IIMC-CRC-2017-23

IIMC C ASE R ESEARCH C ENTRE (IIMCCRC)

I NDRANIL B OSE , P RATIK T ARAFDAR
M ARCH 2018

MANTHAN’S MAYA: ASSISTING EXECUTIVE DECISION

MAKING THROUGH VOICE ANALYTICS
(Atul Batra, Chief Technology Officer (CTO) of Manthan imitated the persona of a fictitious
CEO, Bill Smith, who was at home making breakfast for his kids and wished to have a quick
sense of how business was going)

Bill: Alexa, talk to Maya.

Maya: Good Morning. I am Maya from Manthan Retail Analytics – your virtual assistant. I
can give you insights to power your business decisions. May I know who am I assisting
today?
Bill: This is Bill.
Maya: Hi Bill Smith. How may I assist you today?
Bill: Alexa, ask Maya, what is my business update for today?
Maya: Hi Bill. The business update for January 16th is as follows – net sales stand at 884
thousand dollars; you have achieved 78 percent against the plan, up by 15 percent against
last week with a gross margin of 23 percent up by 7 percent against last week.

On January 15-17, 2017, Manthan, advanced analytics and business intelligence company for
retail and consumer packaged goods (CPG) showcased Maya and its Artificial Intelligence (AI)-
powered decision-making abilities at booth #4111 of the NRF Convention and Expo at Jacob K.

Prof. Indranil Bose and Pratik Tarafdar of the Indian Institute of Management Calcutta developed this case
study as the basis for class discussion rather than to illustrate the effective or ineffective running of an
organization.

This case study is meant for use in Executive Education Program for a course titled “Advanced Programme
in Financial Technologies and Financial Analytics” taught by Prof. Indranil Bose of IIM Ahmedabad starting
16th November 2023. Beyond limited printing rights, copying, distributing or posting of this case study in
any form on any media is strictly prohibited. The limited rights to use this case is only valid for the duration
of the program.

Copyright © 2017, Indian Institute of Management Calcutta.

 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Javits Center, New York. Batra demonstrated Maya in the conference by engaging in business
dialogues (with Maya). The demonstration looked promising and kick-started Maya’s launch in
the retail industry. The 30-member Manthan’s AI R&D team of majors in Natural Language
Processing (NLP), Machine Learning, and Mathematics and Statistics had utilized over $7.5 million
and spent 3 years to build the disruptive innovation, Maya1. It was commended as the ‘coolest
innovation of the year’ in the NRF, 2017 conference. Atul Jalan, Chief Executive Officer of
Manthan, ascertained that Maya was the first and foremost virtual assistant in the world to be
integrated with business information systems, and was virtually indistinguishable from magic. He
envisaged Maya to lead the radical transformation of business operations to zero-touch user
interfaces and voice-based interactions.

At this juncture, Batra had to ride the expectations of his CEO by marketing the product
appropriately to all the C-suite positions (for example, chief executive officer, chief financial
officer, chief operating officer, chief information officer and chief marketing officer). However,
Batra was aware of the challenges ahead. He opined:

“Since we launched Maya, we have spoken to 10-15 top analysts in our industry globally,
and they have acknowledged this as a unique product ... however when you have such
disruptive technology, adoption is not a given. Now that we have opened a can of worms,
the challenges are that people’s expectations have skyrocketed. They reckon Maya to be
able to answer any kind of questions. In order to meet such expectations, you have to deal
with domain specific data and algorithms, but there is no one-size-fits-all approach.”

Some of the fundamental questions that cluttered Batra’s mind were as follows: How could Maya
disrupt the technological arena? How could Manthan market such an innovation? Was the
market ready for such an AI-induced product? If yes, which market segment and roles would be
most responsive to Maya? How could he build awareness for such product and show value to the
customers? Did Maya have the capability to cater to the legacy of innovativeness?

INDUSTRY BACKGROUND – INTELLIGENT VIRTUAL ASSISTANT (IVA)

The origin of the voice assistant dated back to early 1960s when IBM launched the primitive voice
assistant with its Shoebox device that could recognize 16 words and 9 digits. It laid the foundation
of the pre-modern era of voice and virtual assistants. Subsequently, Dragon introduced Dragon

1
Manthan’s Maya can answer CEOs’ questions within seconds, Business Line, The Hindu, January 30,
2017, http://www.thehindubusinessline.com/info-tech/infotech/article9512568.ece

2
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Dictate, the first competent speech recognition and transcription product for the consumers. It
was followed by Microsoft’s text-based virtual assistant, Clippy which did not gain popularity yet
was allegorical towards the evolution of natural language assistants through interactive feedback
mechanisms.

Later, in the year 2010, Siri was launched as an app for iOS which marked the beginning of
contemporary voice assistants through the integration of smart phones and voice assistants. It
was a spinoff from a project instituted by the SRI International Artificial Intelligence Center. The
Siri app was acquired by Apple two months after its launch, and it performed a wide range of
voice queries, including executing phone actions, scheduling events and reminders, searching the
internet, managing the device settings, searching basic information, navigating areas, and
communicating with iOS-integrated apps2. Siri enjoyed a wide following, and soon it was
succeeded by similar technological breakthroughs by Google Now, Microsoft’s Cortana, and
Amazon Alexa. Amazon Alexa, launched in 2014 was first used in the brand of smart speakers,
named Amazon Echo which were capable of voice interactions, streaming podcasts, music
playback, playing audiobooks, scheduling alarms, package tracking, movie listings, calendar
management and providing weather, traffic, and other real-time information3. The detailed
timeline reflecting the history of voice revolution has been provided in Exhibit 1.

Gartner had predicted five technology shifts that would bring about major transformation in both
marketing strategies and organizations. One of it was the virtual agent that Gartner envisioned
would participate in the majority of commercial interactions between people and businesses by
2020. They predicted that the share of mobile search queries actuated through virtual personal
assistants (VPAs), such as Siri, Alexa, and Cortana, would rise from 20% to over 50%. Amazon
Echo and Google Home, room-based VPAs would encompass 10 million homes and a large share
of commercial traffic. Marketers were expected to spend over US$ 250 million on conversational
technologies such as chat bots and virtual agents to expedite and optimize the customer
journeys4.

2
SIRI RISING: The Inside Story of Siri’s Origins – And Why She Could Overshadow The iPhone, Bianca
Bosker, Executive Tech Editor, The Huffington Post, 2013, http://www.huffingtonpost.in/entry/siri-do-
engine-apple-iphone_n_2499165.
3
Alexa Voice Service Overview, https://developer.amazon.com/docs/alexa-voice-service/api-
overview.html.
4
Predicts 2017: Marketers, Expect the Unexpected, Gartner Inc., 2017,
https://www.gartner.com/binaries/content/assets/events/keywords/digital-marketing/gml3/gartner-
2017-marketing-predicts.pdf.

3
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

The advancement in natural language processing had been instrumental in establishing the
power of conversational Artificial Intelligence (AI) solutions across different consumer
touchpoints. The prevalence of tools like Apple’s Siri and Google’s Google Now had ascertained
the consumer engagement with the NLP solutions. Also, there was 20% reduction in service calls
when Aetna initially introduced its virtual agent tool in its online registration process 5.
Subsequently, this figure went down to 50% when the company remodeled the tool and asked
the customer to restate questions if the virtual agents failed to answer the initial question. Nearly
one in five visitors used the virtual agents. According to a survey conducted by Forrester, all
vendors and organizations acknowledged that the barrier to adoption of virtual agents was
gradually on the decline and the well-managed tools were achieving greater than 90% success in
addressing customer inquiries6.

Tractica Research predicted that users of unique active consumer IVAs would grow to 1.8 billion
worldwide by the end of 20217. This pace of adoption would be catapulted by several factors.
Firstly, IVAs had wide access to multiple devices interacting with the Internet of Things (IoT),
which included smart homes, smart automobiles, and smart wearables. Such characteristics
made it ready for daily use thereby increasing the chance of user adoption. Secondly, myriads of
robust APIs were available that could easily integrate to the back-end systems to access the data
and enable personalized customer experience. Thus, the integration of virtual agent programs to
the existing systems in organizations would be hassle free. Thirdly, virtual agents facilitated with
simplified knowledge management interfaces could be managed and maintained at ease with
fewer resources. Lastly, the growth and advancement of disruptive technologies such as cloud
computing, analytics, mobile, NLP, and high-speed internet were making the IVAs affordable to
the organizations as well as to the individuals8.

Conversational AI solutions had addressed the “full services versus self-service” dilemma by
providing full-service experiences through self-services processes. However, according to Frost
and Sullivan, many of the business users found that the lightweight, general purpose virtual
assistants such as Apple’s Siri were incapable of supporting and performing complex or niche

5
The Business Case for Virtual Agents, Brendan Witcher, A Forrester Report, 2014,
https://www.forrester.com/report/The+Business+Case+For+Virtual+Agents/-/E-RES116386.
6
Ibid.
7
The Coming Intelligent Digital Assistant Era and Its Impact on Online Platforms, cognizant 20-20
insights, 2017, https://www.cognizant.com/whitepapers/the-coming-intelligent-digital-assistant-era-
and-its-impact-on-online-platforms-codex2447.pdf.
8
Ibid.

4
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

tasks9. Some IVA providers heavily integrated technology to businesses of the companies they
served and Next IT was one of the them. Amtrak’s Ask Julie, Aetna’s Ask Ann, and the US Army’s
SGT STAR were powered by IVAs developed by Next IT and they delivered substantial value to
the organizations. In particular, Amtrak’s Ask Julie (Amtrak Corporation was the company
providing passenger railroad services within the contiguous United States) had provided 8X ROI
and was reducing the expenses for the railroad by $1 million per year in customer service
emails10. It had generated 25% more bookings, with 30% increase in revenue per booking.
Therefore, the industry related to IVAs had an immense scope of growth and development
through the ubiquity of natural language processing in every sphere of technological
advancement. The pace of change in the IVA space was rapid and accordingly the entire business
ecosystem was undergoing a paradigm shift. This entailed an era where IVAs were disruptive in
supplanting the conventional way of information seeking for decision-making.

COMPANY BACKGROUND – MANTHAN

In 2003, Jalan was living in London and introspecting about business away from software services;
he thought about ventures ranging from robotics to ski resort. One day, he created a folder
named ‘Manthan’ – metaphor for deeper contemplation – for collating his own ideas. He figured
out that math and data science were going to shift the paradigm paving the way for the next
Industrial Revolution11. He moved to Bengaluru and began to explore data science. Finally, in
2004, he founded the business intelligence and analytics company, Manthan Systems. In 2005,
Manthan secured the seed round through a debenture-based funding from Small Industries
Development Bank of India (SIDBI). Later in 2008, it received Series-A funding from IDG Ventures
and DFJ ePlanet Ventures. After that, it raised Series B, C, and D from Fidelity Investments,
Norwest Venture Partners, and Temasek Holdings, contributing a total of about US$ 100 million.
Over the years, Manthan grew at an unprecedented rate with a strong portfolio of analytics
products and solutions. As on 2017, Manthan served the Chief Analytics Officers to over 200
leading retail and consumer group organizations across 21 countries12.
In the words of Batra:

“Manthan has, from the very beginning, invested in building its own technology platforms;

9
Global Conversational A.I. and Intelligent Assistants Technology Innovation Award, Frost and Sullivan,
2016, http://www.nextit.com/downloads/Next-IT-Award-Write-Up.pdf.
10
Ibid.
11
Of Manthan and how Atul Jalan built a multi-million-dollar analytics business, YourStory, 2017,
https://yourstory.com/2017/06/manthan-atul-jalan/.
12
The Manthan Story & Analytics, https://www.manthan.com/manthan/the-manthan-story.

5
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

one of them is the analytics platform which is well recognized and has obtained several
accolades related to excellence in IT and analytics … the analytics platform has a competitive
and differential advantage. Many years back we also rode the wheel of cloud computing by
implementing analytics on the cloud. Across the years, we have perfected it, and now pretty
much every piece of software in the industry is being consumed over the cloud. So, that is
one of the early moves we made, and we did it in conjunction to leveraging big data.”

Manthan had developed an unparalleled understanding of the retail industry and the industry
constraints across geographies. This intricate level of understanding provided them an
opportunity to convert the advancements in cloud computing and big data analytics into
actionable business solutions. Manthan mastered all types of AI-powered analytics for all classes
of business users in retail enterprises. The different types of analytics that Manthan had dealt
with over the years to gain a strong footing in the industry has been presented in Exhibit 2. Such
prowess in business intelligence and analytics had given Manthan several awards and
recognitions including “Gold in Analytics” at Express IT Awards 2015, Eight #1 Rankings in the
2016 RIS Software Leader Board, Frost & Sullivan “Company of the Year in Analytics” award in
2016, IDG Ventures Exceptional Business Growth Award, and Innovation Partner Award by AWS
in both 2016 and 2017.

MAYA – AN INTELLIGENT VIRTUAL ASSISTANT FOR BUSINESS EXECUTIVES

Speaking about the recent technology innovations and the trend of Industry 4.0, Batra said,

“…what we are hearing from some of the top technology people is that one of the biggest
disruptions will be the voice interface revolution where we will be talking to all kinds of
devices such as smartphones, smartwatches, etc. through voice … what Manthan is doing is
to incorporate the technical features from natural language processing capabilities in the
consumer world - like Siri and Google Now - into the business world, and building the world’s
first conversational interface powered by AI for complex business analytics.”

With this vision, Manthan developed Maya, a conversional analytics application supported by
NLP, intent analysis, deep machine learning, residing in Manthan’s analytics platform. It used a
voice interface to communicate with information systems and answer complex business queries
in the natural language to a CXO wanting to know why company’s sales declined last weekend or
how the brand sentiment was for a particular product. The examples of the different kinds of
questions that the initial version of Maya was capable of answering is provided in Exhibit 3.
Therefore, it curtailed the business user’s dependency on IT experts or analysts for business
insights to a great extent. Also, it was integrated with Amazon Alexa for voice-based interactions,

6
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

and it learnt on the go, thereby capable of delivering personalized answers on the basis of user’s
role and context. An equivalent text interface has also been developed. Malaya Kumar
Mohapatra, Principal architect at Manthan and his team were instrumental in developing Maya.

Maya was an AI-powered conversational engine which addresses business questions by not just
merely picking up available information from the database, but also inspecting a variety of
sources, pooling relevant information, analyzing it, applying business intelligence, and coming up
with the best answer appropriate for the context. Exhibit 4 examines the purpose that Maya was
going to serve in the business domain.
IT ARCHITECTURE OF MAYA AND INNOVATION THROUGH DISRUPTION

Maya embodied two basic components –Alexa integrated to Amazon Echo and Manthan’s
Business Intelligence (BI) & Analytics Platform. Alexa parsed the natural language speech into
text and a set of intents. Intent understanding focused on deciphering the information that a user
wanted a computer to obtain or the action that he/she wanted to take, communicated through
text query or utterances. For example, it resembled the process of converting “Please show the
sales of product A for the month of June” to view. sales(month = ‘June,’ product = ‘A’), a language
that the computer understood. This intent when provided to Manthan’s BI & Analytics Platform,
it interpreted and analyzed the content to produce the desired results.

The logical architecture of Mathan’s BI & Analytics Platform rested on the data, which was
extracted from multiple sources, such as social media data, weather and traffic data resembling
non-business generated data available on the cloud often referred to as cloud data, business
generated data coming from business systems like finance systems, Enterprise Resource Planning
(ERP) systems, and Loyalty systems, or Customer Relationship Management (CRM) systems.
Therefore, the platform contained a data integration layer where data from various sources were
curated and integrated into this platform either through batch processing or real-time
processing. Once the data was available on the platform, the data transformation layer
aggregated, transformed, and harmonized the raw data to business logic through facts, numbers,
and metrics linked to business. Subsequently, the transformed data loaded into the data lake or
data warehouse. The data lake represented a storage repository facilitating the organization of
data in various structural forms and also low cost storage for large volumes of data.

After forming the construct of the data lake or data warehouse, datamarts were created. These
data marts consisted of the final, transformed version of the data which was reflective of
functional aspects of the business and was aligned with a specific business line or team. For
example, there could be a finance data mart consisting of all transformed data from finance
systems. Also, there was another layer of technical and business metadata lying on top of data

7
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

marts. It represented the semantic layer over the marts characterizing the type of data, the
business purposes it served, and how it had been organized. On top of this, there was
visualization, reporting, data discovery, and algorithmic toolkits. So, the entire architecture as
explained above defines Manthan’s BI & Analytics Platform. The diagrammatic representation of
the same has been provided in Exhibit 5. There were different products such as Retail Analytics,
Customer360, TargetOne, Vendor Link which sit on this platform. Similarly, this platform also
formed the underlying foundation of Maya.

Another intriguing aspect of Maya was how the intent analysis was done and imparted to the
Manthan’s BI & Analytics Platform. The intent that was derived from the Alexa Skills Kit was
converted to a tokenized language. Further, this tokenized form of intent was consumed by a
Lambda function, which is an Amazon Web Service (AWS) server-less compute component. Since
the intent was generated from the voice platform susceptible to a lot of noises or deformities,
the Lambda function applied machine learning techniques on it to accurately comprehend the
semantics of the speech. For this purpose, Manthan had developed the capabilities of
probabilistic matching, fuzzy logic, etc. within the Lambda function. Also, integrated to Lambda
was DynamoDB, a NoSQL database which stored the user session linked to Alexa. This made Maya
aware of which user was logged into the device, what was his/her profile, and what were his/her
privilege rights.

The output of the Lambda function was passed to a well-formatted RESTful API endpoint which
pushed the request to the Manthan’s BI & Analytics Platform. The platform retrieved the data
from the cache memory if it was available there. Otherwise, it looked in the data warehouse to
address the business queries and surfaced the data out in the form of voice. The architecture of
Maya reflecting on the entire path of execution of the utterances is provided in Exhibit 6.

Beyond the contemporary architecture of Maya, Manthan planned to incorporate several

advancements to improve the efficacy and effectiveness of the product. According to Vijay
Chidambaram, VP of Cloud Engineering for Manthan Systems:

“What we want to do is figure out the scope of device interactions, where I am talking to a
browser, I am talking to a laptop, or I am talking to a mobile phone…it [Amazon Lex-Polly]
allows you to do natural language text search, like how you probably use Google. When I
am asking Maya on a text interface, we want to get the graphs and visualizations pertaining
to that. That is something we are doing using Lex-Polly.”

Apart from the implementation of Lex-Polly to drive in-application text and voice conversations,
Manthan was innovating Maya’s capabilities in answering compound questions consistently as
well as making the capability even more conversational and context-rich. Also, they were working

8
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

on several developments such as voice biometric based authentication, sharper natural language
to metadata matching, Maya notifications on anomaly detection and so on.

The entire architectural overview of Maya threw light on the innovative mindset of Manthan’s
business. The ethos of the organization was conducive to cutting-edge technology products. In
order to describe the cultural dimensions that promoted advanced technology innovations within
Manthan, Lyndon Saldanha, VP of Human Resources for Manthan Systems said:

“Our value propositions in terms of how we are as a company today or have been over the
last seven years lies in the fact that we are able to straddle between the ability of the big
and the agility of the small. On one hand when I say the ability of the big, I am talking about
the intellectual wealth under the roof of Manthan, and on the other hand, the agility of the
small reflects our size. We are not an overly large company where decision making can be
long, arduous and can possibly slow things down.”

He also added:

“We have the ability to provide work that is a product but not service-oriented; there is an
element of introducing inventive work, bringing new ideas to the fore, organizing lots of
brown bag and open house sessions. What we try is not to box the employees, rather we
provide a fair amount of flexibility in work settings which allows people to adjust according
to their comfort.”

Therefore, the trend of innovation at Manthan was the culmination of the proper orientation of
human resources to collecting disruptive innovation ideas and putting them into implementation,
as well as the practice of some integral elements of culture such as risk-taking, flexibility,
creativity, and entrepreneurship. Apart from this, the organizational structure and the
inculcation of right thinking through appropriate allocation of resources augmented its
competency to innovate. Manthan had continued to file patents in large numbers over the years
to showcase their innovating ability. As a result, Manthan grabbed several awards, the prominent
and the most recent of them was the Innovation Partner Award at AWS Summit 2017 for
leveraging the AWS technologies and spanning Maya.

MARKETING OF NLP AS A COMMERCIAL PRODUCT

The introduction of Maya signaled Manthan’s penetration into the high technology NLP space.
Such superior technology achieved marketplace success by collecting, disseminating, and acting
on market-based innovation. Often, technological innovations ran into the risk of not

9
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

understanding the needs and capabilities of the intended user. Maya was a product targeted to
high-tech firms and needed a proper market orientation to attain superior sales growth and
profitability. This required competence in three essential activities – opportunity identification,
product commercialization, and product and process innovation. All these activities required
strong marketing capabilities to discern a wide range of markets and develop customer
applications using the disruptive technology. Manthan was always capable of identifying the
target audience for its product portfolio and innovate according to their needs. Reflecting on this,
Sameer Narula, Chief Product Officer of Manthan Systems said13:

“We are delighted to be named a visionary leader by MarketsandMarkets. We believe AI

and prescriptive analytics are shifting the way retailers use data and act on insights,
enabling precise, intelligent and timely decisions across every retail function. These are key
differentiators for Manthan’s offerings and compelling reasons for modern retailers to
adopt next generation analytics solutions as they seek to unify retail customer experience,”

The Retail Analytics Market Global Forecast for 2022 by MarketsandMarkets showing the
competitive leadership map of the retail analytics market is provided in Exhibit 7. When asked
about their strategy in creating market awareness and understanding the market response for
the product Maya, Vikram Murdeshwar, Senior Product Marketing Manager of Manthan Systems
remarked:

“Our approach has been to firstly reach our existing customers and offer them a demo
service, that is to provide them Amazon Echo and an application that they can just turn on
in no time and get started. That is our first priority, and in terms of building awareness for
it [Maya], what we have done is to use the PR [Public Relations] route, because not everyone
is aware of Maya; everyone is aware of Siri, but in terms of business to business application
of voice platform, we are probably the first in the space.”

Apart from leveraging PR, Manthan organized roadshows, where the company demonstrated
what AI does. They showcased the contribution of AI in various industries and explained how it
could help organizations and finally moved on to introducing Maya. Some organizations
attending the roadshows had been quick to imbibe the concept and agreed to enable their users
around Maya; however, some had been reluctant in adopting the innovation at this point of time.
Rather, they suggested that a text-based conversional agent may suit their purpose.

13
Manthan Named a Visionary Leader in the Retail Analytics Market on the Strength of its Product
Portfolio, Santa Clara, 2017, https://www.manthan.com/manthan/pressreleases/1203-manthan-named-
a-visionary-leader-in-the-retail-analytics-market-on-the-strength-of-its-product-portfolio

10
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

As part of the marketing drive, Manthan’s digital marketing team led by Santosh Kumar ran social
campaigns, spread brand awareness, and established a direct connect with C-level executives
through LinkedIn and Twitter since these people resembled their target audience. Exhibits 8-10
provide some creatives related to digital marketing. In addition, gaining the recognition from top
technology journals/magazines writing about the recent AI trends and developments was
another way of reaching their audience. In this way, they could spread the news about Maya’s
capabilities among the technologically savvy audience. Also, to highlight the different aspects of
Maya like NLP, the underlying machine learning algorithms, and use cases to the marketers or
merchandizers, Manthan was in the process of releasing blogs, articles, and white papers. The
team also ran email campaigns for existing as well as target customers. Though the open rate of
emails was decent, the click-through rate was unexpectedly low. The people who responded
were asked to try a public instance of Maya listed as an Amazon Alexa skill on their own Amazon
echo devices to get a first-hand experience. Also, they showcased Maya at events such as
NASSCOM and AWS Summit, and received a good response. This showed that most of the
marketing strategies undertaken for the product Maya came under the purview of direct
marketing. Interestingly, most of the direct marketing was aimed at a specific set of customers.
With respect to this, Murdeshwar remarked:

“Our campaigns are initially focused on a set of companies that have already reached a
certain maturity in terms of their analytics adoption to consider using Maya. We looked at
$1 billion plus companies that are extensively using BI and capturing large volumes of data
from a variety of sources and narrowed it down to industries like retail, media and
entertainment, hospitality, and telecom. We wanted to reach the Chief Innovation Officers
(CIOs) or Chief Digital Officers (CDOs) who are looking for transformational technologies
that will be beneficial for their business.”

Even among the target audience, there were some that had strong AI mandate and were eager
to implement AI based solutions. Hence, they would be more receptive to Maya. However, there
were others that lacked a deep understanding of AI and therefore needed a different treatment.
Each of these groups had to be tackled differently. Reflecting on the target market in terms of
geography, Murdeshwar said:

“Our primary geography for Maya initially is the United States and India. Those are the two
regions where Manthan has deployed the strongest team to communicate with their
customers. Other than these two regions, we have an established presence in the Middle
East and South-East Asia. We are yet to tap into Europe for Maya specifically, but it seems
like it is increasingly starting to become an interesting space for us because the number of
AI related events/summits and other things around AI suggest that the region is promising.”

11
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

North America formed the key market for Manthan. So, it had a strong sales team of 26 folks
covering east coast, west coast, mid-west, parts of Latin America, and Canada. It was also
establishing a dominant presence in the markets of India, Europe, and the Far East with sales
force cutting their teeth in a similar environment. In each geography, the sales team comprised
mostly local people who were familiar with the needs of the market and customers.

WHAT’S NEXT?
Manthan was faced with several challenges on the road ahead. In terms of innovational aspects
of Maya, customers considered the voice output as a limited use case since they could not
comprehend the output of complex algorithms through visualizations, graphs, charts, etc.
Therefore, the customers demanded the developments in the upgraded version of Maya so that
they could surpass such limitations. The wave of innovation achieved through the initial version
of Maya had raised the expectations within the business communities to such a degree that they
imagined Maya itself capable of bringing in a digital transformation to their business. Amidst this,
there were manifold challenges towards marketing – on one hand, Manthan had to create
awareness among customers, and on the other hand, they had to provide right impetus to the
product by knowing what was on the minds of the customers. Naveen Khattar, Product Manager
for Maya said:

“They [Customers] have many questions when you go to them such as, how can AI help my
business apart from the hype that it has created, who else is using this, what are the pre-
requisites, how long will this take to implement, what metrics can I expect if I implement.
So, there are hundreds of questions that you need to answer, and I think the results of our
POCs, case studies and to some extent articles and blog will provide answers and allay some
of the skepticism.”

Manthan had customers who were amenable to the voice-based approach for their business
analytics through virtual assistants because their business users did not prefer to indulge in active
analytics. Such customers presented a conspicuous understanding of their business needs which
could be comprehensively solved by Maya. However, Manthan had an onerous task of
understanding the latent needs of different segments of customers and also explaining to them
the improved metrics of the shift to the voice based system involving analytics-based decisions
with more accuracy and less gut.

In spite of all these challenges, Batra was exuberant about the future. The reasons he explained
were in the response he received from the business groups. He recalled:

12
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

“One customer, a CEO said that he wants to take it [Maya] to his board meeting so that the
board members can ask all kind of questions related to how business is doing. They can
interact directly with data without digging into a pile of reports and documents. What better
can a situation be than this!”

Therefore, he prognosticated:

“Maya would basically become everything for Manthan; it would become a face of
Manthan, it would be how Manthan is consumed … eventually it should become a business
advisor, it would become your colleague helping to run the business.”

Batra’s vision looked promising but was intertwined with uncertainties. In terms of the potential
market for Maya, the obvious questions were – What needs might be achieved by the new
technology? How would the needs change at a later point in time? How fast would the innovation
flourish? Would the new product function as had been promised? How would they spread the
innovation? How large was the potential market and would it adopt Maya as an industry
standard? Taking Maya forward amidst all of these technological and market uncertainties was
what kept Batra intrigued and engrossed.

13
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

EXHIBITS
Exhibit 1: A Short History of the Voice Revolution.
Source: voicebot.ai.

14
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Exhibit 2: Types of Analytics Handled by Manthan.

Source: Manthan Systems.

Exhibit 3: Maya Version 1.0 for CEOs.

Source: Manthan Systems.

15
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Exhibit 4: Purpose of Maya

Source: Manthan Systems.

16
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Exhibit 5: Manthan Analytics Platform Logical Architecture.

Source: Manthan Systems.

Exhibit 6: Architecture of Maya.

Source: Manthan Systems.

17
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Exhibit 7: Competitive Leadership Map of Retail Analytics.

Source: www.manthan.com.

18
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Exhibit 8: Banner for Sponsored Post on Facebook.

Source: Manthan Systems.

Exhibit 9: Maya Questions Card 2017

Source: Manthan Systems.

19
 IIMC-CRC-2017-23
Manthan’s Maya: Assisting Executive Decision Making Through Voice Analytics

Exhibit 10: Graphics of Maya

Source: Manthan Systems.

20