Using AI to Find Your Best Prospects

Last week, I was talking with an entrepreneur who shared how he was using AI to do a much better job of identifying which prospects he should pursue.

It is easy to get a list of names and companies. It is much harder to determine which ones are the best fit for your product or solution.

This reminded me of one of our projects at Pardot 15 years ago. With a thousand customers, we knew there were certain characteristics that made a company much more likely to be a good fit for our product.

Here were the top three.

First, did the company advertise its own product through Google AdWords?

If so, that meant the company was already spending money on online marketing and direct response demand generation. Online advertising was a strong signal that the company could benefit from marketing automation.

Second, did the company have sales representatives listed on LinkedIn?

A sales team usually implied a more consultative and complex sales process involving multiple steps and touch points. Marketing automation was particularly valuable for companies with this type of sales motion.

Third, did the company’s website include a newsletter signup box or another form of email interaction?

This suggested that the company was already using email marketing, landing pages, automated responses, triggers, or similar tools. It demonstrated a basic level of marketing sophistication. Marketing automation could take those existing efforts and make them much more effective.

If a company was running Google Search Ads, employing sales representatives, and collecting email addresses on its website, the probability that it was a good fit for Pardot went through the roof. It became an obvious prospect.

What did we do next?

We built our own internal software to identify new companies and evaluate lists of potential prospects.

For example, the software would collect links from every article published on sites such as TechCrunch and TechMeme and add them to a database. It would then identify the companies mentioned in those articles, review Google Search results, analyze LinkedIn, and crawl each company’s website. It could even inspect the website’s source code for commonly used email marketing tools.

Finally, the system would produce a prioritized list of companies and make it easy to synchronize those prospects with Pardot.

We spent a good amount of time on this side project, and it was worth every penny.

Today, AI and large language models make this type of work dramatically faster and easier. They also make it possible to perform a much deeper level of analysis on every potential customer.

What words or ideas on a prospect’s website indicate that the company might be a strong fit for your product?

What broader concepts or signals can be identified through social media, job postings, press releases, customer reviews, or other public information?

What additional data points can AI triangulate at scale in an automated way that would have been impractical or impossible before?

My recommendation to entrepreneurs is simple. Study what your best customers have in common. Identify the characteristics shared by your ideal prospects. Then build tools and systems that continually find, evaluate, filter, and prioritize companies based on those characteristics.

You should never cold call or cold email a company again unless it has first been evaluated using your own set of characteristics.