Customer email and AI drafting glossary

Email support

Canned response: A canned response, also called a reply template or saved reply, is prewritten text you paste into an email to answer a common question.
Email triage: Email triage is sorting incoming email before anyone answers it: deciding what needs a reply, what can wait, what goes to someone else and what needs no action at all.
Escalation: Escalation is handing a customer question to someone with more knowledge or authority when the person answering cannot resolve it.
Help center: A help center is a public website of articles that answer customer questions: how to use a product, shipping and return policies, billing, troubleshooting.
Knowledge base: A knowledge base is a collection of information a business keeps so answers can be looked up instead of remembered: policies, product details, how-to steps, known problems and their fixes.
Repeat question: A repeat question is a customer question your business has already answered many times, such as "where is my order", "can I change my address" or "how do I cancel".
Shared mailbox: A shared mailbox is one email address, such as support@ or hello@, that several people can read and answer from.

AI email reply tool: An AI email reply tool reads an incoming email and writes a suggested reply, usually as a draft in your own inbox for you to check, edit and send.
Auto-send: Auto-send is a setting in some AI email tools that sends the AI-written reply to the customer without a person reviewing it first.
Grounding: Grounding means an AI system bases its answer on specific source material, such as your help articles, website pages, past replies or records, instead of on what it absorbed during training.
Hallucination: An AI hallucination is a confident statement from an AI model that is false or made up: a policy that does not exist, a wrong price, an invented feature or a link that leads nowhere.
Human in the loop: Human in the loop means a person reviews and approves what an AI system produces before it takes effect.
Retrieval-augmented generation (RAG): Retrieval-augmented generation, or RAG, is a way of building AI answers in two steps: look up the passages in your own documents that match the question, then write the answer from them.
Tone of voice: Tone of voice is how a business sounds when it writes: formal or casual, brief or chatty, how it greets and signs off, whether it uses first names or jokes.

Coverage: Coverage is the share of incoming customer email that an AI tool handles at all, for example by writing a draft or an answer, out of all the email that needed a reply.
Edit rate: Edit rate is the share of AI-written drafts that a person changed before sending.
First response time: First response time is how long a customer waits between sending an email and getting the first reply from a person at your business.
Resolution rate: A resolution rate is the share of customer conversations a support tool counts as resolved, often without a person getting involved.
Sent-unchanged rate: The sent-unchanged rate is the share of AI-written drafts sent to customers exactly as written, without a single edit.
Time saved: Time saved is an estimate of the writing time a team did not spend because replies were drafted for them.

Audit trail: An audit trail is a record of what happened, when and by whom or what: which email arrived, which sources were read, which draft was written, who edited or sent it.
Data processing agreement (DPA): A data processing agreement, or DPA, is the contract between a business and a service provider that handles personal data on its behalf.
Data residency: Data residency is the country or region where your data is physically stored and processed, for example in data centres inside the EU.
GDPR: The GDPR, the European Union's General Data Protection Regulation, is the law that governs how organisations collect, use, store and share personal data about people in the EU.
No-training clause: A no-training clause is a contract term in which a provider promises not to use your data to train or improve its AI models.
Sub-processor: A sub-processor is a company that a service provider uses to process your data on its behalf, such as a hosting provider, an AI model provider or an email delivery service.

Check note: A Check note is the short internal note ReplyPen places with each Draft, listing what it relied on: the help articles, website pages and past emails behind the reply.
Data source: In ReplyPen, a data source is anything ReplyPen reads to write a Draft.
Draft: In ReplyPen, a Draft is the ready-to-send reply ReplyPen writes for a customer email that needs an answer.
Memory: Memory is where ReplyPen keeps what it has learned about your business, your tone and your answers, written in plain language you can read and correct.
Rule: In ReplyPen, a Rule decides whether a kind of email gets a Draft.