How we verify and test

Short version: every price is checked against the vendor's own pricing page and dated. Hands-on testing (CallBench) is rolling out next, and we only publish scores we can back with evidence.

1. Price verification (live now)

• Every price comes from the vendor's own public pricing page, linked on the tool page.
• Each price shows the date we checked it. The first full check was on 2026-10-13.
• If a vendor doesn't publish prices, we label it "Pricing on request". If we couldn't read the page, we label it "Not verified". We never estimate or guess a price.
• We re-check prices weekly and log changes on each tool's pricing page.

2. CallBench: AI receptionist testing (upcoming)

CallBench is our planned benchmark for AI receptionists. Here's how it works as it rolls out:

• We set up each receptionist for a clearly labeled test company (a fictional HVAC/plumbing business run by TradeTested), configured the way a real shop would.
• We place scripted calls based on real trade situations: no-heat emergencies, burst pipes, storm damage, price shoppers, reschedules, and Spanish-language callers.
• Every call is scored on the same published rubric: answer speed, task success, accuracy of captured details (name, address, issue, urgency), correct escalation, booking success, made-up information (hallucination), and cost per 100 calls.
• We publish the scripts, the rubric, recordings or transcripts, and the limitations of each test.
• We retest each tool at least every 90 days, and vendors can reply to and dispute results with evidence.

3. Field service software

FSM reviews will use a standard trial workflow (create a customer, quote, schedule, dispatch, invoice, collect payment) with timed steps and documented screenshots. Hands-on FSM reviews are rolling out category by category.

Independence

Sponsors can't buy scores or rankings. See our editorial policy and disclosures.