--- title: How to run personalised outbound at scale without sounding generic description: Outbound stops sounding generic when the system does the research and drafts in your rep's voice, and the rep approves every send. Here is how to keep relevance at volume. canonical: https://discoveryoutcomes.com/gtm-intelligence/ai-outbound-prospecting-at-scale-personalization --- # How to run personalised outbound at scale without sounding generic You keep outbound personal at volume by splitting the work: **the system does the research and writes the first draft in your rep's voice, and the rep reads it, edits it, and sends it.** Relevance holds because a human still owns the send. Most teams feel forced into one of two bad options: - **Choose volume.** Run a mail merge to 10,000 people. It looks like personalisation, it is not, and reply rates sit near half a percent while your domain takes the damage. - **Choose quality.** Have reps research and hand write 20 emails a day. It works, and it stops the moment the rep has a busy week. There is a third option, and it is not a robot sending mail on your behalf. It is a system that reasons over your market, prepares the research, and hands each rep a draft that is already specific to the account and the person. The rep approves, adjusts, and sends. ### Why volume usually kills relevance Sales engagement platforms are good at delivery. They are not built to work out what to say. The rep is still the one doing the thinking: open LinkedIn, read the news, pick a sequence, choose a snippet, write the one line that makes the email feel real. That one line matters. Reply rates rise sharply when the opening is genuinely about the buyer rather than about you. But asking a person to do that research 60 times a day is how you get burnout, and how the quality quietly drops to a template by Thursday. ### Where current tooling falls short **1. Writing help is not research help.** A tool that speeds up typing does not remove the bottleneck. The rep still has to find the reason to reach out. That is the expensive part. **2. Snippets are still a mail merge.** Swapping a case study by industry is a rule, not a reason. Buyers can tell. It reads impersonal because it is impersonal. **3. Action without intelligence.** Your sequencer knows it is time to send step two. It does not know what changed at the account this week, who owns the problem, or whether now is a bad time. It is disconnected from the context that makes a message land. ### The model that works: research by the system, judgement by the rep Give the platform the account list and the goal. It builds context on each account and each person from the knowledge graph: what you sell, who you have won and lost, what is happening in the account right now, and who sits in the buying group. Then it reasons about fit and why now, and drafts accordingly: - An account that just raised funding gets a message about the pressure that funding creates, not a generic product pitch. - An account running a competitor gets a message about the specific gap your reps win on. - An account with no clear trigger gets a problem led message, or gets held back until there is a reason to write. Every draft is written in the sending rep's voice, using their phrasing and their level of formality, so approving it takes seconds rather than a rewrite. ### How to run it, step by step - **Fix the list first.** Volume on a bad list is the fastest way to sound generic. Filter to accounts that actually fit before anything is drafted. - **Require a reason per account.** If the system cannot state why this account and why now, do not send. That single rule removes most of the noise. - **Pick the person, not the title.** Reach the person living with the problem, and name the others in the buying group so follow ups make sense. - **Draft in the rep's voice.** Store a short voice profile per rep: sentence length, tone, words they never use. - **Keep approval in the loop.** The rep reviews each draft, edits what is off, and sends. Nothing goes out unread. - **Cap daily sending.** Protect deliverability with a sensible per mailbox limit. Relevance and volume both depend on your email actually arriving. - **Feed replies back.** What earned a reply, what got ignored, and what the buyer said becomes context for the next draft. ### What changes for the team Reps stop spending their morning on research tabs and spend it on conversations. Managers get to see the reason behind every send, so coaching is about judgement rather than activity counts. And because a human approves each message, quality does not decay as volume rises. That is the honest version of outbound at scale: more relevant messages, prepared by a system that has the context, sent by a person who takes responsibility for them. ## Common questions Research each account before writing, then draft from a real reason to act rather than a merge field. The system does the reading and the first draft, and the rep shapes and approves it. No. It researches and drafts, and a rep reviews and approves every message. Nothing leaves your domain without a person seeing it. Far more than by hand, because the research and the first draft are already done. The rep spends their time judging and editing instead of digging. ### Want to see this on your accounts? Bring 20 accounts and watch it work in 15 minutes. No setup, no campaign. #### Keep reading