Use Case: Build an AI-Powered Lead Generation System

Hi there,

As we wrap up 2023, I'm reflecting on the various ways I've helped companies grow their businesses with AI.

In particular, the marketing function has emerged as my favorite playground for AI. Even small, interconnected AI projects can supercharge marketing efforts, leading to faster growth and high ROI right out of the gate.

That's why today, I'm bringing you 3 use cases that can be combined into one powerful system as an early Christmas present.

If you're in marketing, keep reading!

Don't Launch AI with a Big Bang, But a Series of Small Sparks

The beauty of AI in marketing is its flexibility. Rather than committing to large, overwhelming projects, we can ignite transformation with a series of manageable "small sparks."

It blends perfectly with my concept of Atomic AI use cases.

These incremental steps not only make AI more accessible, but also allow for agile adaptation and learning.

It's all about building momentum, one smart, AI-driven step at a time, while at the same time making sure you’re compounding your efforts.

Sounds complicated? It’s not.

Let me walk you through the following AI-powered Lead sequence that will help you make more relevant prospects show up at your doorstep.

The Scenario

Generating quality leads is painful and expensive. You probably already know this, so let's not twist the knife too much.

I’ve been working a lot recently for trade show organizers, so let’s consider the following example:

When a anonymous user comes to our website, we want them to leave their contact information if they are relevant leads (e.g. potential sponsors or exhibitors) so we can follow up.

Of course, this applies to other industries as well. If you're a SaaS company, you want people to sign up for a free trial. If you're a software company, you want to get them into a demo. You get the idea...

Note: The following workflow has been simplified to fit in one newsletter, but the general concepts apply. If you like a deep dive on any of these use cases, hit reply!

Use Case 1: Lead Capture With Chatbots

The problem: People don't just "leave" their contact info on a website (or at least very few do). Most will just browse around and disappear.

The solution: We will use an AI-powered chatbot to give our website visitors something "useful", which is tailored directly to them and encourage to leave their contact details for follow-up.

The process

We’ll deploy a chatbot on our event website like this:

Step 1: Put a chatbot on our "end-of-show report" page (potential clients often look up previous events) and let them greet "Do you have a question on our events? Let me help you."

Step 2: Instruct the chatbot to ONLY answer question regarding previous editions of our show based on PDF reports and other data sources we provided.

Step 3: At some point, the chatbot will ask the user if they like a cost estimate and the brochure for participating in next year’s show.

Step 4: If the user agrees, the chatbot will collect the necessary information (stand size, with or without booth construction, how many open sides, etc.) and hand that over to a pricing calculator in the background.

Step 5: The chatbot will provide the cost estimate on the fly!

Step 6: To follow up, the chatbot will collect company and email address to send more info and the pre-filled application form (show organizers don’t send quotes, they send application forms, let that sink in).

Optional: In GDPR-regulated countries, send a double opt-in email to the email address to get consent.

Here’s how the chat convo could have looked like:

The outcome

Bam! We’ve just built an application that is able to walk a anonymous website visitor through the stages of becoming a known, qualified lead who wants a quotation from us - running 24/7 and in all major languages! We’re collecting the leads in our CRM or - to keep it simple in this demo - a Google Sheet form.

Technology used: OpenAI Assistant APIs for the chatbot, a chatbot platform like Voiceflow for building the app, Google sheets API for the pricing calculator.

Use Case 2: Lead Classification

The problem: Some leads are important customers, while others are just competitors trying to find out our prices. We should handle our top clients personally. If someone asks for a 100k exhibition package, they don't want to land in an email sequence but talk to a real person. How do you find these customers? A naive approach would be to handle quotes below a certain size automatically and everything above manually. That’s a good start. But new customers, even big companies, typically start with a small package first, making them fly under the radar. To make sure we don’t overlook them, we need to bring them to the attention of our sales people.

The solution: We use an AI model to classify lead into manual and automated follow up, based on the quote size and the company info.

The process

Based on the company address, we will try to infer whether this is a potential top customer or not. Here’s a simple way to do this with AI.

Step 1: Take the company address from the Google sheet and send it to the AI.

Step 2: Instruct the AI to classify the address into categories like “Small company <10M”, "Medium company 10-500M”, “Large company >500M”, or “Unknown” and put this info into a new column “Company type”. Example:

Step 4: In a new column called "Follow-Up", we use a rule-based classifier to decide if the lead should be followed up manually or automatically. This decision is based on the quotation size and company type. For example, quotes below $5,000 are automatically processed, unless the inquiring company is medium or large in size. In the example below, we use the stand space as a value indicator.

Step 5: Pass all "manual" leads to your sales team and "automated" leads to an email sequence (see next use case)

The outcome

Sharper lead prioritization, saving time and boosting conversion rates.

Technology used: GPT-3.5 API to classify the company address, Google Sheets to store the data.

Use Case 3: Automated Lead Nurturing

The problem: While our sales people are busy chasing the high-ticket leads (rightly so!) we need to cater the low-ticket leads as well.

The solution: Automated email follow-ups.

The process

All leads marked for "automatic" follow-up would be passed on to the next pipeline:

Step 1: An automated (non-AI) service will generate the filled-in application form for the client based on the spreadsheet data.

Step 2: An AI will write the personalized email message, linking back to the previous chat conversation. Here’s a sample ChatGPT prompt you could adapt for your automated workflow:

Step 3: An email service combines the AI-generated email text with the PDF form and marketing brochure and sends it to the customer. Tip: Start by putting these emails in a draft folder so you can keep a human in the loop and double check the process.

Step 4: Automatically mark the leads in the Google Sheet that have been added to the email sequence.

Step 5: A follow-up email sequence will be sent to the client if they don’t respond within X days.

The outcome: We have a fully-automated 24/7 email follow-up engine that’s nurturing low-ticket leads with minimal effort, sometimes converting them into valuable opportunities.

Technology used: A simple PDF service like PDF.co to generate the custom PDFs, an automation tool like Zapier to handle the workflow, and an AI service like GPT-4 to write the emails.

Conclusion

Marketing is a treasure trove for AI.

Instead of big projects, we can use little but effective AI applications. Start with small ideas and watch them grow into something amazing.

I hope today's use cases have inspired you to kick off your next AI project. Again, if you need help - reach out anytime!

Happy holidays and see you next Friday for a short edition!

I'll share my take on what to expect from AI in 2024.

Best,
Tobias