MarTech360 Interview With Kyle Maddock, SVP Marketing, AppTek

“Time efficiency creates time for more creativity and innovation, and that’s always a good thing.”

Hi, Kyle, can you walk us through your professional journey so far?

My entry into marketing was baptism by fire when I joined landline long distance telephone provider Tel-Save Holdings in the mid-1990s right after college. The company forged a $100 million relationship with America Online (AOL) which ignited the US dot-com explosion of the late 1990s. I was a member of a very small four-person team managing a multimillion-dollar marketing budget which included online pop-up advertising, direct mail and call transfer campaigns. We would reach a million consumers on any given day and achieve conversion-to-sale rates up to 10-15% or higher. The veterans I worked with joked with me at the time that I would never see anything like that as a marketer again in my lifetime – and they were mostly right.

As we expanded outside of AOL, I was able to evolve with the early generation of e-commerce as a first mover and buyer in paid search, affiliate marketing and co-branded partnerships. I built nationwide campaigns for long distance telecommunications as well as ultra-targeted zip+4 localized campaigns for street-level households for local telephone service. I then moved into the multi-carrier wireless industry where I was able to expand those same activities to a new customer base as well as build new channels for distribution on platforms such as QVC and the Home Shopping Network.

From there I pivoted into the direct response television industry where I took on a Creative Director role at Ignite Media Solutions. Ignite is a performance-based marketing agency that handled the inbound phone and web ordering from “As Seen on TV” television advertisements. I contributed in extending the brand to more consumers on the web, by building popular online branding campaigns for products including the Snuggie, Flex Seal, Shake Weight, and many more. One of my favorite projects was building the brand for the Dirty Jobs cleaning product line in conjunction with Mike Rowe, it was an amazing experience working with that team.

As the media landscape, FTC rulings and online shopping behaviors changed the DRTV industry in relation to e-commerce, I started doing some consulting work for AppTek.  AppTek supplies the automatic speech recognition technology to Ignite which is used to automate the buying process for inbound direct response telephone orders, converting customers at a rate higher than live agents at a fraction of the cost. AppTek needed some marketing help to expand its brand and that’s how I began interacting with the team and eventually ended up where I am today.

Tell us about your role at AppTek. What drew you to join the company full time?

In 2017, I was sent to eBay headquarters for a conference on machine learning. This was my first exposure to AI, and I was in absolute awe of what was being accomplished in the areas of computer vision, natural language processing and behavior prediction. I immediately enrolled in an online machine learning developers’ course to further understand the technology and have been fascinated with it since. I had begun to understand what the technology was capable of and knew I wanted to be part of this industry.

Joining AppTek, I was offered direct access to some of the most reputable and academically cited AI and machine learning scientists in the world.  AppTek has over 30 PhD’s on staff who are masters in the complex mathematics that make computers do what they do.  We emphasize that our scientists are toolmakers, not tool-users, which makes a big difference when trying to extend the capabilities of a given machine learning application. I was grateful that all of them are very friendly and personable, and kind enough to answer my questions no matter how simplistic.

Today at AppTek, I manage all marketing activities including go-to-market strategies, public relations, content, product, pricing, event marketing, digital marketing, and also have my hand in sales and business development activities.

Also Read: MarTech360 Interview With Hunter Montgomery, Chief Marketing Officer, ChurnZero

What sets AppTek apart from the competition?

AppTek is the only private full-stack language technology solution provider in the world in that, using our own technology, we offer automatic speech recognition (ASR), neural machine translation(NMT), natural language processing/understanding(NLP/U), and text-to-speech (TTS) under one roof.  The only other companies that can claim the same are the big cloud companies such as Google, AWS, and Microsoft.

The advantage of having all of these technologies under one roof is that our scientists collaborate and build these systems to seamlessly interact with one another for more advanced applications such as speech translation and speech-to-speech automatic dubbing.  If you try to fasten together disparate siloed ASR and MT systems, you won’t get nearly the same accuracy or results in comparison to building them together with the same teams.

AppTek is also very focused on equity and diversity for its AI solutions and models.  We deploy something we call 4D for HLT (Human Language Technology) methodology into our language models which takes into consideration Dialect, Demographics, Domains and Devices/Channels to ensure every voice is accurately captured and represented.  We consistently benchmark the performance of our models using scientific test sets to ensure we are delivering the best-in-class solutions to the market and work diligently to reduce the propensity for any types of bias in our models.

Lastly, AppTek scientists are a rare breed in that they are not only focused on research, but also on solving real customer challenges.  Traditionally, you often see the former when it comes to this level of science but not so much the latter.

How did the COVID-19 pandemic affect your business?

Prior to Covid, we had significant momentum – we had just earned one of the top spots in Washington Business Journal’s Fastest Growing Companies – and planned to keep sustaining growth through conferences and events.  Covid was a bit of a boulder in the road as conversations shifted, budgets changed, and it was much more difficult to maintain personal relationships with clients.  It was still a strong year for us, but it definitely set us a step back in terms of building new relationships.

On the flip side, the pandemic strengthened our relationship with Gallaudet University, the largest university for Deaf/Hard of Hearing (HoH) individuals in the world.  Amidst the sudden exponential rise in videoconferencing services during the pandemic, many services did not offer a turn-key solution to deliver Deaf/HoH users a means to access the captions of conversations, especially when sign language professionals/interpreters were not available due to scheduling or cost restraints.

I worked with a group between Gallaudet and AppTek to help incubate a new business entity named “GoVoBo”, which offers universal access to automated captions and translations via a user interface developed by the Deaf/HoH community.  It was an amazing and humbling experience working with the Gallaudet team. I learned so much about what “accessibility” truly means and the importance of addressing all communities of differing abilities with utmost priority.

How, according to you, are MarTech and marketing automation helping in transforming the marketing landscape?

MarTech and automation are invaluable tools to manage the entire end-to-end lead-to-sales funnel. There are so many great services out there now competing with one another that we as marketers reap the rewards. The value of these technologies, from my perspective, is that they allow you to be more focused on the activities that matter rather than spending excess time manually working through lead and customer management. Time efficiency creates time for more creativity and innovation, and that’s always a good thing.

Why is voice technology growing popular in MarTech and where does it fit into the MarTech landscape?

Voice applications have found their way into almost every part of our lives through the rise of mobile and home-based voice assistants, automotive assistants, IVR systems, IoT devices, healthcare applications, and so much more. Every marketer needs to consider the role and impact of voice on their service or platform as it is the preferred method of interacting for the majority of users, so it shouldn’t be treated as a second thought.

Additionally, voice can serve as a valuable resource for gathering business intelligence and marketing feedback by tapping into a brand’s contact center conversations. By mining customer/agent interactions, marketers can identify customer patterns and trends that offer insights into brand performance, product, returns, customer churn, new revenue opportunities, customer sentiment and so much more. AppTek has a product called Lucid that aggregates conversations from cross-channel interactions and provides actionable intelligence to better inform enterprise decision making.

Where do you see the role of Human Language Technology (HLT) Solutions in today’s economy and the future?

Technology has connected individuals and businesses around the globe, and now, globalization is happening on a new scale with the rising trend of individual content creators and small businesses on social media. The explosion of user-generated content through platforms such as Youtube and TikTok can have a much wider reach globally through localization, and the only means to do that cost-effectively and at scale is through automated HLT solutions to assist with the end-to-end speech-to-speech translation process.

Still, with something like 7000 spoken languages today in the world, there is still a large gap in what we can automatically transcribe and translate. Today at AppTek we support models for approximately 80 different languages and dialects, but we also focus on what is called “low resource languages” which include under-represented languages that have limited data sets. I think it’s important to preserve as many languages as we can, as the diversity of languages across different cultures is what makes us wonderfully human and so we should strive for equality among them from a technological standpoint.

What is AppTek doing to further the research and development of Human Language Technology applications?

Today, AppTek is focused on automatic dubbing, which takes a media file from a given source language and automatically converts it into a target language while presenting it back in similar-sounding voices and emotion as those of the source speakers while matching the time constraints between speaker changes. There is a wide range of complex speech and language services at play in this pipeline, but one of the “secret sauces” that makes this technology work effectively is our meta-aware machine translation with length control features.

To explain by example, if the source language has a speaker speaking a sentence of 30-characters in length within a 3-second clip, and the first direct translation of that same sentence is produced at a 50-character length, the speed of that voice on a speech synthesized output would sound faster and unnatural within that same 3-second time constraint. AppTek’s MT predicts the best translation closest to the original character length (+/- 10%) of the source language without losing context, so as to make the automatically dubbed media file sound more natural when presented back in a new language. It is fascinating how complex languages are and what scientists can do to account for all these nuances.

Beyond dubbing, the AppTek science team continues to research a wide array of topics including cross-lingual emotion transfer, sign language recognition, and many other facets of speech and language recognition and automation.

Could you name some platforms/tools that you use for marketing?

I run various platforms/tools inside AppTek’s marketing stack, including tools for CRM, lead generation, email, social media, analytics and more. If I were to name drop a few applications that I find useful which may not be on everyone’s radar, I would include SEMRush for not just search engine optimization, but also for gathering competitive analysis. There is an incredible amount of useful information you can gather just by analyzing a competitor’s or customer’s backlinks.

I also enjoy using AI-enabled GPT-3 content writing tools such as Frase.IO and Jasper which can really help when trying to formulate new content.  I’m excited for what creativity AI will introduce into the market with AI image generators such as DALL-E 2 and Midjourney.  I highly recommend anyone who hasn’t had exposure to these image generators to try the Dall-E mini models themselves at craiyon.com and see what these tools are capable of.

What would be your advice for all the aspiring marketing professionals out there?

I think everyone drives home the importance of analytics and being data-driven as a marketer, and it is certainly important. But I also believe in today’s world, with so much noise around us, as a marketer it’s more important than ever to connect with others one-on-one, listen and empathize with your prospects and customers on a personal level, deeply understand what they are trying to solve and then tailor your product or platform to deliver on those specific needs. Connection and understanding is key.

Thanks, Kyle!