Foresight LabsA jacket, two journeys

Read the regulation behind each answer

Foresight’s AI works from the regulator’s own documents. Each alert links to the documents it summarises, and each fact the Assistant marks opens to the words it quotes. Your experts decide what to rely on.

Illustrative. The law and its words are real.

Check it before you rely on it

Foresight’s agents write and score the alerts, and the Assistant answers your questions. Each shows where it came from, so your team can check it.

  • Open each alert’s source documents

    Each alert links to the documents it summarises, as the regulator published them. Open one to check the alert’s dates and thresholds against the full text.

  • Check an Assistant fact against the words it quotes

    Open a marked fact, and the Evidence Lens shows the passage behind it and why it supports the claim. Conclusions and next steps carry no mark, because they are the Assistant’s own reasoning.

  • See why an alert scored High for you

    Each alert’s scoring details show how it rated on topic, substances, impact, geography and urgency for your organisation. The score judges relevance. Whether the law applies is your experts’ call.

  • When the source is unclear, the agent stops

    When an agent can’t confirm a fact from its source, it stops there and passes the item on. Foresight’s team publishes alerts and resolves the items its agents pass to them.

  • Tell Foresight when a score is wrong

    Answer “Is this alert relevant to you?” on any alert, rate how relevant and impactful it was, and add a note. Your team’s answers help Foresight tune how it scores alerts for your organisation.

  • Foresight doesn’t train AI on your data

    Foresight trains no AI models. The Assistant sends your questions and files only to AI providers that don’t use them for training.

What your reviewer will ask

Where it came from
Each alert links to its source documents, and each marked Assistant fact opens to the passage it quotes.
Why it matters to us
Each alert’s score shows the reasons it rated High, Medium or Low for your organisation.
Who decided
Foresight’s agents write and score the alerts, and its team publishes them. Your experts decide what to act on.
What happens to our data
All data is encrypted at rest and in transit. SOC 2 Type I is complete, and SOC 2 Type II is in progress.
Read how Foresight protects your data

Questions about Foresight’s AI

Does a person check each alert?

No. Foresight’s agents write each alert from its source documents and score it for your organisation. Foresight’s team publishes alerts and resolves the items its agents pass to them, such as a fact an agent couldn’t confirm from its source.

What does a High score mean?

The alert scored 70 or more for your organisation. Medium is 40 to 69, and Low is below 40. The score measures how relevant the alert is to you, and its scoring details show why.

Where does the Assistant’s evidence come from?

From what it reads: your alerts and their source documents, and official sources on the web. Open a marked fact to see the excerpt, why it supports the claim and a link to the source.

Are the Assistant’s answers checked?

Not by an expert. Open the marked facts before you rely on an answer. When the work has to stand up to scrutiny, Foresight runs it as a project and checks every result against its source.

What if an alert is wrong?

Tell Foresight. If the score is off, answer “Is this alert relevant to you?” on the alert. If a fact is wrong, contact Foresight, and its team looks into it.

Bring an alert your team had to check by hand

In your demo, trace Foresight’s version of a change your team checked back to its source.