Increase the effectiveness of your marketing strategy.
Meet with your marketing and customer service teams and discuss the most common questions and messages they receive.
Take note of any responses your teams find successful.
Outline the customer journey and determine at which stages your customers will engage with your chatbots.
Because of their usefulness, chatbots are appropriate for most marketing activities. These activities include lead generation, onboarding, content distribution, and product adoption.
Identify where you will use your chatbot, as users’ needs and awareness may differ based on the pages or platforms they are visiting.
You will also have varying degrees of functionality depending on whether you use a platform-provided chatbot to engage users or opt for a 3rd party, webpage-oriented experience.
Map the chatbot conversation, using a tool like Draw.io, Miro, or Microsoft Visio to visualize the branching messages on the path to conversion or resolution.
If your goal is to increase transaction volume, you may ask Looking to save 30% on your order?, instead of How can I help you today?
End each chatbot conversation with a CTA. If your playbook promotes a specific product, include a link to get started or learn more.
The easier you make it for the user to take action, the more likely you’ll close the sale or generate a new lead.
For example, a prospect asks a chatbot a pricing question on your product pricing page. After addressing questions specific to the user, the CTA might be: Great, glad I could help clarify the pricing. Would you like to start your free 7-day trial right now? If the user clicks yes, your chatbot might send them to a free trial sign-up page pre-populated with the user’s name and email address.
As part of ending the conversation, consider providing the engaged user with a path to connect with support if they have additional questions. It’s not uncommon for customers to look to your chatbot for answers.
Edit each playbook to account for voice, personality, and conciseness. Ensure each statement, question, or response sounds natural, conversational, and human.
It may be a chatbot, but users still want to feel like they’re engaging with an authentic representation of your brand, your company, and your real-life human teams. However, don’t pretend that the chatbot is a real entity, as the experience can be very frustrating for a customer and also break trust.
It’s also important to be transparent because the user will know that they can get some questions answered but maybe not all. This will allow them to bypass some of the chatbot steps if their situation does require help from a human. To help with this assessment, ask yourself:
- Is this message something I would expect to hear over the phone?
- Are these words that I would use as part of a conversation?
Test your chatbot. As you review your pre-defined chatbot responses, evaluate the ease of consumption.
By the end of your review you should feel comfortable with:
- The screen space each message occupies.
- Message scannability.
- Recommendation relevance and accuracy.
Modify your approach if you see increased bounce rates, or if users are asking questions not included in your existing conversation journeys. It is recommended that you wait 1-2 weeks before assessing the performance of your chatbot.