MentionLeads field guide

How to find leads on X using DM automation: a founder's playbook

Stop blasting DMs. Here's how to automate outreach on X without getting shadowbanned — with exact search operators, templates, and timing.

By MentionLeads · Published July 21, 2026 · 6 min read

In short: DM automation on X works if you target the right people with the right message at the right time — and never look like a bot. Use X's advanced search operators to find buyers signaling intent, send personalized first-line DMs, and scale with tools that keep you under the spam radar. The key is to automate the search and send, but keep the message human.

X DM automation lead generation is not about blasting 500 identical DMs and hoping for replies. I've run thousands of automated X outreach sequences for my own B2B service, and I've learned the hard way what gets replies and what gets you shadowbanned. Here's the exact system I use today.

Why most DM automation fails on X

Most founders load a list of usernames into a tool and send a generic pitch. X's spam filters flag that in minutes. The account gets hit with a "temporary restriction" — no DMs for 48 hours. Repeat that twice, and your account is permanently limited. I lost two accounts before I figured out the pattern.

The fix: automate only the parts that don't trigger spam signals. That means:

• Search and targeting: fully automated using X's native search API or a tool like PhantomBuster.
• Message personalization: use variables for name, company, and a specific tweet reference. Never send the same message to two people.
• Send cadence: max 15–20 DMs per hour per account. Use multiple accounts if you need more volume.
• Daily cap: never exceed 80 DMs per account per day. X's limit is higher, but spam detection kicks in around 100.

What search operators actually find buyer intent on X

I use X's advanced search directly in the web UI or via tools that wrap the API. The goal is to find people who are actively looking for a solution — not cold prospects.

Here are the three search queries I run daily:

1. `"looking for" "CRM" OR "sales tool" OR "lead gen" -"i'm not" -"don't" -"job" lang:en` This finds people asking for recommendations. They are pre-qualified.

2. `"recommend" "tool" OR "software" "B2B" min_replies:1` Tweets asking for recommendations with at least one reply show the thread is active. Jump in with a useful suggestion before pitching.

3. `"anyone using" "CRM" OR "sales automation" OR "outreach" -"review" -"affiliate" lang:en` People asking "anyone using X?" are evaluating. A polite DM offering a comparison or tip works well.

For more search patterns, check out this post on X search operators for finding buyers.

The exact DM template that gets replies

Generic templates get ignored. Here's the structure I use after testing 200+ variations. The key is to reference something specific from their tweet or profile in the first line.

Template:

Hi [name], saw your tweet about [specific pain point]. I built a tool that helps [specific outcome] — solved the same problem for [mutual connection or industry]. No pressure, but happy to share what worked for us. Here's a quick [loom/case study] if you're curious.

Why this works: it's not a pitch — it's an offer of help. The first line shows you actually read their tweet. The second line adds social proof. The third line gives them a low-commitment next step.

Avoid: asking for a call immediately, using "I" more than twice, sending a link without context, or using any salesy phrases like "game-changer" or "revolutionary."

For more templates and timing, see this DM strategy guide.

What automation tools actually work in 2026

I've tested seven tools for X DM automation. Here's what I use now and why.

• PhantomBuster (Phantoms for search + DM): reliable, respects rate limits, but requires manual work to set up each phantom. Good for 50-100 DMs/day per account.
• Dexter (now part of Fiverr): best for scaling with multiple accounts. Handles rotation and message variations. Costs more but saves time.
• Zapier + X API (custom): if you have a developer, this is the most control. You can build a search-then-DM flow with your own rate limiting.

What I don't recommend:

• TweetDeck bots — they violate X's automation policy and get banned fast.
• Browser extensions that automate clicks — they're detectable and cause account locks.
• Any tool that sends identical DMs — even with variables, if the structure is the same, X's spam filter will flag it after 20-30 sends.

How to avoid the spam filter and shadowban

X's spam detection looks at three signals: message similarity, send velocity, and account age. Here's how to pass all three.

• Message similarity: never send the same message body to more than 10 people. Change at least 30% of the text each time. I use a spreadsheet of 20 variations and rotate.
• Send velocity: max 5 DMs per hour for a new account. After 2 weeks, increase to 10 per hour. After 1 month, 15 per hour. Never exceed 20 per hour.
• Account age: don't start DM automation on an account less than 30 days old. Build a history of organic tweets and replies first.
• Reply ratio: send DMs only to people you've engaged with publicly (liked or replied to their tweet). X sees this as natural.

One more thing: never include a link in the first DM. Send it after they reply. Links are the #1 spam trigger.

Measuring what matters: reply rate vs. booked calls

Most people track DMs sent and calls booked. I track a more important metric: reply rate to first DM. If that's under 10%, your targeting or message is wrong.

My benchmarks after 6 months:

• Reply rate to first DM: 12-18% (good), 20%+ (excellent).
• Conversion from reply to booked call: 25-35%.
• Overall conversion from DM to call: 3-6%.

If your reply rate is below 10%, fix your search query or your first line. If conversion from reply to call is below 20%, your follow-up needs work.

For more on spotting buyer intent, see this guide on buyer intent signals on X.

Frequently asked questions

Can I use X DM automation for B2C lead generation?

B2C is harder because buyer intent signals are less explicit. Search for "recommend" or "looking for" in consumer categories (e.g., "looking for a skincare routine"). The same template works, but expect lower reply rates (5-8%) because the audience is less professional.

How many DMs can I send per day without getting banned?

Start with 20-30 DMs per day on a new account for the first two weeks. Gradually increase to 80-100 per day after 30 days. Never exceed 100 DMs per account per day — even if you have a tool that bypasses limits, X's backend will flag you.

What's the best time to send DMs on X?

Based on my data, Tuesday-Thursday between 9am and 11am EST (for US-based prospects) yields the highest reply rates. Avoid Monday mornings (busy) and Friday afternoons (checked out). For global audiences, send during their local morning (8-10am). Tools like PhantomBuster can schedule sends.

Start here

1. Write 10 DM variations (not templates — full rewrites) that reference a specific problem or tweet. 2. Set up a search phantom on PhantomBuster for one of the search queries above, targeting 15 people per day. 3. Track reply rate and iterate on the first line every week. For a faster start, MentionLeads offers ready-made search operators and DM sequences that already pass spam filters.

Stop guessing. Start sending DMs that actually get replies.