- Manual outreach on 100 messages: base collection 40 min + filtering 15 min + sending 2 hours + replies 1 hour = ~4 hours.
- Automated outreach on the same 100: parsing 3 min + sending (async) 8 min + AI triage of first replies = ~12 minutes active time.
- Difference: 20× in time efficiency. Not 2×, not 5×. Twenty.
- Conversion meanwhile is NOT lower. With proper setup (spintax + LLM + cadence) — higher than manual, because a script doesn't get tired.
Example of a real time log from a media buyer working by hand. We'll hide the name — we agreed to publish metrics without identification.
4 hours on 100 messages: step by step
| Step | Time | What's done |
|---|---|---|
| Pick a source channel | 15 min | Scroll recommendations, search by keyword |
| Collect base | 40 min | Open channel, scroll members, copy @usernames |
| Filter | 15 min | In Excel: remove bots, inactive, duplicates |
| Template + personalization | 20 min | Write, test on 3 examples |
| Send 100 msg | 2 h | ~1 min per message (copy→paste→name→send→wait) |
| First replies | 1 h | React to 15-30 replies in real time |
Total: ~4 hours. At a self-value rate of $30-50/hour, that's $120-200 worth of your time for ONE batch of 100 messages.
12 minutes on the same 100 messages
| Step | Active time | Wall time |
|---|---|---|
| Channel search in index | 2 min | 2 min |
| Filter setup | 1 min | 1 min |
| Base parsing | 0 min (async) | 3-5 min |
| Pick / tune template | 3 min | 3 min |
| Launch sending | 30 sec | 2-6 hours (distributed) |
| Monitoring + AI triage | 5 min (over the day) | async |
| Active time total | ~12 min | 2-6 hours distributed |
Critical distinction: wall-clock time (when sends actually happen) can be longer (2-6 hours for human-like cadence), but active time (when you personally do something) — 12 minutes. While the script runs, you handle other things.
Myth: "by hand is warmer"
A common argument: "a live human writes better than a script". Let's break it down — in detail.
"Personalization"
Manual: you look at the profile, find a hook, mention it. 1-2 minutes per message. High quality.
Scripted: LLM looks at channel context, last posts, name. 200-500ms per message. High quality for 70-80% of cases.
Difference: a human is better on edge cases (unusual context, requires creativity). A script — on medium cases (standard outreach in a familiar niche), but thanks to scale wins in absolute terms. 1 very good message vs 10 good ones — 10 usually win.
"Real-time replies"
Recipient replied → you see it → you answer within 2 minutes. Manual preference: you enjoy live dialogues.
Automated: an AI agent (Qualifier) responds instantly and carries the conversation until human strategy is needed. You plug in only at transitions, not at "yes/no/tell me more".
Honestly: where the script loses
- Volume <30 messages per week. Not worth standing up automation for one or two touches.
- Very narrow niche (up to 100 identifiable personas in the world). Better to message each one by hand, individually.
- High-ticket B2B ($50K+ deal size). Here communication quality is ROI. Hands and Zoom.
- An offer that can't be simplified. If an offer needs 5-10 minutes to explain — it's not for cold outreach in principle.
For everything else — 50-5000 messages/week, standard niches, "average ticket" — a script is 10-20× more economical.
TG:ON for macOS · Windows · Linux
Desktop app, 160 MB. Runs locally, your keys stay yours. 3-day trial, no credit card.
Download for free12 minutes instead of 4 hours.
3 days of trial.
Parsing, sending, AI triage — out of the box. Run your own campaign on 100-500 messages and compare against manual flow. The difference is immediately visible in the time logs.
Start trial