AGMM LTD / Privacy
Your information, handled deliberately.
This policy explains what AGMM collects, why it is used, how long it is kept and the choices available to you.
Who is responsible
AGMM LTD, company number 16384156, is responsible for the information described here. For a privacy question or a data request, call 0161 543 7189, or email [email protected].
Website enquiries and bookings
What we keep, for how long, how to stop it: the details you give us, as a business record for as long as it is needed. Contact us if you want us to stop using it or delete it.
If you contact AGMM or request a booking, we use the details you provide to respond, arrange the requested conversation and keep an appropriate business record. Depending on the route you choose, an authorised form, calendar, email, messaging or database provider may process that information for AGMM. The booking forms also ask, optionally, how you heard about us. If you answer, we store the option you pick (for example Google or a referral) with the booking, to see which routes bring real enquiries.
Guides and email updates
What we keep, for how long, how to stop it: your email and what you asked for, until you unsubscribe. Reply unsubscribe to any email.
When you request a working file from an AGMM YouTube video, AGMM stores the email address you submit, the video and resource selected, the page or campaign source, the consent wording shown at the time, the request date and whether the delivery email succeeded. We use this information to deliver the requested file and send occasional practical emails about AI systems and AGMM services. The form states this exchange before you submit it. You can withdraw at any time by replying unsubscribe to any email. We do not sell or share this list with advertisers.
Daily AI tips by email
Your choice: this is a separate email list. Choosing cookies does not sign you up.
If you ask for the daily AI tips, we keep your email address, the signup page, the time and wording of your request, and whether you confirm or unsubscribe. Submitting the form joins the list immediately. You receive a welcome email explaining AGMM, followed by a practical AI business tip. The series may link to relevant AGMM services and the free fit check.
Every email includes an unsubscribe link. Unsubscribing stops this series and keeps a record of that choice so we can respect it. It does not change any other email service you separately requested. You can also ask us to delete your record using the contact details above. Your email address is not added to website analytics or shared with advertisers.
The AI Readiness Assessment
What we keep, for how long, how to stop it: nothing, unless you tick the consent box and submit your email for the four-week plan. After that, your record stays until you unsubscribe with one click or ask us to delete it.
While you answer
The assessment at /ai-readiness-assessment/ runs in your browser and shows you the whole result, including the score, the maps and the steps, without asking for anything. While you are answering, your answers are held only in your own browser tab, so a reload can restore them; they are discarded when you close it or select Start again. Nothing reaches AGMM unless you then tick the consent box and submit your email address to receive the four-week plan.
If you ask for the plan
AGMM stores your email address, the niche and size band you chose, the fifteen answers you gave, the score and band calculated from them, the steps the assessment produced, the time you gave consent and the wording version shown to you, and a random unsubscribe code. We use the current request to send your complete result and next steps in one email. It does not subscribe you to the earlier weekly sequence. We do not sell or share the list with advertisers.
How to stop it
The result email carries a one-click unsubscribe link, which records your withdrawal and needs no sign-in. You can also reply unsubscribe to any of them, or ask us to delete the record entirely using the contact details at the top of this page.
Website measurement
What we keep, for how long, how to stop it: anonymous counts always, anonymous aggregate statistics unless you turn them off, detailed events only if you accept. Original records are kept for 90 days; aggregate daily totals are kept for 13 months. In this section “cookies” means anything the site keeps in your browser, including browser storage (localStorage and sessionStorage). A historical archive retains non-identifying counts and, for the earlier website design, anonymous visit summaries, with identifying and free-text dimensions removed. Select Analytics choice at the foot of any page to decline or change your mind.
| Data | Purpose | Kept for |
|---|---|---|
| Anonymous counts, always on: per five minutes, combined counts by page, campaign tags, referring website host, broad device category, country supplied by Cloudflare, and the name of the network the request came from (for example a home broadband provider or a hosting company) | Count page loads without an analytics cookie or a visitor identifier. The network owner name is recorded only to tell people from automated traffic such as crawlers and hosting networks | 90 days |
| Aggregate statistics, unless you turn them off: one anonymous summary of a visit, sent when you leave, added to daily totals. The summary holds the pages opened, how far each was scrolled, seconds on each, which 5% square of a page each click landed in, whether you pressed a main call to action or opened the calendar, a broad source group, and your device size. It holds no code, no network name and no time finer than the day | Understand how the site is used in total, for example which pages lead on to which. Used for statistics only and not shared. Nothing in a summary can be matched to you, to another summary or to the detailed events | 13 months, as daily totals. The summary itself is kept in this tab only, until you leave |
| Detailed events, only if you accept: what you scroll, click and complete, as described below | Measure how pages, sections and forms are used | 90 days |
| Visitor code and session code, only if you accept | Link the events from one visit and one browser together | The visitor code expires 13 months after it was made, or sooner if you withdraw your choice; the next visit then gets a new code that is not linked to the old one. The session code is discarded when the browser tab closes. |
Anonymous counts, always on
The main website records anonymous aggregate page loads without setting an analytics cookie or creating a visitor identifier. For each five-minute period, AGMM stores combined counts by page, campaign tags, referring website host, broad device category and country supplied by Cloudflare. The analytics database does not store an IP address, raw browser string or full referring URL for these counts.
The network name is the organisation that runs the connection, taken from the request. It is a business or internet provider, not a person, and it is stored on those anonymous counts for the same 90 days.
Aggregate statistics, unless you turn them off
If you have not accepted detailed analytics and have not turned statistics off, the site keeps a short summary of your visit in this tab and sends it once when you leave. It is added to daily totals; the totals say, for example, that 31 visits from search read two pages and 9 pressed a call to action. The summary has no visitor code, session code, IP address or network name, so no total can be traced back to you, and it is never combined with the detailed events. It starts only after the banner has been on screen for a few seconds, or on your second page. Any total smaller than 5 is hidden when the figures are read. It is used only for statistics about how the site is used, is not shared, and is not sent by browsers that signal Global Privacy Control or Do Not Track. Choosing Reject cookies on the banner, or in the Analytics choice control at the foot of any page, stops it at once.
Detailed analytics, only if you accept
The website asks before enabling detailed first-party analytics. The banner states the purpose in short; this section is the full description it links to. If accepted, a random visitor code and session code are stored in the browser so AGMM can measure:
- movement between pages, how far down a page is scrolled and active time on a page;
- which sections of a page come into view, and which links and buttons are clicked;
- where on the page a click landed, recorded as a horizontal and vertical percentage of the page, never as a screen recording or a cursor trail;
- which stage of an enquiry or booking form was reached, and whether a named field was completed;
- broad viewport size and page performance;
- signs of difficulty: several fast clicks on one spot, a click on something that looks pressable and does nothing, and how often a page was scrolled back up;
- that an email address or phone number on the page was copied, as the single word “email” or “phone”, never the text copied;
- how long after the page opened you first did anything or first pressed a call to action, how long a form field was in focus before you typed in it (never what you typed), how many times you left the tab, and how long you had been away since your last visit.
Form progress records only the fixed option selected or the name of a completed field. It never records names, contact details or free-text field contents unless the form is deliberately submitted as an enquiry or booking. No IP address is stored against these events, and nothing is shared with advertisers or any third-party analytics service. Original aggregate records and detailed website events are retained for 90 days. To compare earlier website designs, we preserve non-identifying historical counts after stripping visitor and session codes, contact data, query strings and arbitrary campaign or free-text labels. For the earlier website design only, we also keep one anonymous summary per visit: the pages opened, how far down each page was scrolled, how long each page was visible, the number of clicks, device category, country and broad traffic source, under a new random label that links to nothing else. A summary carries no visitor code, session code, IP address, contact detail or free text. Those counts and summaries are retained for long-term website analysis; Alfred conversations are not retained in that archive.
Your choice
The banner offers two choices of equal weight. Accept cookies starts the detailed analytics above. Reject cookies keeps them off, stores no code and also stops the anonymous aggregate statistics, so nothing is measured from your browser. Until you choose, your visit can add to those anonymous totals. The site works the same whichever you choose. The banner counts itself and the button pressed as an anonymous tally with no code, so we can tell how many people answer it.
Your choice is stored in this browser and you can change it at any time. Select Analytics choice at the foot of any page to bring the request back. Declining stops the measurement, and the visitor and session codes held in your browser are deleted. If your browser sends Global Privacy Control or Do Not Track we treat it as turning statistics off: no aggregate summary is sent, and the detailed analytics start only if you press Accept yourself. Your answer is asked again after 180 days, or sooner if what is measured changes.
The separate campaign pages, formerly called the AI Bottleneck Blueprint campaign pages, use their own measurement choice, described on the campaign analytics privacy page.
Questions you type to Alfred
Alfred is the small companion robot on this website. The questions you type to Alfred are sent to Cloudflare Workers AI to produce an answer from this website's own pages. Saving questions and replies for AGMM review is optional and off until you choose it in the chat. If enabled, AGMM keeps the conversation, page, device category and response time for up to 90 days, accessible to authorised AGMM staff. If you also accept website analytics, the conversation can be associated with your random browser and visit codes. AGMM does not infer your identity from these codes. You can switch saving off or delete this browser conversation using the controls in the chat. Please do not enter sensitive information. Questions already sent to the AI service cannot be recalled by withdrawing this choice. To stop abuse and cap running costs, AGMM also counts how many answers each visitor and the whole site have asked for in the last hour or day, against a one-way code made from the network address and the date. These counters hold no question or reply text and are deleted within two days.
Website measurement stays with AGMM
Current setting: this website, including the AI Constraint Audit ads page, measures visits inside the AGMM Suite only. It does not load Meta Pixel or send booking conversions, contact hashes or browser identifiers to Meta.
The ads page uses the same optional first-party website analytics described above. Anonymous server-side totals count page requests and completed form stages without a visitor identifier. Detailed scroll, active-time and interaction events start only after you allow website analytics. Select Analytics choice in the footer to change it at any time.
Earlier Meta choices
An older page offered a separate Meta measurement choice. That earlier choice does not enable Meta tracking on the current website. Existing browser preferences and signed HttpOnly receipt cookies expire after 180 days. A minimal random revocation record can be retained for no more than 180 days after withdrawal to stop an old receipt being reused. Historical consent evidence remains with its booking under the business-record retention below.
Existing Meta cookies _fbp and _fbc may remain from an earlier visit; this website does not use them for new Meta events. Meta describes its own retention and account controls in its Cookies Policy, Privacy Policy and activity off Meta technologies settings.
TikTok connection
What we keep, for how long, how to stop it: an operator's TikTok account details and short-lived sessions, only for authorised AGMM staff. Remove the connection in TikTok's settings, or ask us to delete the session.
AGMM Auto-Poster uses TikTok Login Kit and the Content Posting API. When an authorised AGMM operator connects TikTok, TikTok supplies an account identifier, basic profile information, permission scopes and access credentials. The service uses these only to show the connected account, present TikTok’s current publishing choices and submit content that the operator has reviewed and approved.
Access and refresh credentials are encrypted and are not exposed to page scripts. The review interface does not sell TikTok data, use it for advertising profiles or share it with unrelated third parties. An operator can remove the connection in TikTok’s settings or ask AGMM to delete the associated session.
AI assistants and the AGMM connector
What we keep, for how long, how to stop it: daily totals by assistant name, tool and outcome for 400 days, a request log for 30 days, and a payment record when an AI agent pays for data. No IP address, user agent text or search words are kept. Remove the connector in your assistant’s settings.
AGMM runs a server that AI assistants can connect to at agmm.co.uk/mcp, and a free, search-only version of it at agmm.co.uk/mcp/learn. Both are described on the connection page. They need no account and set no cookies. When an assistant uses either one, AGMM keeps daily totals: the date, the name and version the assistant gives for itself, whether it connected, listed the tools or called one, which tool, and whether the call worked, asked for payment or was paid. Each total also carries a broad label for the kind of software making the request, such as “Node script” or “curl”, and the gate’s own class for the visitor, never the text of a user agent. A visit to the AI Constraint Audit page from a link marked as sent by an agent is counted in the same way, as a date and the label on the link. Requests to AGMM’s servers can also appear in a request log kept for 30 days, with the time, the path, the country and network name, the kind of visitor, the outcome and a request number.
When an AI agent pays for a data tool, AGMM keeps the payment record: the transaction number, the paying wallet address, which is public on the blockchain, and the item bought. That is a financial record and is kept as one. AGMM does not keep your IP address in these records, the text of your user agent, your search words or any other input sent to a tool, or anything that names you. A connection address is used for a moment as a rate-limit key on payment attempts and is not written to AGMM’s records.
The site runs on Cloudflare, which handles connections as AGMM’s network provider and sees ordinary connection data under its own terms. Hosted assistants such as Claude and ChatGPT connect from their own company’s servers, so AGMM does not receive your address. Tools that run on your own machine connect from it. To ask about, or to ask AGMM to delete, anything connected to you, use the contact details at the top of this page.
The x402 gate on agmm.co.uk
What this means for you: people, search engines and AI answer engines read the site free. Named AI training crawlers are asked to pay a small fee in USDC. Nothing about a person is added to what the section above already lists.
Requests to agmm.co.uk pass a gate that decides, for each request, whether to ask for payment. It reads the visitor’s user agent at that moment to tell a named AI training crawler from everyone else, and keeps the result, which is the class and the operator’s name, not the user agent text. Only a named AI training crawler is asked to pay, using the x402 protocol. Everything else stays free: people in a browser, search engines, AI answer engines, assistants reading a page for a person, link previews, and any visitor the gate cannot identify with confidence. Nobody is charged on a guess.
What is kept about gate traffic is described in the section above: a request log for 30 days and daily totals for 400 days, with no IP address. When a crawler or agent does pay, the payment record in that section applies.
Paid data products and the AI visibility check API
What we keep, for how long, how to stop it: for a payment or a credit pack, the paying wallet address, the transaction and the amount, as a financial record. The website names you check are not written to the request log or the daily totals.
AGMM sells machine-readable data (fact packs, tracked AI terms and daily news digests) and the AI visibility check as an API. Each can be paid for per request with x402, or, for the check, from a prepaid credit pack.
When you buy a credit pack, AGMM keeps the paying wallet address, the transaction number, the size of the pack, the balance left and when it was last used. Each use adds a line with the date, the amount and whether it was one check or a batch. The secret token you are given is not stored: only a one-way fingerprint of it is, so AGMM cannot read it back to you. These are financial records and are kept as such.
The website names you ask about are not written to the request log, the daily totals or the credit lines, and a batch payment records only how many websites were in it. If you pay for a single check per request rather than from credits, the payment’s own record holds the address you paid for, which includes the website name, together with the payment details. To make repeat checks quick, a result is held in a short cache, keyed by website name and not by you, for about ten minutes, and then expires. The website you check will see a few ordinary requests from AGMM for its public files, with a user agent that names AGMM.
Use of the paid products is covered in the terms.
The free AI visibility check
What we keep, for how long, how to stop it: the website name you type and its result are not stored against you. An hourly count per network address is kept only to stop abuse.
The free check at /learn/tools/ai-visibility-check/ needs no sign-up. It fetches a handful of public files from the website you name and shows you the grade. AGMM does not store the website name or the result against you or your device. As above, a result is held for about ten minutes by website name so a repeat check does not hit the site again, then expires. The page may also put the name into your own browser’s address bar so you can share the link; that stays in your browser.
To keep the check from being used to overload other people’s sites, AGMM counts the live checks made from each network address in each hour, and across the whole site. That count holds no website name and is used for nothing else.
If the check offers a box to have the result emailed to you or to hear from AGMM, using it is your choice and it is left empty until you fill it in. If you do, AGMM stores the email address and the website name you checked, to send you the result, and uses them for nothing else unless you also agree to it. You can ask us to delete them using the contact details at the top of this page.
Companies we write to
Where we got your details, why we use them, how to stop it: from Companies House and your own public website, so we can send a limited company a short, private note about it. Reply stop and we do not write again.
AGMM may write to a limited company or LLP with a private receipt: a short note of what our public check of the company’s own website saw, with the date, such as whether its robots.txt, sitemap or llms.txt let AI crawlers in. We write to the company’s generic mailbox or contact form as printed on its own website, not to a named person’s address, and only to limited companies and LLPs.
The company details come from two public sources: Companies House (the company name, number, type and the nature of its business) and the company’s own website (the public files our check reads, and the contact route it prints). Our lawful basis is legitimate interests: we think a company that may be turning AI crawlers away without meaning to would want to know, the note is about its business, and it costs the recipient one reply to stop it. If a message does name a person, you have the right to object to it, and we will stop and keep your address on a do-not-contact list so we do not write again.
To opt out, reply stop to any message, or use the contact details at the top of this page. For the rest of your rights, see below.
Retention and security
AGMM keeps business records only for as long as they are needed for the purpose collected, legal obligations, security and resolving disputes. TikTok browser sessions are short-lived. Production publication receipts retain the post identifier, destination, approved caption and media evidence needed to operate and audit AGMM’s own publishing system; they do not contain the access token. How AGMM handles client information during an audit or a build is written on the security and data protection page.
Your rights
Depending on the circumstances, UK data-protection law may give you rights to access, correct, delete or restrict use of your personal information, object to certain processing, or receive a portable copy. You may also complain to the UK Information Commissioner’s Office. Contact AGMM first if you would like us to investigate or act on a request.
Updates
Last updated 20 October 2026. Material changes will be reflected on this page.