade association; it offers certificate courses in house washing, roof cleaning and flatwork, plus safety and environmental EPA training, mostly online. Taking them is the owner's choice. EPA's RRP rule is another matter: a legal requirement (as checked on 16 October 2026) for firms performing renovation, repair or painting that disturbs lead-based paint in a pre-1978 home, set out in the painting article on RRP records.

The marketing Bot keeps those apart. It drafts before-and-after posts from photos the owner selects and replies to reviews in the owner's voice, naming only courses the owner has actually completed and never implying a credential the firm does not hold. The documentation lists publishing content among the actions to keep behind a clear stop for approval, so each post is the owner's to release.

The estimating Bot is only as good as the price sheet and the photo checklist it reads. If prices still live in the owner's head, or each inquiry gets a different set of questions, writing those two documents is the first piece of work, and it is useful even if no assistant is ever built.

Hand photo quoting to Aevrion's readiness assessment as its one workflow. The assessment is free, sits behind no form and, across twelve fixed dimensions, returns a score, the blockers it found and a suggested next step at once. Unwritten prices will show up as a low mark for rule clarity, and a weak result on rule clarity, ownership clarity or security and compliance readiness caps the overall score. Once the quote workflow rates well, Aevrion Ops can take it into a scoping conversation about the four Bots, the job sheet, and custom integrations to the inquiry form and quoting software, each depending on the vendor's API or terms and on testing. For the spring rush itself, the companion article follows one soft-wash operator through an April backlog.

1. Exterior cleaning is priced from customer photos and answered for at the storm drain, so each quote travels with a pre-job sheet.
2. An estimating Bot drafts line-item quotes from the photos and the owner's price sheet; the owner picks the method and mix and sends every quote.
3. Group handoffs are text only, so the photos go directly from the estimating Bot to the job-sheet Bot that needs to see the drains and beds.
4. Weather reorders the calendar; a routine can propose the new order, and the owner confirms it before any homeowner is told.
5. Lathrop, California is one example: it bars storm drains and septic systems and allows the sanitary sewer only with the property owner's permission.
6. Trade-association courses are the owner's choice; RRP firm certification is a legal requirement for some pre-1978 work, and a person makes that call.