What it is
Most 'AI personalization' is mail merge with extra steps. {first_name}, {company}, maybe {industry} — and a generic message body. That's not personalization, and prospects see through it instantly.
Real AI personalization synthesizes multiple sources into a specific, hand-written-feeling opener. Infonet reads four sources per prospect and picks the strongest single signal — not all four crammed in — to anchor the first message.
How it works
Infonet's AI reads four sources per prospect: their LinkedIn profile (current role, tenure, education, summary), their recent posts and comments (what they've said publicly in the last 60 days), their company's recent news (funding, hiring, product launches), and any mutual-connection context.
It then picks the strongest single signal and writes a one-line opener that references it. The body of the message and the CTA are templated; only the opener varies per prospect.
This 'AI synthesis' approach (Tier 3 in our personalization spectrum) outperforms mail merge by 5x on meeting bookings in head-to-head tests. Read the full data comparison for the numbers.
Voice library
You upload 5-10 examples of how you write — past LinkedIn posts, sample emails, anything that captures your tone. The AI uses those as voice anchors.
Practical effect: messages sound like you, not like an AI. Capital letters where you'd capitalize, contractions where you'd use them, length and rhythm matching your samples.
The voice library can be per-rep (each SDR's individual voice), per-team (the team's shared voice based on top performers), or per-client (in agencies, each client's voice).
Quality guardrails
Every AI message goes through three filters before it's allowed to send.
Factual-citation check. The message must reference a verifiable detail, not a hallucination. If the AI invents a 'recent post' that doesn't exist, the message fails the filter.
No-repeat check. The same opener can't be used across two prospects in the same campaign.
Tone-match check. The message's sentiment must match your voice library. Off-tone messages route to human review instead of sending.
Cost and limits
AI personalization is included in every Infonet plan starting at $39/mo per profile. There's no per-message AI fee — you pay once for the seat, the model API costs are absorbed.
Volume cap: 200 AI-personalized messages per profile per day. Above that we recommend supplementing with snippet-personalization templates.
AI quality varies less than you'd expect across volume. The 200th message of the day is as well-personalized as the first.
What it does and doesn't do
Does: Write the first message. Adapt to the prospect's profile and recent activity. Match your voice. Avoid generic openers like 'I noticed your background.'
Doesn't: Auto-respond to inbound replies. Make sales judgment calls. Hand-off to AE. Replace human review on high-value prospects.
AI is the hammer; you're the carpenter. The AI handles repetition; you handle the conversation.