Suprava Sabat@suprava-sabat-saasleadgen
Opens with an observation
Write from this board Add posts Share
Loading the board
4 posts · curated by mailnamnv
Ian Tracey@ian_dot_so
I resigned from @tryramp today. I spent the last year there doing applied AI work and neither CEO is acting responsibly. They are racing toward self-closing books with no clear plan for what happens when companies have saved too much time and money. More thoughts below.
Opens with an observation
Stanislav Beliaev@stasbel
This guy from Reddit built a job search system with Claude Code, ran 740+ job postings through it, and landed a Head of Applied AI role 🤯
Then he open sourced the whole thing. MIT, no paid tier.
Santiago from Spain got tired of rewriting his CV for every listing, so he turned his terminal into a job search command center.
You paste a job URL. Claude Code reads a 𝗖𝗟𝗔𝗨𝗗𝗘.𝗺𝗱 with 14 skill modes, pulls the posting, reads it against your CV, and scores the role from 1 to 5.
It reasons through the gap instead of matching keywords, so the report shows where you match, where you don't, and whether the listing is even legitimate.
That score gates everything after it. Under 4.0 it tells you to skip. Above 4.0 it goes to work:
- writes an ATS-optimized PDF resume for that listing
- preps STAR interview stories that turn into 5-10 master ones
- finds the right hiring manager and drafts a short intro message
- auto-fills the application form
- tracks your pipeline in one dashboard
- scans 100+ company career pages to find new roles
740+ postings evaluated with it → 12 first interviews → 1 offer.
Most job search automation optimizes for volume and makes you look like a bot to whoever opens the inbox. This one spends compute deciding where not to apply.
It never submits anything without your approve. You make the final call.
Link to the repo: lnkd.in
--
♻️ Find it useful? Leave a reaction and share with your network!
Opens with an observation
Michel Lieben 🧠@michel-lieben
If I had to generate leads with outbound ASAP, I'd start with these 5 workflows:
1. Run waterfall enrichment on existing leads
Simple, but extremely efficient.
If you find 80% of your leads with multiple providers vs. only enriching 40%, you're doubling the leads you can generate.
→ Use a provider like FullEnrich or build your own waterfall with solutions such as Prospeo, GetLeads, LeadMagic & Apollo.io.
2. Map full TAM & enrich with signals
Build the most exhaustive list of potential clients possible.
And then layer signals on top, to find out which one to priorize.
The key here is to mix in:
- first-party signals (i.e: leads already within your ecosystem)
- secondary signals (i.e: leads interracting with common connections & partners) and
- third-party signals (i.e: publicly available signals that showcase buying intent)
The closest to you, the higher the intent.
The more matching signals, the higher the intent.
Accounts that showcase several buying signals, and who already know you + partners, should be contacted ASAP.
→ Use your own call transcripts (via Claap) & internal data for 1st party signals.
→ Use PredictLeads or Explorium for 3rd party signals.
3. Warm social re-engagement
> Extract engagement from successful publications (better if they're yours)
> Tier, qualify & segments engagers/likers/commenters/profile viewers
> Re-engage tier 1 leads via several channels & others via email.
→ Use Limadata or Trigify.io to monitor social signals.
→ Use lemlist for multichannel outreach or Instantly.ai for email outreach.
4. Champion tracking
> Look at past customers in your CRM (I recommend folk)
> Monitor the ones who switched jobs recently.
> Reach out on their new email address.
They're usually great leads if they were happy with your product at their past company.
5. De-anonymized website visitors
> Use Knock2.ai, Warm AI to de-anonymize website visitors
> Enrich & filter the leads and identify decision makers
> Reach out w/ the context of the page they visited.
P.S: You can build all these plays via ColdIQ's MCP with 'only' a few natural language prompts. Check us out :)
Opens with an observation