Simple AI Blog

Business Research on ‘Steroids’ using Powerful Artificial Intelligence (AI)

By June 5, 2017 No Comments

AI Conversation with Shahzad Khan, President and CTO of Gnowit.com

SHORT RECAP

FREE: Exclusive Time-Based Offer from Gnowit

Today I spoke with Shahzad Khan, President and CTO of Gnowit. Using powerful AI algorithms, the company’s platform provides intelligent research by gathering data from the web, emails from social media and then uses semantic analysis and other techniques to summarize the research information. Think of it as Google Alerts on steroids. As a marketer, you’re going to love this platform.

TIMELINE

  • 01:22 Shahzad Khan Background
  • 03:24 Big Challenges for People Doing Research
  • 03:30 Finding the Right Information
  • 04:13 Extracting Meaningful Pieces of Information
  • 08:02 Google Alert Comparison
  • 09:09 Gnowit Data Sources
  • 09:25 Gnowit for Social Media
  • 11:22 AI inside Gnowit
  • 14:30 Quick Industry Insights at a Glance
  • 15:36 AI Trends
  • 17:57 AI Predictions in Relations to Business and Marketing
  • 20:35 Advise for Businesses Looking to Implement AI for the First Time

KEY POINTS

  • Any organization that is not taking advantage of AI capabilities within their business are going to have a serious problem.
  • The first thing to do is identify when and where decisions are made within the business process. Especially the ones that involve clients. When you figure out where the decisions are made, it becomes easy to identify what data is required. If they have enough data available, whether it’s a ledger, excel sheets, or emails – then you have something which can be mined.
  • Gnowit is like Google Alerts on Steroids. It pulls data from more sources than Google, but through advanced AI systematically filters out poor quality. Then it reads, interprets and summarizes all the information for you.

TRANSCRIPT

Justin Jones: I’m speaking with Shahzad Khan, President and CTO of Gnowit. Using powerful AI algorithms, the company’s platform provides intelligent research by gathering data from the web, emails from social media and then uses semantic analysis and other techniques to summarize the research information. Think of it as Google Alerts on steroids. As a marketer, you’re going to love this platform.

Tell us a little bit about your yourself.

Shahzad Khan: My career started as an economist rather than a scientist back in 1993. I studied economics for a year before realizing I enjoyed the math, I enjoyed the fact that I could identify patterns that could potentially predict the future – even though I recognized it was not going to happen in my lifetime.

I began taking a couple of courses in computer science as I had an interest in programming. For me, programming was one of the last great crafts that you could use your hands and see instant results.

I went on to do my Master’s Degree in the United States and PhD at the University Cambridge in the United Kingdom. I kept going back to the fact that with programing as a skill I had the potential to create software that could augmented my everyday tasks.

The first program I built allowed me to track my to do list. This laid the foundation upon which I developed my interest and career. Today we have an advanced technology that ‘understands’ what my interests are and automatically sources and curate’s information (from everywhere) to help me achieve my research objective.

Justin Jones: What challenges do researchers have and how exactly does Gnowit help them overcome these challenges?

Shahzad Khan: One of the three biggest challenges facing researchers has to do with finding the right information. We have a software platform that is pulling information from everywhere – newspapers, press releases, blogs, and social media etc. On average, we’re making about 3 billion web calls every month, and see about a million unique documents every day. This far outstrips the capability of any individual for sourcing information. This is just one of the three problems facing research and marketing data.

The second problem you have is filtering this vast amount of data so you can identify and focus on the most relevant information. There are typically three techniques for achieving this. One way is to do it by hand. As you can imagine, this is a very onerous task to perform on a day by day basis. The average person could probably go through thirty to forty sources, but that’s not a very good use of anyone’s time.

The second way is to focus on a few trusted information sources that cover the topic you’re interested in. The downside is that your information is limited to mainstream sources which are heavily editorialized. Often you will find valuable information in sources that are not mainstream, which means you would not be exposed to these insights following the second approach.

Thirdly, you can use a service like Google Alerts. You identify the keywords of interest and Google will provide you with a ton of information. Using Google Alerts is similar to the manual research approach, except now you have exponentially more information to read, sort and digest. This is not a sustainable research approach.

With Gnowit, your unfair advantage is that we are pulling data from the ‘global information pipeline’, however it is being curated specifically for your research needs. You can cover a lot more research in a lot less time, and will not miss any key information because it’s not in the mainstream media.

One of our clients was receiving roughly 80,000 unique documents every day, all of which were relevant to their needs. Once Gnowit added a semantic layer and simple intuitive dashboard, the client become significantly more effective with less effort. The end result was basically 100 well summarized documents with excellent workflow.

Justin Jones: You actually summarize the data as well?

Shahzad Khan: Absolutely. Our summaries are of three sorts. Firstly, we summarize based on the actual topic. So, if you have lots of documents that cover the same event, we will collapse them into the most relevant.

Justin Jones: I’ve been using Google Alerts for many years, for both personal and marketing research. Although it is significantly easier than the manual approach, I have found a lot of duplication in terms of articles.

Shahzad Khan: We recently carried out an experiment where we compared Gnowit data with Google Alerts.

One of the things we noticed was that there was very little overlap between the two services. Because our sources are highly curated (at core level) we know that these are the best sources around the world.

During our experiment, we were collecting roughly twice as many results as Google Alerts with and only about a 15 percent overlap between the services. What this basically means for people who need quick access to critical information – Google Alerts is overworking you.

Google Alerts delivers a lot of unnecessary data that researchers need to wade through to reach the valuable information. For example, they will pull data from job postings, advertisements on Kijiji and Craig’s List, and generally pulling data from sources that are not authoritative. All this data is pushed to you because it contains a few keywords that are important to you.

On the other hand, Gnowit pulls data from authoritative sources of which some are less mainstream. Sometimes the biggest insights come from the most obscure sources. Poor quality is systematically excluded so Gnowit helps you uncover the hidden gems wherever they may be hidden.

Justin Jones: I assume Gnowit is great for social media content research?

Shahzad Khan: Gnowit is absolutely amazing for social media. Firstly, you are pulling in information from sources that are not necessarily in your feed.

Secondly, we also pull data from people in your feed who share information which suggests that there is value in what they shared and may be worth reviewing.

We run this data through the same analysis tool-chain that we use on the other sources and tag it specifically to your Feed. So, you’re getting real-time data from people that you respect, and the authorities they recognize – who may in fact be putting out information for the first time on the web. This type of information will often appear in your inbox before it is picked up by mainstream media.

Thirdly, as our platform provides workflow you have the ability to choose which articles you push out to your various networks. Your feed is curated using AI, so you just select the articles you want to share. You also have the power to create custom newsletters. This literally takes a couple of minutes and with one click you can push it out to your website, social networks or email list.

Justin Jones: How does Gnowit leverage AI?

Shahzad Khan: In terms of Artificial Intelligence, we use it in many different places throughout our platform. The first is the curation.

We use AI to automatically identify, monitor and scan websites and their articles. With machine learning we are able to automatically detect things like advertisements, obituaries, About Us pages and other things that are not newly added. Our system also has a mechanism by which we can frequently scan a website and identify all the places where change has happened. We essentially reverse engineer the sitemap and that gives us the means to monitor sites every couple of minutes without ever being banned.

We then have pattern analysis technology that automatically takes new documents and induces a scraper to identify the best text sections while removing irrelevant content like navigation, advertising, unnecessary comments and links to other articles. So, you essentially end up with a plain Word Document with all the noise and irrelevant data removed.

We then analyze the clean text, generate the summary, and identify the key concepts which are then placed in the supporting metadata and metrics within the live data warehouse. This is an example of just one of the many processes in our system.

Then we have a process by which we gather feedback from customers to identify preferred content. Our system then learns to exclude or promote certain types of information based on the feedback. With Gnowit we you essentially stay on top of the latest trends and information, including content of interest that has not been discovered by the mainstream.

Justin Jones: AI has quietly permeated many aspects of our life. As marketers, we are not giving AI the attention it deserves. If we don’t actively start adopting AI in key marketing processes, we’re putting ourselves at a disadvantage in the next 2 to 3 years. Most likely finding ourselves in a similar situation to where traditional retail businesses find themselves – unable to effectively compete with the Walmart’s and the Amazon’s of the world.

What do you see as the defining Artificial Intelligence trends marketer need to be focusing on?

Shahzad Khan: One of them is the whole idea of automation and digital disruption. We did a calculation to determine how many people it would take to deliver comparable results to Gnowit. According to our calculations you would about a billion-people working 24/7 to deliver full functionality. We would need about 70,000 people just to process the data coming into our system right now.

Any organization that is not taking advantage AI capabilities within their business are going to have a serious problem.

The second element is personalization. We’re continually working on improving our curation so people only see the information that they’re interested in when looking at emails and websites. This kind of personalization adds value, engages and builds trust; essentially the function of marketing.

Justin Jones: What do you see happening with regard Artificial Intelligence in business?

Shahzad Khan: Because of all the excitement around Artificial Intelligence, people are talking about it but they don’t really understand what it means for them and their businesses. I believe in a couple of years there will be a few key AI offerings that people will understand and begin to use, like personal agents.

These personal agents will be more than Amazon’s Alexa or Google’s Home. It will do what our smartphone does today (and more), without the need to explicitly and proactively identify stuff and interact at the level we currently do.

There is also places that will silently use AI to create value, especially on the business side. People are still putting a lot of effort and sweat into tasks that if properly structured machine could do. People are also creating a lot of data in their daily lives – every time they go to the supermarket, buy an item, use GPS navigation, or visit a website. All this data gathered from individuals can be used to help make recommendations and predictions to streamline our decision-making efforts. AI will learn from our past actions and decisions.

Justin Jones: What advice would you give a business that’s looking to implement AI for the first time.

Shahzad Khan: The first thing to do is identify when and where decisions are made within the business process. Especially the ones that involve clients. When you figure out where the decisions are made, it becomes easy to identify what data is required. If they have enough data available, whether it’s a ledger, excel sheets, or emails – then you have something which can be mined.

You need to have someone who is an expert, someone who understands how to convert the data (and value) into a machine algorithm. Perhaps semi-automatically to begin with, but automatically doing at least 80 percent of what was previously done by hand. A Great source for these types of capabilities are local universities, I find them to be good partners. If they are too busy teaching and writing papers, they will at the very least be able to provide you with solid, well informed advise.

RELATED LINKS

Gnowit.com
Gnowit is a tool designed for professionals who desire to know more. Our news database allows you to save time, uncover hidden information, and effectively serve your business’ or clients’ needs. Set your search, and let us do the rest.