AiThority Interview with Chris Maeda, Co-Founder & CTO at Botco.ai

Hi, Chris. Welcome to our Interview Series. Please tell us a little bit about your journey in the industry. How did you arrive at the idea of starting Botco.ai?

I was involved with AI in the symbolic AI era of the 1980s until the so-called AI Winter. At that point, I went to grad school for a computer science Ph.D. and then had the good fortune to co-found a successful marketing automation software company (with co-founder Anu Shukla) in the 90s.

Around 2015, I was on a business trip to Hong Kong when someone pulled me aside and showed me WeChat, which — in the Chinese-speaking world — was a replacement for websites, mobile apps, and email marketing. I knew that historically, every time a new customer interaction channel emerged, there was an opportunity for new marketing automation vendors to focus on it. We founded Botco.ai to provide tools for this new conversational marketing channel.

How do you see the global SaaS industry for contact centers evolving in the post-pandemic era?

The contact center industry has always been focused on reducing costs and increasing customer satisfaction by allowing customers to self-service and by making contact center agents more productive. AI chatbots are a natural evolution for contact centers in that they are the ultimate self-service tool, which allows a much wider range of use cases to be supported with no involvement from human agents.

Recommended: AiThority Interview with Andy Pitre, Executive Vice President of Product at HubSpot

What are the perennial challenges that product developers have to face in the fast-moving marketplace for conversational automation?

This market is highly fragmented, both horizontally by technology, and vertically by data and know-how. Product developers must carefully make horizontal build-vs-buy decisions while also carefully planning their path through the vertical segments.

To illustrate, building foundational language models is clearly a bad horizontal decision today, but this wasn’t always the case. Vertically, product developers must narrow their focus to amass data and win markets, but cannot become overfitted to a specific vertical unless they are willing to limit their addressable market. The fragmented market means that vendors have to negotiate many more tradeoffs.

What are your core offerings? Which industries use Botco.ai for their business?

Botco.ai serves a number of industries, including health and wellness, senior living, and governments at the local and state level. The Botco.ai platform provides our partners with an intuitive way to improve the customer journey — with no coding needed. This allows companies to automate repetitive workflows, whether they’re using a template or building their own custom flow. Our platform also unifies messaging across the web, mobile, and Facebook, providing customers with a seamless customer service experience, while a comprehensive dashboard provides real-time insights on customer behavior.

In 2022, Botco.ai launched Instachat Builder, the first conversational AI platform to use text context to train chat applications. Instachat Builder is trained on reams of data and automatically produces questions and answers, enabling faster deployment. In other words, businesses can quickly launch an intelligent chatbot that can answer hundreds of questions in a fraction of the time that was previously possible. Historically, launching a chat function has been an extremely time-consuming process, but Instachat Builder offers a shortcut through that process while still delivering 98% accuracy on day one.

Please tell us more about the role of AI and machine learning in health services.

Healthcare is a very complex domain because it involves systems that are not well understood, like human psychology and physiology, so the technologies that operate on these systems are also not always well understood. Moreover, a lot of conversations in the healthcare space leave a lot of key information unsaid because of privacy concerns, empathy, or embarrassment. Humans have no trouble “reading between the lines” in situations like these, but this is typically a challenge for AI programs.

So our belief is that AI can solve this problem by using causal models to “read between the lines” and determine what is being unsaid in delicate situations. In addition, dialogue managers and generative models can be used to make conversing with AI more comfortable and natural.

Recommended: AiThority Interview with Alan Holland, CEO and Founder of Keelvar

In the era of ChatGPT and generative AI, how does Botco.ai continue to lead the market with its breakthrough AI? Tell us more about the AI engine at the backdrop of your recent AI developments.

The challenge for generative AI in our market is that there is no tolerance for wrong answers. After all, we’re talking about AI systems that are being used to make healthcare decisions. We have harnessed generative AI where appropriate to train and maintain chatbots, but with the appropriate guardrails that ensure the chatbots are always producing correct answers. Examples of this are our Instachat Builder, which understands the customer documents and data needed to create the chatbots, and our Training System, which automatically analyzes chatbot interactions to perform ongoing revisions of the chatbots. Additionally, there is no place in our architecture today where we allow a generative AI to answer a question without some level of human review.

ChatGPT conversations are everywhere. How do you see ChatGPT and other generative AI apps playing a larger role in your industry?

The big breakthrough of ChatGPT (and similar task-trained models) is that they are able to generalize their knowledge to perform new tasks with little or no new training. So any NLP task that required a specially trained model can now be performed by a general-purpose model like ChatGPT. This dramatically reduces the cost of using these AI models, so I expect we will see them popping up everywhere.

How can we make AI innovations more inclusive and equitable?

Large language models (LLMs) are inclusive in the sense that today they are freely available and easy to use. On the other hand, they are products of their training data. So we should ensure both that LLMs continue to be widely available, and also that the data used to train them is as free of bias as possible.

What is your take on AI ethics and the democratization of data science ecosystems?

I feel good about democratization because the LLMs are freely available and inexpensive for everyone to use. The ethical dilemma is harder to parse because any large technological breakthrough changes society in ways that feel ethically challenging (such as the Industrial Revolution and the computer revolution). In the case of healthcare, I take comfort in the fact that we already have stringent privacy laws governing the use of health information, and AI systems are subject to these laws. This begs the question of whether we need to strengthen our privacy laws in other spaces.

Recommended: AiThority Interview with Lori Anne, Director of Product Development & Management at Verizon

An event/ conference or podcast that you have subscribed to consume information about B2B technology industry: If invited, would you like to be part of a podcast episode on CX and B2B SaaS?

I’m a fan of the All-In podcast. I generally agree with their perspective that the value of LLMs and other horizontal AI technologies will be low, and that value will accrue to companies that can build up proprietary data sets.  It’s also interesting to hear their perspective as tech investors in the B2B space.

And yes, it would be great to be part of a podcast episode.

Thank you, Chris ! That was fun and we hope to see you back on AiThority.com soon.