Build your first chatbot for lead generation
Build your first chatbot for lead generation
1. Select a messaging channel that your chatbot will live on, based on the messaging apps your target audience already uses.
To generate leads, implement your chatbot where your audience exists, instead of trying to get your audience to move to a new channel. Channels include: Your standalone website Facebook Messenger Slack Kik Twitter You can also opt for a multi-channel strategy. For example, cosmetics company Sephora’s chatbot on Facebook Messenger is focused on booking store reservations, getting help with makeup color matching, and other customer service-focused needs. In contrast, their Kik channel is focused on finding specific products.
2. Choose a chatbot platform to build your first lead-generating chatbot, or take a DIY approach if you have the resources to code it.
If you have the coding expertise, you can build your own chatbot. Some messaging channels, such as Facebook Messenger and Slack, also provide resources for building a chatbot natively using their own tools. However, most people opt for a done-for-you, low-coding approach using chatbot building platforms like: Motion AI Chatfuel Google DialogFlow Botsify Beep Boop Bot Kit Octane.ai SAP Conversational AI. If you chose a multi-channel approach earlier, use the same chatbot platform across channels instead of using a different chatbot platform for each channel.
3. In your chatbot settings, select display and audience targeting parameters. These will control who sees the chatbot option and when or where.
Include: Web pages you would like your chatbot to appear on. Characteristics of the audiences you want to target. IP address exclusions. The point at which the chatbot should appear on the chosen page or channel.
4. Identify a lead generation goal that you want the user to achieve when chatting with your chatbot.
You will use this goal to design the conversation flow and measure success of the chatbot. For example, you might want the user to: Provide their name and email address. Opt in to a lead funnel. Download a specific PDF, white paper, case study or similar resource. Start a free trial.
5. Interview your sales team and review your target persona and customer journey to find common user questions and objections.
While your chatbot’s AI can learn over time, it must be prepared for the most commonly encountered questions and concerns that your existing leads have. Interviews with your sales team can identify the types of queries they receive on social media, emails and the phone, while your customer journey can shed more light on wider customer needs and questions at the top of your funnel.
6. Create a new automated conversation flow, starting with a greeting and a question like How can I help? and ending with a call to action related to your lead generation goal.
Most chatbot-building platforms, such as ChatBot and SAP, use a simple drag-and-drop format. You can create conversational branches based on the user’s answers, but try to bring each back to a call to action based on your lead generation goal.
7. Test your chatbot’s flow with internal and external testers. Fix confusing messages or fractured exchanges.
Chatbots are meant to mimic natural human conversation, and that looks and sounds different to everyone. Before deploying your chatbot, do several tests with internal and external testers. Look specifically for situations where the chatbot’s replies don’t align with what the user intended to say or ask.
8. Review your chatbot’s logs and edit or adjust the conversation flow as necessary. Fix errors or issues flagged in your chatbot’s logs.
Look for issues like: Drop offs: If users are exiting the chatbot midway through the conversation, they may have found their answer elsewhere. Or the chatbot might not be providing the information and answers they need as quickly as expected. Poor completion rate: If users get to the final CTA but don’t take action, you may need to add additional information about the benefits or value of the CTA. It may also indicate that the offer is not as valuable to the user as you thought.