AI Use Policy

Where Haulprint uses AI today

Haulprint’s reports, findings, scores and alerts are produced by fixed rules that we write and test against public records. They are not produced by generative AI models, so the same record always gives the same result, and a finding can be explained by the record behind it.

This policy sets the rules we follow before and after we add any AI feature, so customers know in advance what to expect.

What AI will and won’t do

• AI in Haulprint will assist people, never decide for them. Underwriting, coverage, placement and claims decisions stay with our customers.
• We will not use AI to make decisions about individuals, or to make decisions with legal or similar effects on anyone.
• Content generated by AI will be labelled as such in the product.
• We will tell customers what each AI feature is for, what it is not for, and its known limits. We will never claim an AI feature is error-free.

Customer data and AI

• We do not train AI models on Customer Data, and we use AI providers only on terms that do not let them train on it.
• We send an AI provider only the information a feature needs.
• Each AI provider is listed on our Security page, and customers get at least 30 days’ notice before a new provider handles their data, as our Subscription Agreement requires.
• Where an AI feature would use a customer’s own data, the customer’s administrator can switch it off for the account.

Before an AI feature launches

Each AI feature is recorded in an internal register with its purpose, the data it uses, its provider and a risk level. Before launch we review:

• what happens when it is wrong, and how a user will notice and correct it;
• whether it could treat people or companies unfairly, including through stand-ins for protected characteristics such as names or locations;
• how it performs on real records where we already know the right answer;
• where a person must review its output before anything relies on it.

Features that could affect a decision about a trucking company’s insurance get the closest review, and are released only with a human review step.

After launch

We monitor AI features for errors and drift, act on reports from customers, and fix, restrict or withdraw a feature that does not meet this policy. We keep records of reviews, tests and changes.

Laws we design for

We design AI features to meet applicable law, including state rules on automated decision-making and high-risk AI, consumer protection law and the Fair Credit Reporting Act limits in our Subscription Agreement.

Accountability and contact

Our chief executive is responsible for this policy. We review it at least once a year and whenever we add an AI feature. To ask a question or report a problem with an AI feature, email [email protected].