Fifteen minutes per company. Open the website, find the contact page, copy the address, read the About page, work out what they actually do, and only then start writing an email that does not sound like every other pitch they received that week. Two hundred companies meant fifty hours of that, on repeat. This is how we replaced those fifty hours with AI cold email outreach that still sounds personal.
The problem every agency and SME hits
Cold outreach works when it is personal. Everyone agrees on that part, and it is exactly the trap. For a small business, "personal" means one human doing the research, one company at a time, fifteen minutes at a time. Send two hundred emails, and you have spent fifty hours before writing a single one.
Most agencies and SMEs resolve this in one of two ways. They send generic emails to everyone and get the reply rate generic emails deserve. Or, they keep the personal approach and never reach more than a few dozen companies a week, because that is all the hours allow.
Kappal hit the same wall with our own B2B outreach. No matter how carefully we scheduled the work, one person could not research and write for more than a handful of companies in a day without the quality dropping.
What happened when we automated it
We automated it. The first version sent emails fast, hundreds of them, in the time it used to take to do a handful by hand.
And the reply rate was close to zero.
That silence was the honest result. We had removed the effort but kept the flaw. The emails had the company's name in them, and that was about it. Same opening, same pitch, same closing, with a different name on top. Anyone who reads cold emails all day can tell in the first sentence, and our recipients clearly did. Nobody replied, because nobody believed the email was actually for them.
We spent weeks trying to fix the writing. But the problem was never the writing. It was the research underneath it.
The realization that changed everything
An AI can write a personal email for any company on the internet. But it can only write a good one about a company it actually knows something about.
Feed it a name and a website address and nothing else, and it will produce something that looks personal but reads generic, because that is all the material it was given.
The fix was not a smarter email model. It was research, done automatically, before a single word gets written.
That is what Kappal-Marketing-Engine does. You describe who you want to reach. The system reads each company's website and pulls out the facts that matter: what the business does, who it serves, what it sells, how it talks about itself. Then, the AI writes one email per company from that material. Not a template with a name swapped in, but a uniquely drafted email based on a different set of facts.
How it works, in plain terms
The loop fits on one screen. Set your target market or upload a list of companies. The system goes out and reads their websites (the part that used to eat the fifty hours). It builds a profile of each company and drafts a personalized email for each one. You read the drafts, edit anything you want, and send the batch.
Nothing goes out without your eyes on it first. The system drafts; you decide.
Under the hood, there is a simple pipeline: crawl, extract, draft, send. Each stage is kept deliberately separate so it stays fast and cheap, even when scaling to thousands of companies. But the technical details matter less than the result: what used to take a week of evenings now takes an afternoon.
A real-world scenario: The expo bag full of business cards
To see where this engine really earns its keep, consider the bag of cards you bring home from an industry expo.
Hundreds of direct contacts: marketing heads, sales managers, the people who actually decide. You met them, you shook hands, and the window of opportunity is short. An expo contact expects a follow-up while they still remember the conversation.
The cards are gold, but the time is gone. Collecting a hundred cards takes an afternoon. Writing a hundred personalized follow-up emails takes a week, and by then, the conversation has gone cold.
With the engine, you feed it the cards in whatever form they came: a scan of the pile, a PDF, an Excel export, or a CRM list. Built-in OCR reads the scanned cards, the uploader handles the rest, and the system turns the pile into a clean list of named companies. Then, it reads each website, works out what they do, and drafts a unique follow-up email for each one.
What this looks like in practice
- Bulk-Import Your Leads: Upload an Excel sheet, a PDF, a directory export, or scanned business cards (read via OCR). The system turns whatever you have into a clean company list with zero manual data entry.
- Automated Website Research: Give it a market or a list of companies, and it reads their websites for you. It learns what they do, who they sell to, and how they describe themselves. This is your fifty hours, compressed into minutes.
- One Unique Email Per Company: The AI drafts a different email for each company based on what it learned. The context of their actual business is included, bridging the gap between "mail merge" and "personal."
- You Steer the AI: Different campaigns need different angles. You set the goal, whether it is a specific offer, a warmer tone, a sharp pitch, or a re-engagement push, and the AI adapts.
- Human Approval: Every draft is generated for you to review and edit. Nothing sends itself without your final sign-off.
- Built for Small Teams: If you can describe who you want to reach, the system handles the research and the drafting. No massive marketing department or copywriter on retainer required.
- Controlled Sending: Batches run with rate control so you don't land in spam folders. You can pause, cancel, retry, and track every single email.
The takeaway
Remember the core rule of AI outreach: An AI can only write a good email about a company it actually knows something about.
The fifty hours are gone, but the emails still get written one company at a time, because that is the only way cold outreach earns replies. There is simply no human spending the fifty hours on the research anymore.
About Kappal Software and Kappal-Marketing-Engine
Kappal-Marketing-Engine is an internal product of Kappal Software Private Limited (kappal.in), and it runs our own outreach in production. We built it because the manual approach stopped working at scale, and we now offer it to agencies and SMEs hitting the same wall.
If you sell to other businesses and want cold outreach that doesn't sound cold, write to us through the contact page at kappal.in. Tell us what you sell and who you need to reach. Because the system runs our own outreach, fit assessments are based on production use, not a polished demo.
Kappal Software
Building tomorrow's enterprise solutions