AI Responsibility

Using AI responsibly

We use AI because it makes our work better and less expensive for clients. It also runs on real electricity and real water. Here’s how we keep our use in proportion, and what we’re doing about the footprint.

Last updated: October 23, 2026

Why this page exists

AI runs in data centers that draw a lot of power and use water for cooling, plus the water used to generate that power. The International Energy Agency estimates data centers used about 415 TWh of electricity in 2024, around 1.5% of the world’s total, and that AI is the fastest-growing part of it (IEA, Energy and AI). Estimates of the energy and water per AI request vary widely, so we don’t quote a single number. The direction is clear enough to act on.

How we use AI sparingly

We treat AI like any other expensive resource: use it where it earns its place.

  • The smallest model that does the job. Routine work such as tagging, classification and short summaries runs on small, fast models. The large models are for the hard reasoning that actually needs them.
  • AI only where it saves real work. If a plain script, a search index or a person does the job just as well, we use that. Sometimes the boring fix beats AI, and we say so.
  • No wasted runs. We cache results, batch jobs and avoid re-running the same request. Our free site scan is limited to three scans per visitor per day.
  • People approve what matters. The agents we build do specific jobs, and a person approves their output before it reaches a customer or an ERP. Agents never approve their own work.
  • We watch the meter. We monitor our own AI usage and cost every day, so waste shows up quickly.
  • Your data stays yours. We use AI providers’ business services, which do not train on customer data by default, and we never use client data to train models.

Offsetting our footprint

We’re starting a program to offset the carbon and the water behind our AI use. We’ll estimate that use from our provider bills and the energy figures providers publish, offset it through verified carbon projects and water restoration programs, and publish what we offset on this page, with the numbers, once the first round is done.

Some history, because it matters: in earlier years we planted trees to offset our footprint. We didn’t keep good enough records to put a number on it, so we’re not claiming one. When carbon pricing stalled, the offset market around it went quiet, and so did our program. This time we’re keeping the receipts.

Where we stand on superintelligence

We build narrow, supervised agents, not anything close to superintelligence. Our view on where AI is heading, and why the human approval step matters more as models get more capable, is on our AI page. The terms are defined in our glossary.

Questions

If you want to know how we use AI on your project, or how we’d keep its footprint down, ask us: [email protected] or 877.609.9029.