AI in Auto Industry:

Harbingers of Change In the Driverless World, In-Transit Innovation will define

market leadership

MUMBAI | SATURDAY, 16 SEPTEMBER 2023 Level 4 and 5 autonomous vehicles now have become
more prevalent. This has lead to increased personal
Indian Automotive Industry has now grown to the productivity for car occupants, who would be free to spend
size of $254 Billion clocking a CAGR of ~15%. their transit time focusing on activities other than driving.
Autonomous driving has evolved with growing use
of Artificial Intelligence (AI) as depicted in the table
Experience over Product Ownership
below. Successful vehicle makers have already
The customers today demand more from the mobility
launched their own connected vehicle platforms.
experience, and value experience over product ownership.
According to SIAM there are about 15 Mn
OEMs are focussing on car design, development and
connected vehicles are plying on the road in India
marketing around customized in-car experiences. OEMs are
today.
taking a deep dive into technologies that enable cars to
The Autonomous Vehicle Evolution using AI: taking care of both their own maintenance needs and the
needs of their occupants during the drive.

Level 0 The driver controls everything inside

Upto the vehicle – steering, brakes,
2010 throttle, power

Level 1 The Driver mostly has control over

(2011- everything, but some specific
2014) functions are performed
automatically by car such as steering,
accelerating
Level 2 The car includes at least one
(2014- automated driver assistance system
2018) for steering and
acceleration/deceleration, using
information about the driving A key
environment. The driver must still feature of this evolved business model would be the use of
always be ready to take control of the
deep insights into customer behaviors, to design a plethora
vehicle when needed.
of services that are truly hyper-personalized. Examples
Level 3 Drivers are still necessary, but critical include cars with automated-payment interfaces,
(2019- safety functions can be completely
augmented reality capabilities or work collaboration
2021) shifted to the vehicle.
features. Different players will leverage the emerging
Level 4 Fully autonomous, vehicles are opportunities to different extents, depending on their
(2021- designed to perform all critical safety readiness for the new era and strength of their partnerships
2023) driving functions and monitor road (see Figure)
conditions for entire trip.
Level 5 A fully-autonomous system enables
the vehicle to perform at the same
(2023- level as a human driver, in every
driving scenario. Care are developed
without steering wheels or pedals
 Artificial Intelligence is Transforming Automotive collected from every vehicle – trips, telemetry, engine
Industry Monetization Models: RPMs, speeding up and slowing down, accidents, etc. –
should be used to make people’s lives easier, including
AI is based on experience creating the Connected users and industry professionals.
Car Platform, which obtains data from hundreds of
thousands of connected drivers. All the information The automobile development cycle is long enough for AI,
big data analysis, and machine learning to enrich
the product-planning process. Data collection and
The role of AI
analysis provides clear direction about the nature of
Carmakers have been collecting vast amounts of customers and their needs as consumers.
data on traffic conditions and human responses. These solutions can be monetized through the sale
Data such as avoiding obstacles and collisions, on of services, accessories, and vehicles.
how people react, on what a driver is likely to do
when faced with a certain situation, on road AI can tell when it’s a good time for a driver to
replace their car, understand that their lifestyle has
networks and weather conditions.
evolved, and offer them a new car pre-customized
Once the data is collected, it needs to be to meet their needs. By using AI we can analyze and
processed in a dynamic fashion. AI algorithms predict drivers’ consumer behavior (what they like
to do, how they like to use their cars) and, most
come into picture for just that purpose. Self-driven
importantly, understand their future needs and
cars (and even assistive features) are controlled by
what kind of car or accessory they’re going to need
algorithms that have learnt from processing the tomorrow if their lifestyle changes. For example, if
massive amounts of data collected by carmakers. Is a driver has a baby, we need to offer them a
a cyclist about to suddenly turn in front of your somewhat larger car.
car? Will the truck on a highway onramp wait for
In addition, AI and machine learning will allow OEMs
you to pass before he merges? If it’s raining, is the to control planning features in the car (buttons,
cab in front of you more likely to brake abruptly? displays, and indicators). Big data and big data
analysis can provide concrete information about the
And since we all know from experience how no natures of customers and their needs regarding
specific functions. AI-based industrial analysis and 9
two road situations are ever the same, you can’t forecasting allows OEMs to add or remove certain
set rigid rules. Instead, you need self-learning, features and functional buttons today, identify
adaptive systems that learn from the hundreds of defective component suppliers, and, of course,
reduce costs
millions of miles of driving data that has been
collected, in order to act faster (and in a safer More Than Autonomous Driving:
fashion) than even the most quick-witted human
driver. Sometime in the future, we’ll also see cars While self-driven vehicles might be the most visible
share data, with the AI in your car warning the AI beneficiaries of AI, there’s more to this. From user
in the car behind you that it’s about to brake. interfaces that recognise individual drivers’
preferences, to self-learning diagnostic systems that
Science-fiction? Not at all!
warn of impending failures, to navigation systems
that use real-time data and traffic modelling to make
journeys shorter, to ride-hailing apps that maximise
the utilisation of a cab aggregator’s fleet, change is
happening on all fronts. In fact, analysts at Tractica
estimate that the market for automotive AI could
touch $14 billion by 2025.

Future where carmakers don’t sell cars, but maintain

fleets of self-driving vehicles, allowing customers to
use ride-on-demand services. You’ll hail a cab
controlled by an AI through an AI-driven fleet.
AI Advancement will have multiple implications for offered as–a-service for older vehicles. Car designs will
the Industry Value Chain: move toward representing a moving entertainment hub, a
mobile office or a hotel on wheels
OEMs and suppliers: The mainstreaming of
autonomous cars will force OEMs to rethink and Insurance companies: Autonomous cars promise to
redesign how they develop cars. The car of the introduce massive shifts for insurance companies, as well.
future will be connected with a gamut of services For one, insurers will need to rethink their coverage models.
and infrastructure facilities that consumers will use The statistical algorithms and methods currently in use to
on a daily basis. This will require OEMs to create broad driver categories and averages, which then
collaborate with traffic control, public utility determine coverage and set rates, will be replaced by
services, infrastructure providers, telecom service personal one-to-one profiles that tailor coverage and rates
providers, telemetry equipment providers, as well to one driver. Further, over 80% of collisions that occur
as data analysis and machine-learning service today are expected to be prevented once humans are taken
providers to ensure that in-vehicle hardware and out of the driver’s seat. This will undoubtedly lead to
software integrates seamlessly with multiple casualties in the auto insurance industry, as well as a
systems and platforms massive reduction in fees for the insured, to as little as 10%
of what customers pay now. As competition rises, it will
Automakers will also need to develop partnerships increasingly be a struggle for insurers to find a way to
with entertainment subscription services, such as remain relevant
Netflix or Amazon Prime. Such collaboration could
also provide a common pool of customer data Looking Ahead: Final Thoughts
shared by all players in the ecosystem to customize
Like the consumer electronics industry, the auto market
products and services for specific cus-tomer
segments. The capabilities of these third-party could see demand for a new model of vehicle shift to every
players will need to be considered throughout the three to six months rather than the traditional annual
product lifecycle, from product development to model refreshes and multi-year gaps between new product
delivery designs.

Dealerships: Showrooms will evolve from focusing Dealerships will need to overcome the hurdle of declining
on sales to providing a vehicle usage experience, revenues from traditional vehicle sales and service by
using augmented or virtual reality technologies and developing new business models aligned with the shared
personalizing the experience to the needs and mobility ecosystem. While dealers will continue to serve as
wants of the customer. Post-sale service and sales and service points, they will also be the nodes for the
extended markets of fleet management, subscription
repairs, meanwhile, will focus on software upgrades
and defect corrections, in addition to mechanical services and new vehicle servicing models that will overturn
repairs. With increased self-healing capabilities, decades of precedence in the industry
especially for minor problems, dealer service Overall, the auto industry will undergo an unprecedented
centers would largely focus on customized shift by 2025, changing the nature of every player and
accessories, major overhauls and repairs resulting relationship throughout the value chain. The winners will be
from collisions those that overcome inertia to redefine themselves over
Technology companies: Amid these changes, the next decade
technology players and new market entrants will
play a key role. Reliable plug-and-play systems will
emerge that could be retrofitted in older cars,
enabling autonomous capabilities (see Quick Take,
below).6 Machine learning algorithms7 that read
driving patterns, trip conditions, terrain contours
and navigation could enable self-driving features,