Personalization at scale

What counts as personal

A first name and a company name in the right places is a mail merge. A message is personal when it says something true about that person now: a new role, a post they wrote, a hire on their team, a tool their company runs. At scale means every row of a list gets that treatment without someone researching each person by hand.

The research comes from data, the first draft from a prompt over that data, and a person decides the rules and checks a sample before anything sends.

Levels of personalization

• Merge fields. Name, company and title. Needed, and nobody reads them as personal.
• Segment lines. One line per industry or role, written once and reused for the whole group.
• Row-level lines. A first line written for each person from the facts in their row.
• Timing from signals. The message goes out when something changes, such as a new job, and says why you are writing now.

How to personalize at scale

1. Put the facts in the row. Enrich each person and company with the fields your line will use before you write anything.
2. Score the fit first. A weak fit gets no line and no email, which keeps the effort on the people who fit.
3. Write the prompt like a brief. Name the fields to use, the length of the line and what it must never claim.
4. Read a sample. Run the prompt on a few rows, fix it, then run the rest.

Data from public sources is still personal data. The ICO says the UK GDPR applies when you use publicly available details of a named person for B2B marketing. Unless an exception applies, people whose data you got from someone other than them must get privacy information within a reasonable period, one month at the latest.1 When you use the data to contact them, Article 14 of the UK GDPR sets the latest point as your first message to them.2

In GetSales the table does this row by row. Type / in a prompt to use the person, their company or another column's answer. An AI score says how well each person fits. Set the first line's column to run only on a good fit, and each one who fits gets an opening line written from their own row while a weak fit gets none. Auto-run starts off, so a score or a first line waits until you run it, and you can try a column on one row first.

Related

• Data enrichment: work emails, a fit score and a first line for each fit, in one table.
• Buying signals: what a person or company does that tells you when to write.
• AI sales assistant: Creator plans the search, the work emails, the first lines and the steps from one sentence.
• Cold email outreach: where the first line goes, and the follow-ups after it.

Sources

1. Information Commissioner's Office, Business-to-business marketing, dated 22 May 2025. Read 28 October 2026.
2. legislation.gov.uk, United Kingdom General Data Protection Regulation, Article 14, Regulation (EU) 2016/679 as it applies in the UK, the version in force since 30 September 2026. Read 28 October 2026.