What is Squeegify? The complete guide to AI window-cleaning quotes

Squeegify is a quoting engine that sits on a UK or Irish window-cleaner's website and gives every visitor a real, locked price in ten seconds. The customer types their address, the AI reads the property from satellite and street-view imagery, and a £ figure appears with the time-slot picker right underneath. The operator never has to be in the loop · most replies happen while the cleaner is still on a round.

This guide is the long version of what that means in practice, what it costs, who actually benefits, where it falls short, and how you'd install it on your own site if you wanted to try it.

The problem Squeegify is built around

The single biggest leak in a UK window-cleaning round isn't the price. It's the wait. A homeowner finds your website, fills in your contact form, and then sits there. Most operators reply same-day if they're lucky, next-day usually, sometimes never. In the meantime that homeowner has filled in two more contact forms from competitors. Whoever comes back first usually wins. The honest number on this · in my own round and from talking to dozens of operators · is somewhere between 20% and 40% of enquiries convert. Sixty percent of the work it took to get them to your form just walks.

The fix is to take the operator out of the quoting loop entirely. Not pricing intelligence, not lead-scoring dashboards, not five-step funnels. Just: visitor lands, types address, sees a locked £ price they can book. No 24-hour callback. No "I'll get back to you with a quote tomorrow." The job is on your calendar before the customer has finished their coffee.

That's what Squeegify is. Everything else is supporting machinery.

How the AI actually reads a property

When a customer types an address, three things fire in parallel. Geocoding turns the address into latitude and longitude. Then we pull three aerial satellite views at different zoom levels · close enough to count rooflights, wide enough to see neighbours and plot shape · and two ground-level Street View frames framed at the building. Those five images plus the address get sent to Anthropic's Claude Haiku 4.5 vision model with a structured prompt that walks each elevation, counts the windows visible per storey, infers the rest by symmetry, and identifies the building era and access difficulty.

Claude returns a JSON object: window count per elevation, total, floor count, building type, confidence per field, an access risk note, and a job time estimate in minutes. The deterministic pricing engine takes the counts plus the operator's per-window rate, base call-out fee, floor multiplier, and extras pricing and produces the line-itemised breakdown the customer sees.

The whole thing takes between four and eight seconds. The widget shows a progress bar while it works and falls back to a calibrated default if the AI exceeds a six-second budget. The customer never sees a spinner; they see a quote.

A real example. A 12-window, two-floor semi-detached UK home at default Squeegify pricing produces a £29.11 monthly quote · £8 base call-out, £21 for the windows, £5.25 floor uplift, then 15% off for the recurring discount. The operator can customise every knob in that calculation from their dashboard. Most do, once they've seen how their own market prices the same property.

Honest accuracy caveats

The AI gets the count within ±2 windows on standard residential UK and Irish housing most of the time. It's noticeably worse on Victorian flat conversions with hidden rear extensions, properties with major tree occlusion, and anything with a non-standard build (oasthouses, converted barns, boat houses). For those it lowers its confidence score and flags the lead so the operator knows to phone-verify before locking the price.

The customer's own count is always authoritative on the customer-facing quote · the AI read sits in the operator's pre-arrival brief as an audit trail and is never used to overrule what the customer typed. That decision is intentional. We don't want operators arguing with customers over a window the AI miscounted from a satellite tile.

If the demo on /demo gets the count wrong on your own house by more than two windows, email me. The model improves on every miss someone reports back.

Who it's for

The narrow case: your bottleneck is the moment between a homeowner's enquiry and your reply. Concretely, you have a website with a contact form or phone number. Most enquiries arrive outside working hours. You reply same-day if you're lucky. Your conversion is somewhere between 20% and 40%. You charge somewhere between £15 and £40 for a typical monthly clean.

Less narrowly: small to medium UK and Irish window-cleaning operators · solo cleaners with a Carrd or Wix site, two-to-five-van round operators using WordPress, regional fleets running multi-van rounds through Squeegee or Jobber for scheduling. Squeegify sits at the front of the funnel and feeds the booked job downstream to whatever scheduling tool you already use.

It's not for: commercial-only operators (the model is calibrated on residential UK housing stock), operators who only take phone enquiries (a widget needs a website to live on, though we shipped a QR-code path in v15dw that fixes this), or operators in markets outside the UK and Ireland (the AI is calibrated on UK housing types; outside that the count gets less reliable).

What it actually costs

Solo is £14.99 a month with 30 quotes included, £0.40 per quote after, one user, one widget site. Email + booking confirmation. The break-even is one extra job a month at a £30 quote · most operators clear that in week one.

Growth is £24.99 a month with 100 quotes, £0.30 per quote after. Up to five users, five widget sites, team dashboard, conversion analytics. Sized for a couple of vans.

Scale is £39.99 a month with 300 quotes, £0.20 per quote after. Up to 20 users, branded portal with custom CSS, priority phone support, monthly call with the founder. Built for regional operators.

Fleet is £59.99 a month with 625 quotes, £0.10 per quote after. Up to 100 users, custom integrations, REST API, 99.9% uptime SLA, dedicated founder channel. Built for fleet operations.

All plans share: 7-day free trial with no card required (just pick a plan, get full dashboard access for 7 days, add payment to keep going), monthly or annual billing (annual prepay saves two months), prices exclude VAT, instant access to every feature.

The first ten operators who sign up get FOUNDER25 · 25% off any plan, locked for the lifetime of the subscription. As of writing this is still open. After ten it closes permanently. Solo drops to £11.24 a month with FOUNDER25 applied, which is genuinely under twelve quid before VAT for the engine that books the customer for you.

How you install it

One line of HTML on any website that lets you paste a script tag. The same snippet works on WordPress, Wix, Squarespace, Webflow, Shopify, Carrd, GoDaddy, plain HTML, anything. The script lazy-loads the widget on first interaction so it doesn't slow down your site's initial render. Your brand colour, button label, and call-out copy come from your dashboard config and propagate to every embed within two seconds of you saving.

If you don't have a website at all, we ship a QR-code install path. Print the QR (PNG for A3, SVG for vinyl) and stick it on the side of your van, on business cards, on leaflets, in your Facebook bio. Every scan lands the customer on your branded quote URL at squeegify.co.uk/q/your-slug. About thirty percent of UK window cleaners run van-only with no website, and the QR path is built for them.

If your website is managed by a developer or a website-builder agency, the dashboard has a one-tap "email everything to my web person" button that opens a pre-filled mailto with the snippet, install location hint, and Squeegify support contact. They paste it once, they're done.

Where Squeegify fits with your existing tools

Squeegify isn't a CRM. It doesn't do route planning, recurring schedules with payment history, GPS van tracking, or invoice generation. If your bottleneck is co-ordinating five cleaners across a city, Squeegee or Commusoft is the right buy for that part of your stack and Squeegify is a separate front-end layer on top.

A common setup: Squeegify on the public website for instant quoting, Squeegee or Jobber for back-office scheduling and recurring billing. The booked job flows from Squeegify into the scheduling tool via the Zapier webhook or the REST API on Fleet. Customer details, scheduled slot, quoted price, AI access notes, all pre-filled. The operator's morning isn't "open my CRM and re-enter what came in overnight"; it's "look at the calendar, drive to the first job."

If you don't have a CRM and don't want one yet, Squeegify's lead inbox is enough for solo operators · search, filter by status, customer history, weather forecast for the scheduled slot, recurring schedule detection, calendar sync, photo log, customer messaging. The dashboard isn't trying to be a full field-management product, but it covers the rounds-management surface a solo or two-van operator needs.

The founder honesty section

I'm Robin Oruman. I run Squeegify as a sole trader from the UK. The product is opinionated because I had to ship it under a launch deadline with no co-founder and no investors. That has tradeoffs.

What I'd want a prospective customer to know honestly. The AI is calibrated and improves measurably every week, but it isn't perfect. The operator dashboard is functional and shipped end-to-end · I dogfood it daily on the Fleet plan · but it's narrower than Squeegee or Jobber. The pricing engine is deterministic and tested but it has only seven extras (large window, skylight, bay, French door, conservatory, fascia, solar) · enough for 95% of UK residential jobs, not enough for niche commercial work.

The 7-day trial is real · full dashboard access for 7 days with no card on file. Add payment at day 7 to keep going, or just walk away. If you cancel and want to come back later, your widget config, lead history, and round are all preserved.

The founder email is robin@squeegify.co.uk. I read every message and reply personally. If the demo gets your house wrong, if the install fails on your specific WordPress theme, if the dashboard misses something you need · email me. Squeegify exists because I was annoyed at watching enquiries die in inbox limbo on my own round, and I'd rather build the thing operators actually need than the thing that maximises a metric.

What to try next

The honest test of whether Squeegify is for you takes ninety seconds. Open squeegify.co.uk/demo. Type a real address you know · your own, a customer's, anywhere on the four pre-typed sample buttons. See what the engine says. If the count is within ±2 windows of what you'd quote in person, you've answered the only question that matters.

If the demo lands the count: start the 7-day trial at /signup, paste the snippet on your site (sixty seconds), and watch the next overnight enquiry come in already booked. If the demo misses by 5+ windows on your house, email me before you sign up · I'll either fix the model on that specific property type or save you the trial.

The 25% founder discount with code FOUNDER25 still has spots open as of this article. Once they're gone, they're gone.