Terms of use

Terms of use and how we estimate

Agentic Translator is a tool that turns a plain-language description of a business into automation ideas, with rough estimates of the time each idea could save. These terms explain what the tool is, what it is not, how the numbers are produced, and the ground rules for using it. By using the site you agree to them.

Last updated 11 September 2026 · Operated by Legion AI LLC, Texas, USA

1. What the tool is, and is not

You describe work your team does by hand. The tool matches that description against a catalogue of automation patterns and uses an AI model to decide which ones fit, then shows you cards describing each idea, what it would need, and an estimate of the time it could take on. Some cards are "matches" to work you described; others are ideas to explore, shown with an "If…" condition and no savings figure.

The output is generated by software, including AI language models, and is meant as a starting point for a conversation. It is not a quotation, an audit of your business, a promise of results, or a commitment by anyone to build anything. Nothing on the site creates a contract to deliver an automation. If you ask for help through the site, any work that follows is agreed separately, in writing.

The site is currently a private pilot. Features, wording and estimates will change while we learn from testers, and the service may be paused or withdrawn at any time.

2. How we get the numbers

Every hour or dollar figure on a results page is an estimate built as follows.

Hours a month

  • Each automation in our catalogue carries a typical range of monthly hours that small businesses spend on that kind of work by hand. The ranges are our own catalogue estimates, informed by industry benchmarks and experience. They are not measurements of your business.
  • We pick a point in that range based on the size of your team: the number you give us on the form if you give one, otherwise head-counts in your own words ("two dentists and four staff"), otherwise we assume a small to mid-size business. The card says which of these we used.
  • We then apply the share of that work the automation can realistically take on. That share is never shown as more than 92 percent, because someone still has to handle exceptions.
  • The form asks how often the work you described comes up and how long it takes on those days. If you answer both, the matched cards are scaled so their monthly hours add up to the middle of the range your answers give (with the same automatable share applied to each), and they are marked "from your answers". If you leave the questions blank, a matched card offers the same two questions for that task alone ("Make this number yours"). Your answers are kept beside the catalogue figure.
  • Once five or more businesses have given their own figures for the same catalogue entry, and those figures broadly agree, new cards for that entry show the middle of those figures instead of the catalogue range, and say so ("based on N businesses' own figures"). Where the figures disagree widely the catalogue figure stands. Figures are only ever combined this way; no business's individual answer is shown to anyone else.

Dollars a month

  • A dollar figure appears only when you enter an hourly pay rate. It is the automatable hours multiplied by that rate, reduced by 15 percent because freed-up time is never redeployed in full.
  • Where a range is shown, the lower figure is 60 percent of that amount (the conservative case) and the higher figure 140 percent (the optimistic case). Both come from the same hours.
  • These are staff-capacity figures, not cash. Freed hours only turn into money if you use them for other work or reduce overtime or hiring.

Things to keep in mind

  • Estimates for different ideas can overlap. The same front-desk hour can appear under two ideas, so the headline total is an upper bound on one team's recoverable time, not a sum you should bank.
  • Ideas to explore are not counted in the total until you tick "This fits" on one; then its typical figure, or your own answer to the two questions, is added and the page says so.
  • Setup effort, ongoing costs, software licences, change management and the cost of any developer's time are not included unless a card says so.
  • Your actual results depend on your systems, your data, your people and the developer you work with. We do not guarantee that any estimate will be achieved.

3. Not professional advice

The site provides general information about automation. It does not provide legal, tax, accounting, financial, employment, medical, insurance or other professional advice, and nothing on it should be relied on as such. Decisions about staffing, spending, compliance or the handling of regulated information remain yours, and you should take advice from a qualified professional where it matters.

Mentions of particular software, tools or vendors are examples. We are not affiliated with them unless we say so, and we do not endorse them.

4. What you share with us

You keep ownership of everything you type or upload. By submitting it you give Legion AI LLC permission to process it, send it to the AI and hosting providers we use, keep attached files with your run, and keep it all as described in the privacy policy (opens in a new tab). We also keep descriptions, the recommendations they produced, and your thumbs-up or thumbs-down responses to train and improve our matching system, as the privacy policy describes.

You are responsible for what you share. Please do not submit:

  • passwords, account credentials, card numbers, bank details or government identifiers;
  • medical records, patient information, or anything protected by health-privacy law;
  • material covered by attorney–client privilege or a confidentiality obligation you owe to someone else;
  • personal information about other people (employees, customers, clients) beyond what is needed to describe the work in general terms;
  • anything you do not have the right to share.

The tool does not need any of this to work. If a file contains no recognisable description of a work process, we tell you so and do not use it.

5. Acceptable use

You agree not to use the site to break the law, to probe or overload it, to try to reach other people's reports or the staff area, to submit content designed to manipulate the AI into producing harmful or misleading output, or to copy the catalogue or results for resale. We may limit the number of analyses you can run in a period so the service stays available to everyone.

6. Accounts and the preview password

During the pilot the site sits behind a preview password. It is for the person we gave it to; please do not share it or post it. If you sign in with Google or an emailed code to save reports, you are responsible for activity under that account, and you may only see reports you created. We may close accounts that break these terms.

7. Developer network applications

Applying to the developer network is an expression of interest. Submitting an application does not create employment, a contractor engagement, an agency relationship or any promise of work or payment. Any engagement, including referral or fee arrangements, would be set out in a separate written agreement.

8. Services we rely on

The site is hosted on Vercel, stores data with Supabase, generates its analysis with Anthropic's AI models, uses OpenAI for the rollout roadmap (not shown during the pilot), uses Google for optional sign-in and web fonts, loads one script library from the jsDelivr network, and sends email through a mail provider. Each of those providers handles data under its own terms. The privacy policy lists what each one receives.

9. No warranty, and limits on our liability

The site and everything it produces are provided "as is" and "as available", without warranties of any kind, express or implied, including any warranty of accuracy, fitness for a particular purpose, merchantability or non-infringement.

To the fullest extent permitted by law, Legion AI LLC and its members, staff and contractors will not be liable for any indirect, incidental, special, consequential or punitive damages, or for lost profits, lost savings, lost data or business interruption, arising from your use of the site or reliance on anything it produces, even if we were told such damages were possible. Our total liability for any claim relating to the site is limited to the amount you paid us for it in the twelve months before the claim, which during the pilot is nothing.

Some jurisdictions do not allow certain warranty exclusions or liability limits, so parts of this section may not apply to you. Nothing here limits liability that cannot be limited by law.

You agree to hold Legion AI LLC harmless from claims arising from content you submit in breach of section 4 or from your breach of these terms.

10. Changes, suspension and ending use

We may change these terms and the site at any time. The date at the top tells you when the terms last changed; continuing to use the site after a change means you accept it. We may suspend or end access to the site, or to any account, at any time, including to protect the service or other users. You may stop using the site at any time, and you can ask us to delete your saved reports as described in the privacy policy.

11. Governing law

These terms are governed by the laws of the State of Texas, without regard to its conflict-of-law rules. Any dispute that cannot be resolved informally will be brought in the state or federal courts located in Texas, and you consent to their jurisdiction. If any part of these terms is found unenforceable, the rest remains in effect.

12. Contact

Questions about these terms, or about a result you were shown, go through the enquiry form. Please do not include sensitive information in the message.