Chatbots have emerged as one of the most prominent platforms for the banking, e-commerce and retail sector. We spoke to the Head of Operations of Amplify.ai, a Californian startup which has an office at Noida, Kartik Walia, to know about Amplify.ai startup and how it has expanded its operations and customer base in India since 2016.
How Did It Begin?
The journey of Amplify.ai began in the year 2016 when the CEO, Mahi de Silva and the CTO, Manoj Malhotra while travelling to China on a business for a prior company, were blown away by how messaging had become the primary way consumers interacted with businesses, and how well that worked for both parties. After witnessing the chat boom in the East, the duo started Amplify.ai to ride the massive wave of B2C messaging outside of China, leveraging AI to enable conversational engagement at a superhuman scale.
How Is Amplify.ai Unique?
Walia said that chatbots are changing the changing the functioning of different industries for good. With chatbots and conversational AI, we are mastering the art of communication and are able to take it beyond what is humanly possible. With it, we can create rich consumer profiles and use them to understand a user’s intent, sentiment, and preferences based on his or her behavior – both in the short term as well as the longer term. “It is just like having a perfect candidate to do a job, who is familiar with all of your existing customers, quickly becomes familiar with your new or potential customers, never gets fatigued, and has an unquenchable thirst to learn. So, this candidate creates impeccable sales pitches, helps understanding the customer better, and makes up for any inadequacy left by a business in servicing the customer. It naturally increases the productivity of verticals that it is deployed in and the broader market at large”, says Walia.
Talking about how their startup differs from the competition in the market, he said that they see first-generation chatbots as a step in the right direction, allowing automated interaction over messaging channels. But the ultimate goal is to achieve conversational AI, with which brands can have 1:1, personalised, and persistent conversational relationships with every prospect and customer, across the entire customer journey. To that end, Amplify.ai has developed the industry’s first and only enterprise-class conversational AI platform.
The technical team behind Amplify.ai currently has 30 people. Much of the core team has been together for over 15 years building high scale tech platforms, driving three $100+ million revenue businesses. The technical team is very seasoned, and covers back-end, front-end, APIs, and core AI tech.
It natural language processing techniques and machine learning algorithms to solve for intent classification, sentiment extraction, contextual learning, content recommendations, customer service routing and many other use cases. A typical day of the technical team looks a lot like any software startup, and varies greatly by role (sales vs. engineering vs. marketing). Walia said, “I suppose the area where we’ve got uniqueness is that we have two teams of “builders” who use our platform to build AI-powered branded virtual assistants for our top-tier clients.” Currently the chatbot has implementations in English and a number of Indic languages, including Hindi, Bengali, Marathi, Gujrati, Urdu, Kannada, Malayalam.
The company had to figure out how to derive meaning from very short phrases, overcome challenges of typos, alternate spelling, for example “u” as shorthand for “you”, and emoji. Conventional NLP systems were developed to work with longer blocks of conventional prose, so they had to build their own messaging-optimized NLP system and amass a large enough messaging-based conversational dataset to train it via ML. Walia said that this was a bit of a chicken-and-egg problem, but they managed to solve it with some pricing creativity.
When asked about the skills required in the industry, Walia said that the obvious answer is computer science for the AI piece of the puzzle. The not as obvious answer is English, creative writing, and design, as UX (user experience) is a critical component of developing compelling conversational experiences.
Setting Up Workforce In India
India has been an attractive proposition to the startup for a number of reasons. As the largest mobile-first market outside of China, they felt that India would be an ideal market for AI messaging in terms of both early adoption and scale. And scale is really important for ML datasets. Their early successes in India, getting us to our first billion consumer engagements and first 100 million users, gave them conversational dataset was critical to training our NLP that has now allowed to get to over 2.5 billion engagements with over 175 million consumers worldwide.
India is also a hub for us serving the broader South Asia and SouthEast Asia markets. They chose Noida in particular for proximity to a lot of the companies that they sought to partner with or sell to. Amplify.ai is a partner of several globally leading brands including Facebook, Google, Google’s AdLingo team, Kohler, Quandoo, and Change.org. They also work with Zee Media, Network18, ABP News, The Quint and Reliance Jio.
E-commerce is one of the chatbot’s top markets. Prior to their conversational AI solution, the state of the art in e-commerce marketing was email. All flows were optimized to get a user’s email address. But email has very low open rates. Messaging-based, AI-driven conversational engagement is clearly a superior channel, offering dramatically higher open rates, retention rates, and conversion rates. In addition, conversational AI drives higher LTV (lifetime value) of the customer by allowing greater personalization over time by applying machine learning to the history of interactions. For e-commerce brands, Amplify.ai helps to drive higher sales with conversational marketing over messaging channels. They have shown greater than 2X lift in conversion and 10X improvement in 6-month retention vs. web/email.
That chatbot also has worked for media companies and help them to get an ever-growing audience on Facebook Messenger, which they can use to amplify engagement, consumption, virality, and ad-based monetization. It has also helped the political parties and campaigns to them drive higher engagement, which leads directly to more donations.
The Amplify.ai platform today supports conversational engagement on platforms of Facebook, Facebook Messenger, WhatsApp, Instagram, RCS (Rich Communication Services), desktop and mobile web, mobile application (through in-app integrations), and conversational ad units, including Facebook’s Click-to-Messenger and Google’s AdLingo. By turning a click into a conversation, they are able to move prospects down the funnel before handing them over to the advertiser’s website.
They have also partnered with media company ABP to create an informed and engaged society through its consumer’s interaction, analysis and customized content solutions, and are developing AI virtual assistants for ABP live FB properties and are live on three of its most prominent social channels - ABP News, ABP Live and ABP Majha.
Amplify.ai, for the most of 2018 were focused on getting to scale on the conversational dataset front in order to train their NLP. Their platform has now processed over 2.25 billion engagements with over 175 million users, yielding a massive dataset. Having achieved that data goal, the startup has now shifted their top focus to monetization, and their key goal for 2019 is revenue growth.
Amplify.ai thinks of itself as the “Switzerland of conversational AI.” At a business development level, it seeks to partner with any platform where B2C conversational engagement can happen, so additional messaging platforms, social platforms. Examples include Twitter and Apple Business Chat. On a business level, they seek to partner with more of the leading media, e-commerce, and telecom companies, as well as expand into new verticals.
As long as their long-term goals are considered, Walia said, “As proud as we are of developing the industry’s first enterprise-class conversational AI platform and having it drive amazing business results for our clients and partners, we are far from done.” The startup wants to make AI-driven conversational engagement the primary way that brands interact with their prospects and customers. What that will look like is better user experiences, deeper long-term relationships with favorite brands, marketing that is way more efficient and effective, and customer support that is truly better at a lower cost.
Register for our upcoming events:
- Meetup: NVIDIA RAPIDS GPU-Accelerated Data Analytics & Machine Learning Workshop, 18th Oct, Bangalore
- Join the Grand Finale of Intel Python HackFury2: 21st Oct, Bangalore
- Machine Learning Developers Summit 2020: 22-23rd Jan, Bangalore | 30-31st Jan, Hyderabad