The Future of Query has arrived...it's Chata.ai

by

The future of query has arrived, at least for users of QuickBooks Online. I believe that future rests with Chata.ai, or at least the technology that Chata brings to the table.  Chata.ai is a simple-to-use analytics and reporting tool that does your data 'dirty work' for you through the fantastic power of artificial intelligence. It's "AI-tastic."

With Chata.ai you will make better business decisions because you will turn your QBO 'data' into information you need to know.  There won't be any exporting of your data to Excel or some other program where you have to key in formulas and produce pivot tables to see what your data means, with Chata.ai you will use the App's artificially intelligent natural language understanding engine to simply query for the information you want. 

If all of this seems like science-fiction you might be right.  In my mini-series titled, The Future of Query in an AI World I review the history of computing and more specifically Query. I even give you my own (Murph) definition of Query and then describe all the bother and hassle computer users have gone through to pull, pry and steal their information back from the computers to which we entrusted it.

Learning a complex computer language to write SQL (structured query language) queries, or migrating data into some complicated software like Excel or Crystal Reports just to try to get the number you need is, to put it in one word, 'old fashioned' (OK, that's two words, but you get my point).

If I can Google the Internet and find almost anything, why can't I simply bring similar technology to my data, my QuickBooks data, and just ask a sophisticated query engine for exactly the data I want? Well now I can, that's exactly what Chata.ai was designed to do.

Kelly Cherniwchan, the CEO of Chata.ai told Insightful Accountant in an interview earlier this year, "We came up with the idea of creating a conversational user interface where someone could just type in a question and the system would know where to find that data, retrieve it, perform any computation necessary, and then display it for the user."  He went on to say, "Even in cases where complex computation is necessary, users should still only need to ask for the data they want, the system should do the rest."

Cherniwchan went on to say, "Our vision at Chata was to add real value to small and medium business owners by empowering them with easy access to the information they needed from QuickBooks to make better business decisions."

I asked Kelly about the use of artificial intelligence and machine learning in developing the natural language query engine upon which Chata.ai is based and he responded, "Natural language processing is one of the hardest areas in all of machine learning and artificial intelligence because it works with very obscure, inconsistent, and unstructured data. While it is easy enough for humans to abstract that stuff, it's extremely difficult for computers."  He then told me, "What drew us to natural language processing was the very fact that it was so 'very difficult' because with such difficulty comes great opportunity."

To overcome the difficulties Chata.ai took a two-fold approach. The first was to create a scalable proprietary natural language understanding engine that could be rapidly trained and customized to make use of user specific language models. The second was their methodology allowing 'Chata' to use the natural language query so as to easily find and retrieve the data needed for further processing.

Chata.ai chose QuickBooks Online to be their first real-to-life integration, and quickly undertook the task of developing their natural language model to focus on things like products, inventory, customers, suppliers, and accounts, the kind of data users work with in QBO.  But 'Chata' also had to learn the custom language of accounting, with unique terminology that professionals might use even if their clients, the QBO end users didn't. Now 'Chata' loves accountants, and QBO users alike.

In a subsequent interview a few weeks ago, just before Chata actually 'went live' on the QuickBooks App Center, Kelly told me that their 'Beta testing program' had been one of the greatest benefits to 'teaching Chata a thing or two (or thousands of things actually).'  Not only was Chata exposed to more expansive 'data sets' than the QBO test data used initially for machine learning, but the types of questions that users kept wanting to ask helped develop a far broader language model than they started out with. 

There was so much 'learning going on' from the beta participation that the company actually went back and made major revisions to the processing engine to allow it to handle the expanded modelling and respond more quickly.  Chata also gained some lessons in 'interpreting' the ambiguities of how we humans differ, especially when we ask about our finances.

For example, one person might ask "How much did I make last quarter?" What they mean is 'how much was their total revenue last quarter'. But another person asking the exact same question, the same way, might really be wanting 'how much their actual profit was last quarter', in other words revenue minus expenses is their measure of 'how much they made.'  Over time, Chata can learn these distinctions based upon how the user responds to the options Chata displays. 

The idea was to insure that users don't get stuck in their queries, not knowing how to go on if an initial response isn't what they expect. But even if Chata 'fails to deliver' on a specific query the company wanted a product that would learn from its mistakes, so that it got better at refining the language model, and thereby return a better response to future queries from the user.

But enough about the 'how' and 'why' of Chata.ai, let's talk now about the 'what' of Chata. If you are a business owner, ProAdvisor or Accountant using QuickBooks (online) you are probably interested in what the accounting data means, the answers it can give, and the directions it can provide.  You rely on QuickBooks data to provide an understanding of your business, or your client's business. 

For example, if a business owner wanted to understand how much they spent with their primary suppliers they would traditionally access QBO, and then try to find the right report to give them the information they were looking for. If they wanted details they probably might do a little filtering in QuickBooks and then send the data to Excel so they could do some more sorting, computations and ultimately graph the results.

With Chata.ai, they could simply request: "give me a pie chart of how much I spent with my suppliers."  In moments Chata displays the results.

Want to see the specifics on your top 10 customers, just ask Chata, "give me my top 10 customers by revenue." Chata responds with your results.

You could easily see all your customer revenue in QBO, you could even sort the entire list by amount from top to bottom, but there is no way you could get a list of just the highest 10 revenue customers, not without shipping the data to Excel or another reporter and setting-up a query the old fashioned way.

What about your customers by location, if you are looking for a graphic display of your A/R balances 'by State', just tell Chata, "give me a bar graph of total customer balances by state."  Tick tock goes the second-hand on the clock, and Chata gives you the results.

You would need to do some pretty significant query design to get this out of QBO, but Chata makes it look easy. Like I said, Chata does the 'dirty work' of query so you don't have to.  How 'AI-tastic' is that?

I could give you one example after another of just how sci-fi Chata has turned QuickBooks Query, but that's not the best way to see Chata at work.  The best way is to experience it for yourself with a Chata.ai subscription.

But, if you aren't quite ready to begin your own 'future of Query' experience, then sign-up for a 'chat with Chata' in a free one-on-one Chata.ai demonstration. Otherwise, check out the Chata.ai website for all the details I was too excited to tell you about.

By the way, if you are a ProAdvisor, Accountant, Bookkeeper or Consultant be sure to check out our Advisor Alliance article on Chata's partner program.