Calgary tech entrepreneur passionate on how AI can help small business boost productivity

By: Cherine Hlady, Freelance Writer 

It’s an unseasonably warm fall afternoon and chata.ai CEO Kelly Cherniwchan is dressed in only a Star Wars t-shirt and jeans as he sinks into a deep leather chair in the meeting room at the top of The Edison in downtown Calgary.

The Edison – which is quickly filling its floors with a roster of high profile technology tenants in one of the city’s fastest growing sectors (source: Calgary Economic Development) – used to be occupied by a local oil and gas giant. Cherniwchan starts the interview by pointing out a magnificent, ancient-looking leather-bound reference set that sits on one of the room’s bookshelves. “See those books?” he says, “They’re not cheap. And they were just left here.”

It’s a small detail, but one that someone with Cherniwchan’s sensibilities for worth would immediately notice. With well-honed instincts to realize value and maximize benefit, an unclaimed, neglected, valuable antique is an affront to both.

His latest venture, chata.ai, is an extension of this interest in generating value. Through artificial intelligence focused in natural language processing – Cherniwchan and his development team at chata.ai are narrowing in on the untapped value of most business data. Specifically, re-thinking how we access, search and analyze it to boost productivity and enable better-decision-making.

 

The genesis of a business idea

Cherniwchan’s commercial experience with machine learning harks back to one of his previous companies. As one of the co-founders of Circle Cardiovascular Imaging Inc, he worked with a highly talented team using machine learning and statistical techniques to help computers understand cardiovascular images and analyze what they saw.

“I’d been taking some machine learning courses myself and discussed with the team on how we could apply these techniques for computer vision purposes. The team at Circle really took this to a world class level.  From there, my personal passions led me in to the natural language processing world, which is such a difficult and complex domain.”

This led to the start-up of a new business. At first, the focus was on developing chatbots (a type of conversational interface for a system that conducts conversations with users by text or voice), but when carrying out user experience research to inform his business idea, a key finding was his potential customers needed a solution to help them quickly and easily access their business data, not something that provided friendly chats for their customers.

What drew me to natural language processing is that it’s excessively difficult, but with difficulty there is opportunity as well.

“What we ended up creating is a conversational user interface where someone can just type in a question and the system knows where to find that data and bring it back for further processing. Even in cases where complex calculations are required, all the user has to do is ask. This is where we can add real value to small and medium size business owners. It’s kind of like having your own personal assistant for analyzing data.”

AI, at its best, empowers human workers by taking over the most time-consuming, manual aspects of their jobs, and freeing them up to use higher-value skills like selling new products, building relationships with customers, or creating operational efficiencies.


Finding the niche

Says Cherniwchan: “Natural language processing is definitely one of the harder areas in all machine learning and AI. It’s working with very obscure, inconsistent, and unstructured data. Humans can abstract* that stuff, although human miscommunication still readily happens, but for a computer it’s difficult. What drew me to natural language processing is that it’s excessively difficult, but with difficulty there is opportunity as well.”

The approach chata.ai has taken to overcome this difficulty is two-fold, first the company has created a scalable proprietary NLU (natural language understanding) engine that can rapidly train and customize user specific language models. Second, chata.ai can use this output to determine, on the fly, where to find and retrieve data for further processing.

Their first among many integrations is with QuickBooks Online.“So, if you are a QuickBooks online or other online accounting software user, we already have the language model trained,” says Cherniwchan. “The system can focus on training components specific to your business such as products, inventory, suppliers, etc. Then we would create a custom language model specific to you and your needs.”

* Abstraction is a conceptual process where general rules and concepts are derived from the usage and classification of examples.

There’s a higher degree of accuracy in this more specialized context, when someone types in a query the machine will translate it accurately. This is especially important when dealing with real-world business data.


Creating a great user experience

When I mistakenly refer to chata.ai as a “chatbot”, Cherniwchan is quick to educate me on what a chatabot is.  “a chatbot is the conversational interface the user types or dictates to and then gets some response.  Just like your mobile phone or webbrowser is an interface, so is a chatbot. It is the NLU service behind the scenes that does the real work.”

“With a lot of chatbots, very few have been able to have a good experience monetizing it” says Cherniwchan because what the usual experience is with NLU systems is responses such as “I don’t understand what you’re saying,” or “click here for help” or “can you please rephrase the question?” Users get frustrated and don’t want to use these systems any longer. With chata.ai, if the system doesn’t understand what you’re talking about, it’s going to use a portfolio of technologies to support the user with recommendations or targeted help.

The purpose of our technology is to support people by taking on cumbersome data search and analysis tasks

“Our whole product development strategy is to have the user never get stuck,” said Cherniwchan. “We are thinking of all the different ways that someone could ask the same accounting question (for example, For each customer in Calgary what are the total sales per product for the last 3 months?) and testing that our application can understand the question accurately every time. Often we force our system to fail, because it helps us move forward and the system can refine its language model.”

 

AI as gateway to better data interaction, boosting productivity and creating value

Overarchingly, chata.ai is designed to be easy-to-use, intuitive, accessible and fast. Today’s business owners are overwhelmed and need the ability to make faster, better decisions. And while they rely on data to understand their business, getting fast and easy answers from that data, says Cherniwchan, isn’t easy.

For example, say a business manager wants to understand how much they spent with a certain supplier in the past year for a presentation he or she is working on. The ‘traditional’ way of accessing and analyzing this data would be for the owner (or, more likely, his or her bookkeeper or accountant) to find the data on spending with suppliers in a spreadsheet, sort it for the specific supplier they are looking for and the specified timeframe, export it, create a brand-new table of data, hit charting function, create the chart, save it as a jpeg file and put it into a PowerPoint.

With chata.ai, they could just ask: give me a bar chart of how much I spent with supplier xyz per month for the past year and boom the application would send back that bar chart in seconds. If it can take you five minutes or five seconds to do something, most people will take the five second option Cherniwchan contends. chata.ai’s pricing is consumable for everyday users, with the fee more than recovered by the potential time savings realized by using the application.

“Business owners and advisors for those businesses need real-time access to their data and more importantly, they want to spend more time doing business, and less time searching endlessly for answers.”

The same need to be productive extends to accountants and bookkeepers: their workdays can be constantly interrupted by calls from their clients (how much did I send with supplier XYZ in the past month?) when what they should be doing is focusing more on higher value work, obtaining more clients and being more productive.

Says Cherniwchan: “The purpose of our technology is to support people by taking on cumbersome data search and analysis tasks—AI can free you up to be more productive and add greater value.”

As the interview concludes, Cherniwchan returns to his company’s office space across the hall—a space that used to be occupied by two oil and gas executives and now houses chata.ai’s entire team of fourteen employees. Always as ever, he’s maximizing value from what he has.

It’s an unseasonably warm fall afternoon and chata.ai CEO Kelly Cherniwchan is dressed in only a Star Wars t-shirt and jeans as he sinks into a deep leather chair in the meeting room at the top of The Edison in downtown Calgary.

The Edison – which is quickly filling its floors with a roster of high profile technology tenants in one of the city’s fastest growing sectors (source: Calgary Economic Development) – used to be occupied by a local oil and gas giant. Cherniwchan starts the interview by pointing out a magnificent, ancient-looking leather-bound reference set that sits on one of the room’s bookshelves. “See those books?” he says, “They’re not cheap. And they were just left here.”

It’s a small detail, but one that someone with Cherniwchan’s sensibilities for worth would immediately notice. With well-honed instincts to realize value and maximize benefit, an unclaimed, neglected, valuable antique is an affront to both.

His latest venture, chata.ai, is an extension of this interest in generating value. Through artificial intelligence focused in natural language processing – Cherniwchan and his development team at chata.ai are narrowing in on the untapped value of most business data. Specifically, re-thinking how we access, search and analyze it to boost productivity and enable better-decision-making.

 

The genesis of a business idea

Cherniwchan’s commercial experience with machine learning harks back to one of his previous companies. As one of the co-founders of Circle Cardiovascular Imaging Inc, he worked with a highly talented team using machine learning and statistical techniques to help computers understand cardiovascular images and analyze what they saw.

“I’d been taking some machine learning courses myself and discussed with the team on how we could apply these techniques for computer vision purposes. The team at Circle really took this to a world class level.  From there, my personal passions led me in to the natural language processing world, which is such a difficult and complex domain.”

This led to the start-up of a new business. At first, the focus was on developing chatbots (a type of conversational interface for a system that conducts conversations with users by text or voice), but when carrying out user experience research to inform his business idea, a key finding was his potential customers needed a solution to help them quickly and easily access their business data, not something that provided friendly chats for their customers.

What drew me to natural language processing is that it’s excessively difficult, but with difficulty there is opportunity as well.

“What we ended up creating is a conversational user interface where someone can just type in a question and the system knows where to find that data and bring it back for further processing. Even in cases where complex calculations are required, all the user has to do is ask. This is where we can add real value to small and medium size business owners. It’s kind of like having your own personal assistant for analyzing data.”

AI, at its best, empowers human workers by taking over the most time-consuming, manual aspects of their jobs, and freeing them up to use higher-value skills like selling new products, building relationships with customers, or creating operational efficiencies.


Finding the niche

Says Cherniwchan: “Natural language processing is definitely one of the harder areas in all machine learning and AI. It’s working with very obscure, inconsistent, and unstructured data. Humans can abstract* that stuff, although human miscommunication still readily happens, but for a computer it’s difficult. What drew me to natural language processing is that it’s excessively difficult, but with difficulty there is opportunity as well.”

The approach chata.ai has taken to overcome this difficulty is two-fold, first the company has created a scalable proprietary NLU (natural language understanding) engine that can rapidly train and customize user specific language models. Second, chata.ai can use this output to determine, on the fly, where to find and retrieve data for further processing.

Their first among many integrations is with QuickBooks Online.“So, if you are a QuickBooks online or other online accounting software user, we already have the language model trained,” says Cherniwchan. “The system can focus on training components specific to your business such as products, inventory, suppliers, etc. Then we would create a custom language model specific to you and your needs.”

* Abstraction is a conceptual process where general rules and concepts are derived from the usage and classification of examples.

There’s a higher degree of accuracy in this more specialized context, when someone types in a query the machine will translate it accurately. This is especially important when dealing with real-world business data.


Creating a great user experience

When I mistakenly refer to chata.ai as a “chatbot”, Cherniwchan is quick to educate me on what a chatabot is.  “a chatbot is the conversational interface the user types or dictates to and then gets some response.  Just like your mobile phone or webbrowser is an interface, so is a chatbot. It is the NLU service behind the scenes that does the real work.”

“With a lot of chatbots, very few have been able to have a good experience monetizing it” says Cherniwchan because what the usual experience is with NLU systems is responses such as “I don’t understand what you’re saying,” or “click here for help” or “can you please rephrase the question?” Users get frustrated and don’t want to use these systems any longer. With chata.ai, if the system doesn’t understand what you’re talking about, it’s going to use a portfolio of technologies to support the user with recommendations or targeted help.

The purpose of our technology is to support people by taking on cumbersome data search and analysis tasks

“Our whole product development strategy is to have the user never get stuck,” said Cherniwchan. “We are thinking of all the different ways that someone could ask the same accounting question (for example, For each customer in Calgary what are the total sales per product for the last 3 months?) and testing that our application can understand the question accurately every time. Often we force our system to fail, because it helps us move forward and the system can refine its language model.”


AI as gateway to better data interaction, boosting productivity and creating value

Overarchingly, chata.ai is designed to be easy-to-use, intuitive, accessible and fast. Today’s business owners are overwhelmed and need the ability to make faster, better decisions. And while they rely on data to understand their business, getting fast and easy answers from that data, says Cherniwchan, isn’t easy.

For example, say a business manager wants to understand how much they spent with a certain supplier in the past year for a presentation he or she is working on. The ‘traditional’ way of accessing and analyzing this data would be for the owner (or, more likely, his or her bookkeeper or accountant) to find the data on spending with suppliers in a spreadsheet, sort it for the specific supplier they are looking for and the specified timeframe, export it, create a brand-new table of data, hit charting function, create the chart, save it as a jpeg file and put it into a PowerPoint.

With chata.ai, they could just ask: give me a bar chart of how much I spent with supplier xyz per month for the past year and boom the application would send back that bar chart in seconds. If it can take you five minutes or five seconds to do something, most people will take the five second option Cherniwchan contends. chata.ai’s pricing is consumable for everyday users, with the fee more than recovered by the potential time savings realized by using the application.

“Business owners and advisors for those businesses need real-time access to their data and more importantly, they want to spend more time doing business, and less time searching endlessly for answers.”

The same need to be productive extends to accountants and bookkeepers: their workdays can be constantly interrupted by calls from their clients (how much did I send with supplier XYZ in the past month?) when what they should be doing is focusing more on higher value work, obtaining more clients and being more productive.

Says Cherniwchan: “The purpose of our technology is to support people by taking on cumbersome data search and analysis tasks—AI can free you up to be more productive and add greater value.”

As the interview concludes, Cherniwchan returns to his company’s office space across the hall—a space that used to be occupied by two oil and gas executives and now houses chata.ai’s entire team of fourteen employees. Always as ever, he’s maximizing value from what he has.

Cherine Hlady is a freelance writer, editor and communications strategist. She develops and edits content – across a variety of print and digital mediums – for clients primarily in the high tech, government and not-for-profit sectors. She lives in Calgary, Alberta with her husband and two young children.
Media Contact – Kari Viccars; kari@qxw.f61.mwp.accessdomain.com | 587.435.5274
Photo credit – Lindsay Westren
Cherine Hlady

Author Cherine Hlady

Cherine Hlady is a freelance writer, editor and communications strategist. She develops and edits content – across a variety of print and digital mediums – for clients primarily in the high tech, government and not-for-profit sectors. She lives in Calgary, Alberta with her husband and two young children.

More posts by Cherine Hlady
X