Could you tell us about your interaction with the new-age technologies like AI, Machine Learning and Automation in the IT industry?

As the CTO for a technology intelligence company trying to help our customers accelerate business outcomes, we are always experimenting with AI to find ways to bring innovation to our customers. Much of our intelligence is mined from document-based data and allows us to explore both “sides” of AI which is what makes my job so exciting. What I mean by this is the application of both Machine/Deep Learning and Natural Language Processing (NLP). There are so many tools now to allow for rapid ideation and experimentation with AI, which for us is extremely important because of the sheer size and complexity of our data pipeline and the cost to take Data Science experiments full scale. These tools have proven invaluable to us.

What is your role at HG Insights and what is your background in AI?

I am the CTO at HG Insights. This has been both the most challenging and rewarding role of my career. At HG Insights, I am responsible for Engineering, Data Science, and Data Operations. My background in data goes back over 20 years. Prior to HG Insights, I spent a lot of time as an engineer in the Data Transformation, Integration and Data Management space. I was introduced to Data Science while working at Liaison (recently acquired by OpenText) on their Alloy Data Platform, particularly in the Healthcare Informatics and Clinical Trial data space. Later on, I helped to build a consumer-based Data Analytics platform for MuleSoft. However, coming into HG Insights, I had to quickly ramp up my Data Science and Machine Learning skill set. Working with the scale of data that we process monthly, I learned many lessons along the way, particularly around productizing Data Science beyond experimenting with Python, AI libraries, and notebooks.

How has AI changed Account-Based Marketing (ABM)?

AI has revolutionized ABM. The entire philosophy of ABM is underpinned by the ability to identify and focus on fewer accounts by uncovering which ones are of highest value to an organization. This has caused a wave of AI-based offerings. Examples of applied AI to help ABM can be seen by the proliferation of lead scoring models and propensity models to help predict which accounts are most likely to buy. In addition, AI is being used for look-alike modeling to find more companies like the ones that are most successful for a business — what Sales and Marketing teams call the Ideal Customer Profile (ICP). Deeper AI solutions that can process richer sets of data to chart the customer or buyer journey are also on the rise.

However, AI has also brought with it a wave of disappointment and black box fatigue. Many vendors touting “AI” are failing to live up to the promise of AI and the results have many buyers skeptical. With the recent rapid adoption of AI there isn’t always the same level of discipline (and dare I say ethics) to ensure elimination of bias or skew in producing a result. At the end of the day, the tools make it easier for Sales and Marketing teams to setup and conduct experiments with large sets of data, but if the insights they generate are flawed this will have a negative impact on the business, leading them to waste time on opportunities that don’t really exist.

How has NLP helped HG Insights provide meaningful outcomes to its customers?

HG Insights provides the most comprehensive and accurate picture of a company’s technology profile, which means we can tell you exactly what software and hardware products a company has installed. We do this by mining billions of documents. NLP is absolutely key to our success because without it we would not have context. Without context, we might, for example, not be able to disambiguate when a document referencing the world “Salesforce” is referring to  the CRM or to a department at an organization.

What is HG Insights doing in the area of AI to further its offering to customers?

We’ve recently filed two new patents in the AI-space. We are working on various new ways to predict what products companies are going to purchase. We are also working on a new method of named-entity recognition to allow us to continue to improve our ability to mine technology intelligence with greater precision and recall. Lastly, I am really excited by a series of persona-based intelligence initiatives using several different Machine Learning approaches which will give our customers a much deeper look at how technology is being adopted across an organization.

Do you see a difference in how companies approach AI as a product offering versus for internal analysis purposes?

Yes, very much so. Many companies who employee Data Scientists, do so to mine their data exhaust. Usually so they can make some kind of strategic/business decision to help increase revenue or decrease cost. Many of these are one-time experiments that may not have the same cost or runtime concerns as compared to Machine Learning-based “products.” At times, they aren’t even well-defined projects (for example, “Here’s a data lake, what can we learn from it?”). That’s not to say all Data Science projects are throwaways. Many times, even after learning has occurred that is used to inform the business, the insights derived may still be productized into something a customer can consume.

Another critical key related point is the fact that internalized Machine Learning is intended to have a direct impact on the business. This impact can generally be directly observed and with it comes accountability and transparency. When a company sells an “AI-powered” solution, there’s an impending promise of impact, which many times cannot be directly measured by the company producing the solution in the same way. It is the company who purchases the product (your customer) who ultimately determines whether your solution delivers on its promise. This fundamentally changes the dynamic of trust.

A lot of customers are getting ‘black box fatigue’ with third-party AI solutions. How do you tackle this challenge?

It starts with transparency. Transparency creates trust. There’s a push towards cognitive computing, which is the next level of AI. Humans want to know the “Why” of something to feel a connection with it. This creates trust in the system. In essence, I always want a user to be able to ask the question “why was this prediction made?” The best way to do that is to allow the user the ability to “double-click” or peer inside the black box so-to-speak. In the past, vendors make promises of ROI if the intelligence is applied correctly.

ROI can take time, and if not realized, a lot of finger-pointing can occur. I want our customers to have confidence long before their experiments yield fruit, because they can make a mental connection with the output. This creates trust. One of my all time favorite TED talks is by Simon Sinek entitled “How great leaders inspire action.” In it, Simon delves into how the human brain works, and that at the center of our trust and decision-making process lies the “why” of a thing. Connecting AI to “why” is truly the next wave of AI in my opinion.

How do you see the global SaaS scenario evolving around AI? Which technologies have been the biggest disruptors in this industry?

First and foremost, the accessibility of Cloud-based AI-tools that allow for rapid experimentation is a game changer. For me, this is the real disruptor that is enabling AI to appear across all industries. Tools like TensorFlow, AWS SageMaker/Comprehend, Jupyter, H2O.ai to name a few, are the real disruptors here in what we traditionally think of accessibility to AI. This is in essence the democratization of AI making it much more accessible. This is having an interesting effect of “AI-everywhere,” particularly when it comes to SaaS, but it is also having the side-effect of making us skeptical as we distinguish innovation from many false promises and let downs as well. I believe there is an obligation to be transparent in the application and use of AI as we have seen some of the adverse effects across marketing, and social media platforms (for example, “deep fakes”).

Being somewhat close to Sales and Marketing use cases, AI has really improved the accuracy of digital ad targeting, customer intent, and provided a greater overall understanding of the buyer journey, but once again, too much of a “good thing” can create an AI-triggered hangover on consumers who may be fatigued form over-targeting. I believe AI will continue to evolve and disrupt in Sales and Marketing, which is exciting to me being in this space.

What are your predictions on the future of AI? How can CTOs safeguard their teams from dramatic shake-ups in the industry?

Currently, in most cases, when AI is employed in a customer-facing solution, it is called out. Customers are generally aware of the presence of AI when engaging with a particular service or solution. This is rapidly changing. We are already seeing it particularly in social media. The lines will continue to blur as AI is adopted more mainstream. As a result of this, detecting fraud will become increasingly more difficult which will lead to heavy investments in cybersecurity to detect AI using, you guessed it – AI.

In terms of CTOs and safeguards, I have always preferred to work for disruptive companies, and as such, I welcome change. We already know that companies will struggle to stay relevant without having an AI strategy in all technical fields and most others (including Sales and Marketing). This pressure is very exciting to me as it continues to evolve and forces us to evolve with it, and in some cases – even take the lead. I believe the best thing a CTO can do is to learn how to incorporate AI development as part of their overall development process and philosophy, vs segmenting and siloing data science which can be challenging to integrate if not planned out.

What digital technology start-ups and labs are you keenly following?

Cortical.io and their retina project. Tesla, simply because I love seeing how disruptive their use of AI is proving to be to the automotive and energy sectors.

What technologies within your industry are you interested in?

H2O, TensorFlow, spaCy, AWS Comprehend/SageMaker (generalized tools), Apache Spark

Which superhero character/movie do you most profoundly relate to-

Iron Man

As a CTO of a MarTech company, what industries you think would be fastest to adopting AI to make Marketing and Sales operate with smooth efficiency? What are the new emerging techs supporting this growth?

Presently software/hardware companies have the highest level of maturity and propensity to spend money in Marketing and Sales technology. I don’t see that changing anytime soon. Within this particular space, SaaS/Cloud-based solution providers will continue to see adoption and growth. We are seeing a whole host of niche players in the Cloud/SaaS space emerging as a by-product of the Cloud+API (microservice) market revolution, something MuleSoft coined as the “Composable Enterprise”.

This creates a lot of specificalized competition across many players in the market where it’s much easier to be displaced. The result of this is an increasing need for technology intelligence-led ABM inside the sales and marketing stack at these companies to gain a competitive advantage so they can stay relevant (and in many cases, survive). It also means larger enterprises also have to respond with agility in order to keep or increase their market share.

Tag the one person in the industry whose answers to these questions you would love to read:

Andrew Ng