DEV Community: Mohamed Kazaz

An Automated Cold Email Generator That Uses Real Website Data

Mohamed Kazaz — Tue, 27 Oct 2026 10:32:20 +0000

Most AI cold emails fail for the same reason: the model has nothing real to say about the recipient, so it invents flattery. The fix is simple. Feed the generator facts from the company's own website and forbid it from making things up.
What makes a cold email feel personal
A good first line shows you looked. Not "I love your website" but a specific, checkable observation: what the company focuses on, which platform it runs on, or a gap you can see from the outside. Specific beats flattering every time.
The data you need
For each prospect, collect a short company description, the top two or three topics from the site, the technologies it uses, and one opportunity signal such as no CRM detected. That is enough for two or three sentences of genuine personalization.
A template that works
Keep it under 110 words. One observation, one relevant pain point, one sentence about what you do, and a low-friction question at the end. No attachments, no long pitch.
Automating the data collection
Disclosure: I built the AI B2B Lead Intelligence Actor for this. You give it a URL and it returns the keywords, tech stack, contact details, opportunity signals and a draft email with a subject line. By default the email comes from a template filled with the extracted signals. You can optionally add your own Anthropic API key to have an LLM rewrite it, with the instruction to use only the extracted facts.
Rules for using generated emails responsibly
• Read every draft before sending. Automation should prepare the email, not replace your judgment.
• Do not invent claims. If the data does not say it, the email should not either.
• Verify addresses and honor opt-outs. Comply with CAN-SPAM, GDPR and local rules.
• Send small batches. Low volume with relevance protects your domain reputation.
Measure what matters
Track reply rate, not open rate. Test two subject lines against each other across the same segment and keep the winner. If replies are low, the problem is usually targeting, not wording.
Bottom line

An automated cold email generator is only as good as the data behind it. Start from real website signals, keep a human in the loop, and your outreach will read like research rather than spam.

Disclosure: I built the AI B2B Lead Intelligence Actor on Apify. It generates these emails from real extracted data. You can try it for free here: https://apify.com/kazaz_organization/my-actor-1

How to Find the Tech Stack of Any Website (and Why Sales Teams Care)

Mohamed Kazaz — Tue, 27 Oct 2026 10:30:50 +0000

What a company's website runs on is one of the most useful pieces of sales intelligence you can get for free. A business on WordPress with no marketing automation needs a different pitch from a Shopify store already using Klaviyo. Here is how to find the tech stack of a website, manually and automatically.
Manual methods
Right-click a page and choose View Source, then search for telltale strings. Paths like /wp-content/ point to WordPress. A script from cdn.shopify.com means Shopify. js.hs-scripts.com means HubSpot tracking. Response headers can reveal hosting, such as Cloudflare or Vercel. Browser extensions automate part of this, one site at a time.
Why it matters for outreach
Tech signals turn generic cold emails into relevant ones.
• No CRM or marketing automation detected: leads from the site may be handled manually.
• No analytics tag: the company may not know which pages drive enquiries.
• Heavy WordPress plugin use: page speed and maintenance are common pain points.
• Already on HubSpot: you can talk about getting more from existing tools instead of replacing them.
Doing it at scale
Checking 10 sites by hand is fine. Checking 500 is not. An Actor that analyzes a list of URLs can return the detected CMS, e-commerce platform, CRM, analytics, chat widgets and hosting for each one in a single run.
Disclosure: I built the AI B2B Lead Intelligence Actor, which does this alongside keyword extraction and contact discovery. It detects what is visible in the page and its headers, using signature matching across roughly 40 technologies.
Limits to know
Detection only sees what the page exposes. Server-side tools and anything loaded after heavy JavaScript rendering can be missed. A missing tag does not prove the company lacks the tool, only that you could not see it from the outside. Phrase your outreach accordingly, for example: "From the outside, I could not spot..."
A simple workflow

Build a target list of 100 companies in one niche.
Run a tech-stack analysis on all of them.
Filter to companies with a specific gap, such as no CRM tag.
Write one tailored message per segment instead of one generic blast. Segmenting by stack typically beats segmenting by company size alone, because it tells you what the company cares about right now. --- Disclosure: I built the AI B2B Lead Intelligence Actor on Apify. You can try it for free here: https://apify.com/kazaz_organization/my-actor-1

How to Extract B2B Leads Without Coding

Mohamed Kazaz — Tue, 27 Oct 2026 10:26:44 +0000

Every sales team knows the routine: open a company website, hunt for a contact page, copy an email into a spreadsheet, repeat. Doing that for 20 companies is tedious. Doing it for 500 is a lost week. The good news is that you can extract B2B leads without coding, and you do not need to build a scraper to do it.
What you actually need from a company website
For outbound prospecting, most teams need four things: a contact email, a phone number, a LinkedIn or social profile, and a reason to reach out. Most of this is published openly on the company's own site, usually on the homepage footer or the contact and about pages.
The no-code way
Apify hosts ready-made tools called Actors that you run from a browser form. You paste a list of URLs, press Start, and download the results as CSV, Excel or JSON. No Python, no servers, no proxies to configure.
Disclosure: I built one of these, the B2B Lead Enricher, so I will use it as the example. The steps are the same for any contact-finding Actor.

Open the Actor page and click Try for free.
Paste your list of company websites into the input.
Click Start and wait a few seconds per site.
Open the Dataset tab and export to CSV. The output includes emails, phone numbers and social links found on the site, plus the contact and about pages, which is where most direct emails live. Plugging it into your workflow If you use Google Sheets, Make, Zapier or n8n, you can run the Actor automatically whenever a new row appears in a sheet. That turns a manual research task into a background job. What this will not do Be realistic about limits. Only publicly listed contact details can be found. Some sites block automated access, and large enterprise sites often do. A generic address like info@ is common, so treat results as a starting point and verify emails before any bulk send. Stay on the right side of the rules Finding a business contact is not the same as having permission to email them. Follow CAN-SPAM, GDPR and local laws, include an unsubscribe option, and keep outreach relevant and low-volume. Bottom line If your time is worth more than a few dollars per thousand lookups, automating contact discovery is one of the cheapest productivity wins in sales. Try it on 10 companies first and compare the result with your manual process. --- Disclosure: I built the B2B Lead Enricher Actor on Apify. You can try it for free here: https://apify.com/kazaz_organization/my-actor