Subscribe | Free Chatbot | CRO Audit | Cool Tools

3 Tests for Your AI Chatbot

AI chatbots have become more complex in the last two years, but with complexity comes inaccuracy. This does not stem from bad technology, but rather from technology being easier to use. In other words, the ceiling has been lowered in such a way that anyone can create a chatbot. However, the lack of understanding of how AI works creates poor results. Here are three areas to think about when you create your AI chatbot.

1. Use Case Accuracy
When a visitor comes to your site, it’s because they have a need or problem to solve. The AI needs to solve each of those use cases. For example, are they finding a product, looking for an answer about their company vacation policy, or want to understand the difference between the services you offer? List these use cases, then test them from the user’s perspective. There is nothing more frustrating than an AI that can’t answer anything.

2. Engagement and Problem Solving
We are all human (for now, at least), and humans like to connect and bond via language. The AI needs to engage with the visitor like a human. Does your AI ask questions, provide a scenario to help problem solve, or provide alternatives for the visitor to consider? Test to make sure the engagement and problem-solving are in alignment with what the visitors are in need of. When it is not, ensure the AI recognizes it is not helping and provides help from a human.

3. Ethics and Legal Considerations
AI law and AI ethics are changing rapidly. Here in Colorado, a new law will be in place by early 2026 about AI discrimination and security. Review your state and federal laws regarding AI and personal data. Test to make sure your AI does not discriminate, has access to confidential data, and does not answer questions that are against FDA regulations or other industry-specific laws. Lastly, test to make sure all data is secure.

Marketing News

• AI Visitors Visit Fewer Pages and Bounce More Often Than Traditional Search Visitors. Google is forever telling us that AI search is of better quality… (Link) (Link)

• My Exact 7-Step Framework for Brand SEO, With Templates. (Link)

• If you struggle to explain your product in a sentence or two, this post is for you (Link)

• How to Future-Proof Your SEO Strategy with Relevance Engineering (Link)

• Writing in the Age of LLMs (Link)

• How to make your AI-generated content sound more human (Link)

• How to build a traffic-first strategy in a fragmented search world (Link)

• Total YouTube Subscribers (Link)

• Decoding crawl frequency: How Googlebot behavior reflects site health (Link)

• How To Build a Lasting Marketing Career: 25+ Tips for Newbies and Pros (Link)

• Report By Product Type -Blogs, Websites, Advertising and Promotions, Social Media, Audio and Video, Others (Link)

• AI Optimization: The Next Generational Scam (Link)

• How Google’s AI Mode Compares to Traditional Search and Other LLMs - AI Mode Study (Link)

AI News

• Checking In on AI and the Big Five (Link)

• Gemini CLI: your open-source AI agent. Brings Gemini directly into developers’ terminals (Link)

• New data highlights the race to build more empathetic language models (Link) (Link)

• Bill seeks to ban Chinese AI from federal agencies, as the US vows to win the AI race (Link)

• AlphaGenome: AI for better understanding the genome (Link)

• The résumé is dying, and AI is holding the smoking gun (Link)

• How to Extract ChatGPT’s Fan-Out Queries (Link)

• Gemini 2.5 for robotics and embodied intelligence (Link)

• The stakes of AI moral status (Link)

• Exclusive: How Claude became an emotional support bot (Link)

• Yes, I will judge you for using AI... (Link)

Cool Tools

• MySite - AI will build a site for you (Link)

• Mood Gallery - turn emotions into art (Link)

• Optimal - AI code review (Link)

• BlurBackground - AI tool to blur (Link)

• Voice Cloning - Free AI voice cloning (Link)

• Magic Patterns - AI prototyping platform (Link)