ClueDig

Legal

AI Transparency Notice

Effective date: [EFFECTIVE DATE] · Last updated: [LAST UPDATED DATE]

Bracketed text in [ALL CAPS] marks a placeholder that must be filled in with real, verified information before this document is relied upon as a binding legal agreement. Do not remove this notice until every placeholder below has been replaced.

1. Purpose of This Notice

Artificial intelligence is core to how ClueDig works: it is how we turn a competitor's web page into a plain-English analysis. This Notice explains, in plain terms, what AI ClueDig uses, what data it sees, how its output is produced, and — most importantly — its limitations, so you can use it responsibly. This Notice is incorporated into our Terms of Use by reference.

2. What AI ClueDig Uses

ClueDig may use several large language models to generate the analyses, summaries, and battlecards described below. AI processes the specific inputs described in Section 3 as our subprocessor. The AI models we use do not train on the data we send it through ClueDig's use of its API. We may change which underlying AI model or provider we use at any time, including to a different model from the same or a different provider, without prior notice, as described in Section 10.

3. What Data Is Sent to the AI

For a standard competitor analysis or an on-demand battlecard, we send the AI model:

  • the competitor name and URL you provided;
  • any optional focus notes you entered for that competitor;
  • the publicly available text content of the competitor's web page at the URL you submitted, fetched by our system at the time of analysis; and
  • the display language your account is currently set to, so the output can be generated in that language.

We do notsend your account credentials, payment or billing information, other users' data, or any information about your own business to the AI model as part of generating a competitor analysis. If the competitor's web page cannot be fetched (for example, because the site blocks automated requests), the AI model is instead given only the competitor's name, URL, and your notes, and is instructed to keep its output general rather than inventing specific details it did not actually observe.

4. How AI Output Is Produced

The process is fully automated, without a ClueDig staff member reviewing the output before it reaches you:

  1. We fetch the publicly available content of the competitor URL you submitted.
  2. That content is stripped of HTML markup and reduced to plain text, up to a fixed length limit.
  3. We send that text, together with the inputs described in Section 3, to the AI model with a fixed set of instructions we author (a "system prompt") that defines the required output fields and asks the model to ground its output strictly in the supplied content rather than inventing facts.
  4. For a standard analysis, the model returns six fields: a summary, pricing signals, hiring/job signals, messaging shifts, a single key takeaway, and suggested action steps. For an on-demand battlecard, the model returns five fields: positioning, pricing, strengths, weaknesses, and how-to-win recommendations written as advice to you, not as promotion of ClueDig itself.
  5. That output is parsed, validated for the expected structure, and stored on your account and, where applicable, included in an email or PDF sent to you.

5. Limitations and Accuracy Disclaimer

AI Output is generated by a statistical language model, not by a human analyst, and is inherently probabilistic. Even with instructions designed to keep it grounded in the supplied content, the model may:

  • misread, oversimplify, or omit material information present on the source page;
  • occasionally state something with unwarranted confidence that is inaccurate, outdated, or not actually supported by the source content (a phenomenon sometimes called "hallucination");
  • reflect only what was visible on the fetched page at the moment of analysis, which may not reflect the competitor's current pricing, positioning, or hiring activity by the time you read it; and
  • vary in wording or emphasis between runs, even for the same input, and may change in style or behavior over time as we update our prompts or underlying model.

AI Output does not constitute professional advice of any kind, including legal, financial, investment, tax, or business strategy advice, and must not be treated as a substitute for your own independent judgment or professional consultation. You are solely responsible for independently verifying any AI Output — including any pricing, hiring, or messaging claim about a third party — before relying on it for a business decision, publishing it, or sharing it with others.

6. No Human Review Before Delivery

AI Output is generated and delivered to you automatically. ClueDig does not review, fact-check, or edit each individual analysis or battlecard before it is generated for, or sent to, you. ClueDig is a software provider and does snot assume any responsibility for the responses generated by the AI models we use.

7. Originality and Similarity to Other Content

Language models are trained on large volumes of text and can occasionally produce phrasing that resembles other text found elsewhere, even when instructed to work only from the content we supply. We do not represent or warrant that any AI Output is wholly original or free of unintentional similarity to other sources, and disclaim any such warranty.

8. Data Retention in the AI Pipeline

As described in our Privacy Policy, the raw web page content we fetch to generate an analysis is not stored after processing completes — it exists only transiently to produce the output. Only the structured AI Output itself (the summary, pricing signal, and similar generated fields, or a battlecard's fields) is saved to your account.

9. Your Responsibilities When Using AI Output

You agree that you will:

  • independently verify any factual claim in AI Output before relying on it, republishing it, or using it to make representations to others (including if you enable a Shared Report);
  • not present AI Output as a definitive, human-verified fact about a competitor without your own diligence;
  • not use AI Output to make defamatory, harassing, or knowingly false statements about any individual or business; and
  • only submit competitor URLs you have a lawful basis to monitor, as described in Section 6 of our Terms of Use.

10. Updates to Our AI Systems

We may update the AI model we use, the instructions we give it, or the fields and format of AI Output at any time, without prior notice, to improve quality, cost, speed, or reliability. Such changes may cause the tone, structure, or level of detail of future AI Output to differ from output you have previously received. We will update this Notice to reflect material changes to how AI Output is produced.

11. No Opt-Out of Core AI Processing

Generating AI Output from the competitor content you submit is the core function of the Service; there is currently no setting to use ClueDig without it. If you do not want a specific piece of content processed by our AI subprocessor, do not submit that content to the Service. Separately, making a specific analysis publicly viewable via a Shared Report is entirely optional and requires your affirmative action, as described in Section 10 of our Terms of Use.

12. Contact Us

Questions about how ClueDig uses AI may be directed to hello@cluedig.com.