Increase user engagement.
Ensure that your CRM and chatbot are connected. Confirm that you have access to first name, last name, and contact information like email address and phone number.
If you are a SaaS company, useful information may include company name, subscriptions, monthly spend, and products adopted. If you use Marketo, a platform that manages many popular chatbot integrations, you will also need to verify that your contact attributes connect to the chatbot’s known customer and prospect targeting.
Review your playbook targeting parameters to confirm that you aren’t excluding a population you had intended to engage.
If you have a catch-all playbook that engages with both prospects and visitors, targeting parameters should include known customer and known prospect attributes. These attributes can be generated from previous engagement with the chatbot, CRM connections, and Marketo smart lists.
Enhance user targeting with CRM integrations or data-enrichment services that track your users across your websites.
Commonly used data-enrichment services include:
These services can be used to capture business information, SaaS usage, and social media presence. In the case of social media, this information can be effective in determining which social media platforms to prioritize for advertisement.
Create new chatbot playbooks for high intent pages.
For SaaS businesses, these pages may include pricing pages, case studies, get in touch portals, and newsletter subscription forms. For ecommerce businesses, these pages may include a homepage, customer dashboard, and specific category or product pages. You can also create chatbots for high intent customer support pages.
Review the existing chatbot playbook and script, and make adjustments to use natural language.
While your visitors may understand that chatbots are not real humans, using natural language improves user engagement. However, many chatbot integrations and platforms come with pre-written playbooks that sound too formal or technical. Review the playbooks and change formal phrasing to more natural language. For example:
- Also… instead of Additionally…
- Have you thought about… instead of Have you considered…
- One moment please… instead of Please hold…
Adjust each chatbot path and script for your brand voice and appropriate tone.
Consistency across communication channels is important for conversion, user trust, and a coherent brand experience. Similar to the natural language issue, edit your playbook and chatbot responses to have the same voice established by your brand guide and editorial style guide. Finally, make sure that the language is appropriate for the situation.
For example, if you are troubleshooting issues, you may want to adjust chatbot responses to be more solemn and empathetic. But if it’s in response to a user’s recent purchase, your chatbot language should be upbeat and congratulatory.
Replace the default chatbot avatar with a visually distinct image or an employee headshot.
This gives your chatbot a human face or a branded icon. Consult your brand and legal guidelines for information around the usage of color, illustrations, and employee headshots. Make sure to A/B test these avatar changes. The avatar you end up selecting could drive positive or negative results based on the expectations and personalities of your audience.
Translate or localize chatbot language based on region.
Small details show your audience that you’re aligned with their needs and goals. This even applies for an English only audience spread across North America and Europe. Something as simple as accounting for British, American, and Canadian spelling can make a difference with prospects or customers in the region.
Add free form text to your chatbot playbook.
You can specify phrases or keywords to look for as a way to navigate relevant information. Because this more advanced feature requires some guesswork, we recommend using it in combination with multiple-choice functionality.
Create error responses to invalid inputs, and have a path and script for potential errors.
For example, if an email address supplied by the user fails to meet predefined criteria, your chatbot should respond with something along the lines of I didn’t quite get that… or I can’t seem to find your email address on file….