Conversational AI: 5 tips to go from pitch to product

On Thursday September 19th, Michelle Parayil, a Conversational AI Lead and Course Producer & Academy Lead at CDI, delivered a breakout session titled ‘Conversational AI: 5 tips to go from pitch to product’ at UX India. 

Having designed over 100 virtual assistants along with designing the world’s largest WhatsApp bot, Michelle’s session included practical insights on how to transition a Conversational AI project from concept to reality. 

The areas of focus included how to establish a solid foundation, building a skilled team, overcoming challenges, practical implementation strategies, and measuring success.

We spoke briefly with Michelle after her session to chat about the presentation and to capture a few of the key takeaways for those who couldn’t be in attendance.

CDI: How are you Michelle? How has the audience and experience of UX India been so far?

It's been such a whirlwind of new stories, people, presentations, and ideas for me. Seeing how different designers, companies, and brands are adapting conversational AI and GenAi into their workflows and products has been fascinating. Interacting with such a wide range of professionals from different industries, brands, and roles was also quite interesting.

CDI: One of the aspects you noted in your presentation was understanding the conversational AI Spectrum and the importance of this for organizations before they begin to build their products. How well do you think organizations today are understanding this spectrum to inform their goals and ambitions?

Organizations' grasp of the conversational AI spectrum is mixed. Many still think they need the most advanced AI for everything, but that's not always true. Simpler bots often work great for basic tasks like directing customers or collecting leads.

The good news is, more companies are realizing that not every problem needs cutting-edge AI. They're getting better at matching their AI ambitions with their actual business needs. It's all about finding the right balance between cool tech and practical solutions. This shift in thinking is leading to more successful AI implementations.

CDI: You also covered the 5 key decisions that are central to success. For those that were unable to attend, can you provide an overview of these key decisions and why they are so imperative to success?

Sure. In my talk, I broke down the process of building a solid conversational AI platform into five key decisions. First, you've got to pick your platform. It's a choice between something user-friendly with drag-and-drop features or a more sophisticated, code-heavy option. Really depends on your team's tech skills.

Next up is language. You need to figure out if your bot's going to speak one language, multiple languages, or even handle mixed languages like Spanglish. It's all about matching your audience. Then there's the mode of interaction. Will your bot chat via text, speak out loud, or do both? This should align with how your customers prefer to communicate.

Fourth on the list is deciding on channels. Are you aiming for a bot that works across the board - web, social media, voice assistants, you name it? If so, you'll want to tailor the experience for each platform.

Last but not least, think about integrations. Your bot should play nice with your databases, CRM systems, and APIs. Don't forget to consider how it'll hand off to human agents when needed and how you'll track its performance.

These decisions are crucial because they directly shape how your bot functions and how customers experience it. Get these right, and you're well on your way to a successful conversational AI setup.

CDI: You also covered off some points around implementation and monitoring and managing effectiveness. It feels like these elements are often less considered than other parts of the journey. What was the main takeaway you were hoping your audience would leave with with respect to this part of your presentation?

When it comes to chatbots, launching is just the start. Here's what really matters:

  1. Look beyond bot data - check emails and calls too.
  2. Fix quick, high-impact issues first.
  3. Track key metrics from day one.
  4. Don't panic over early fluctuations - give it time.

 

IMG-20240919-WA0013

The main point? Keep improving based on data, long after launch. It's an ongoing process, not a one-time thing.

CDI: Well congratulations on a great session. Is there anything else exciting happening with yourself and CDI at the moment?

Absolutely. We've got some exciting things happening right now. First off, you can grab my "Maximizing Chatbot Success" course for free if you act quickly, but this offer won't last long. But wait, there's more! 

We're also working on a couple of new courses that I'm really excited about. One is all about Digital Humans - those super realistic AI avatars that are revolutionizing customer service and more. The other course focuses on GenAI Design for RAG, which is generating a lot of buzz in the AI world. It's an exciting time to be involved in conversational AI, and we're right in the thick of it at CDI. We're constantly pushing the envelope to bring you the newest and best. Hope everyone finds these new courses useful!

To learn more about CDI’s courses, you can visit the website, and until 31st of September, take advantage of our 35% Back-to-School discount using the code "Backtoschool-35" at the checkout. 

If your organization would like to learn more about successfully delivering conversational AI project and/or is seeking support, please don’t hesitate to reach out via [email protected]

IMG-20240919-WA0012-1