Glossary: AI words in plain English
What do these AI words mean for a small business?
They describe who does a job and how much choice it has. An AI agent can take steps on its own. An AI receptionist is an agent that answers calls and books work. Automation is a fixed set of steps that runs on its own. Lead scoring ranks new leads so the best ones get a reply first. Each term below links to real cases we have published.
AI agent
Software that can do a job in steps, not just answer a question. It reads the information it needs. It picks the next step. It acts: it can send a message, book a meeting, or update a record. A person sets the goal and the rules. The agent does the steps.
In our cases: a small agency uses five agents that scan for prospects, draft outreach, classify replies, and book meetings; a construction owner's agent handles the office work behind jobs, such as bidding and billing.
AI receptionist
An AI agent set up to answer the phone or a message line for a business. It answers common questions from the business's own information. It checks the diary or calendar. It books or moves appointments. It takes a message when it cannot help. It can work after hours, so a caller does not hit voicemail.
In our cases: a salon agent answers questions from the services and price list, checks the diary, and books appointments; an HVAC firm's agent answers callers around the clock, asks what they need, checks for real open slots, and books their pick.
Automation
A fixed set of steps that runs on its own when a trigger fires. A person builds the steps once. After that, the same steps run each time, in the same order, with no choices in between. If a step fails or the data does not match, a good setup sends the item to a person to check.
In our cases: when a purchase order is approved in one tool, a set of steps creates the matching order and bill in the accounting tool; supplier invoices are read, checked, and filed, and any invoice with a problem goes to a review queue for a person.
Lead scoring
A way to rank new leads. Each lead gets a score or a rank based on rules the business sets, such as the job type, the place, or how soon they want work. High leads get a fast reply. Low leads get a slow reply or a polite no. In many of our cases, the same job is called qualifying: an agent asks a few set questions, then sorts the lead.
In our cases: a roofing contractor's agent asks four set questions by text and books a site visit for real leads; a sales builder describes scoring agents that sit next to a research agent in the same pipeline; a Make setup with AI modules qualifies new leads as they arrive.
More: Answers to owner questions · Stats from all our cases · Cases by topic
5 use cases every morning
Get 5 like these every morning, each with its source and a prompt you can copy. One email, 6:30am.