Ask an AI automation agency what it costs and you tend to get one of three things: a number on the page, a form that asks for your email before it shows you a number, or nothing at all beyond "book a call". We wanted to know which of those is actually true across the category, not which one agencies imply is true, so we read the live pages rather than the copy written about the live pages. What follows is what we found, with the source and the date we looked, because a number without either is not a number worth repeating.
This is not a ranking. We looked at two agencies with different approaches to publishing price, and at the wider pattern around them, and wrote down exactly what each page shows on the date we visited it. Prices and page layouts change, so every figure below carries its source and its date, and we have set a reminder to recheck them.
MQLFlow publishes a full price list
MQLFlow is a UK automation consultancy that puts prices on the page itself rather than behind a call. The same list repeats across the homepage and all four service pages (mqlflow.com, accessed 2026-08-22):
| Item | Published price |
|---|---|
| Automation Strategy | From £3,200 |
| Automation Set-Up | From £800 |
| Retained Services | From £200/month |
| AI Agents | From £4,000 |
| Live Data Dashboards | From £2,400 |
| Personalised Outreach | From £1,200 |
| Day rate | £800 (excludes software usage fees) |
There is no dedicated pricing page; mqlflow.com/pricing returns a 404 (checked 2026-08-22). The list is a recurring page module instead, so a visitor reads the budget conversation before they ever reach a booking form. Of the agencies with any published numbers at all, this is the more direct mechanic: the price sits where the offer is described, not one click further on.
Agency pricing moves, so we will re-verify this list quarterly. Next check due November 2026.
GetAutomated frames itself as flat-price, then gates the price
GetAutomated sells a monthly automation subscription and leads with "one flat monthly price" as the core promise (getautomated.agency, accessed 2026-08-22). The homepage pricing section shows several tiers, and each one renders in the page's own HTML as a placeholder rather than a real figure. The actual number only appears after a visitor types a work email address into a form and clicks through.
So the one thing the page states as its whole pitch, a flat price with no hidden surprises, is the one thing you cannot see without handing over contact details first. Everything else about the tiers, such as what complexity of build each covers and how support is delivered, is public on the page; the price itself is not. That gap between the stated policy and what the page actually shows is worth naming plainly: a subscription pitched on transparency and a subscription that shows you the number are not automatically the same product.
Most agencies publish neither
MQLFlow and GetAutomated sit at two different points on a small spectrum, and most of the category we looked at sits off that spectrum entirely: no price list, no gated tiers, just a contact form and a call. That is ordinary for services businesses generally, and we are not claiming otherwise about the market as a whole. It does mean that if you are trying to shortlist an AI automation agency by price alone before you speak to anyone, the open web currently gives you very little to go on, and a first call usually has to do the work a price page would otherwise do elsewhere.
We are one of the agencies in that third group. We do not publish tiered prices, because what a function costs to run depends on what the function actually is once we have mapped it, and a from-price on a page cannot carry that. Instead we publish the audit that produces the number: a free audit call that ends with a named function, a cost basis, and a plain recommendation, whether or not it points at us. That is a different transparency mechanic from a price list, not a replacement for having one, and it is worth stating outright rather than leaving it implied.
What to check instead of comparing brand names
A published number is useful, but it answers less than it looks like it answers. Before you compare agencies on price, and before you compare us against any of them, these are the questions a price list does not settle:
What you are actually buying. A day rate buys hours. A build fee buys an artefact, delivered once. A retainer can mean "we keep the thing we built running" or it can mean "we run the function for you, on a cadence." Those are different products that can carry the same-sounding label.
What happens after handover. If the retainer is maintenance, the agency is keeping something alive that they built once. If the retainer is verification, the agency is proving, on a schedule, that the outcome still holds. Ask which one is being sold; the word "retainer" covers both, and the invoice looks the same either way.
Whether the proof is checkable. An hours-saved figure or an ROI percentage is a claim. A number with a source, a date, and a method you could re-run yourself is a receipt. Ask any agency, us included, what happens if you ask to see the working behind a number on their site.
What a risk-reversal clause actually covers. A refund window, a pause-anytime option, and a guarantee tied to a scorecard are three different promises that get described with similar-sounding phrases. Read what triggers the protection, not just that a protection exists.
None of this is specific to MQLFlow or GetAutomated; it applies to any agency you are weighing up. We would rather show our own working here than ask you to take it on trust: receipts carries the numbers we publish alongside the commands that reproduce them, and case studies describes engagements without inventing outcomes we cannot show.
If you want a direct answer on what running a function like this would cost for your business, and what proof you would get for it along the way, book the free audit. No price list, ours included, stands in for that conversation.