1. Examples, Trends, and Potential for Application
Daniel Faggella, CEO at TechEmergence
AI in Retail - Where it Matters
@danfaggella
2. Background Brief
I’m Dan Faggella, CEO/Founder atTechEmergence.com
We’re a market research and media firm with one goal:To cut
through hype and show business leaders the implications,
applications, and important companies in artificial intelligence.
We have business readers all over the world (biggest following
in SF, NYC, Bangalore, London).
@danfaggella
3. Outline of the Talk
1. The State of AI in the EnterpriseToday
2. Forward-LookingTrends of AI in Retail
3. Issues with Applying AI in Retail
4. Becoming “Hype-Proof” - What to Pay
Attention to in the Future
@danfaggella
4. Why it Matters
• Machine learning will likely overhaul entire industries in the next 15
years (will be essential in security, customer service, marketing, BI/
analytics)
• We interview hundreds and hundreds of execs and researchers, and
while these all have different opinions on timelines, they agree on the
inevitability of AI transforming industry, much the same way the
internet did
• AI will impact your business, but you need to know what to pay
attention to.You will be bombarded with hype and news about AI and
you should know how to “prime your antennae”, paying attention only
to what matters, and that’s what I’ll help you with in this presentation
@danfaggella
5. The State of AI in Business
• Make no mistake about it: It’s mostly pilots, testing (not
concrete ROI)
• For every 100 “AI companies”, we’ve found that only 1/3 is
actually leveraging AI in any serious way, and only 1/3 of those
companies are past the stage of “piloting” their product or
service (Maybe 1 in 10 “AI” companies is actually selling
something that has had a positive impact on a business)
@danfaggella
6. What Smart Executives Do
• Stay on top of what the retail GIANTS are investing in for AI,
and pay attention to what actually generates ROI
• Assume that 90% is PR garbage, 10% may have promise
• AtTechEmergence.com this is the research we do every day,
this is our sole focus (Case studies, industry comparisons, etc)
@danfaggella
7. Everyone can load their data into Hadoop… but doing something with
that data (in terms of ML applications) often reveals structural problems:
• No ML talent in place (and consultants can’t do it all for you - see
Machiavelli’s quotes on auxiliary troops)
• More importantly, no management structures in place to (a) deal with
uncertainty of AI applications (there is no guarantee on if they will
work, or when), (b) deal with the lengthy R/D process of AI, (c) getting
buy-in or understanding from the top
• Some vendors / consultants I’ve spoken with think that acquisitions will
be the way that enterprise innovates, not R/D
Biggest Challenges
@danfaggella
9. Use-Cases of AI in Enterprise
Note that retail is not listed specifically, though customer service and
marketing applications do apply to more retail businesses.
10. Retail - Meta-Trend
• Amazon / Online Giants: Perfectly instrumented environment,
all behavior tracked, all features customize-able for customer
preference and improving customer lifetime value
@danfaggella
“Online” Capabilities in the “Offline” World
11. Retail - Meta-Trend
• Offline Giants:Trying to catch up to online, by using many of
their same tactics in the “real world”, including:
• Testing / adjusting in-store product placements (for
maximizing sales, or margins)
• Using cameras to determine customer sentiment around
products, or track behavior and movement patterns
• Recognizing customer identity to prompt them to the best
offerings when they enter the store, or to follow-up with the
best offers after they leave
@danfaggella
“Online” Capabilities in the “Offline” World
12. Retail - Meta-Trend
What Does This Mean?
• “Smart” vendors and service providers will develop in the
ecosystem (Machine vision / camera companies, consultancies
that help set up agile testing / marketing strategies)
• ^ Much of this will be in computer vision and marketing
tech
• Vendors will increasingly win by offering retailers ways to
collect data or work along side their data collection efforts
@danfaggella
“Online” Capabilities in the “Offline” World
13. Retail - Meta-Trend
What Does This Mean?
• Vendors Might Ask: How does (or how could) my products help retailers
collect date / optimize in-store experience? Examples:
• Is the product easy to move around the store, or can it be assembled in
multiple ways to test customer response?
• Is the signage easily replaceable or changeable, or is the display digital so
that many messages / prices can be tested?
• Worse case:Are you at least familiar with the technologies / initiatives of
the customer and you can tailor your service to help reach those aims?
@danfaggella
“Online” Capabilities in the “Offline” World
14. Retail - Meta-Trend
Future Peek:
• Eventually (maybe in 8 years), all big-box retailers will need to be
optimizing their marketing and their in-store experience if they want to
stay competitive.This might include:
• Understanding the interactions / products that seem to correlate to
customer happiness (by using facial recognition)
• Understanding the in-store experiences that tend to garner repeat
visits for further purchasing (by tracking customers at an individual
level)
• Customizing layouts and offerings per each store geo-location,
demographics, seasons, or even based on weather or day of the week
@danfaggella
“Online” Capabilities in the “Offline” World
15. Retail - Meta-Trend
• A few points to bear in mind:
• It will mostly be LARGE retailers (Macy’s, Best Buy,Wal-
Mart) that implement AI in the next 5 years, not SMBs
• Camera technology (and “machine vision”AI tech) will be
among the most important AI technologies, because it
helps to create those “instrumented” environments (for
inventory, for product placement, theft prevention, etc)
@danfaggella
“Online” Capabilities in the “Offline” World
17. Retail - Trends (1)
• AtTechEmergence we don’t believe that retail robotics will be
viable or worthwhile in next 5 years.We believe that the
“instrumented retail environment” trend will be much more
meaningful in changing the retail experience
• Warehouse robotics (like those used at Amazon) hold
greater near-term promise
• Simbe Robotics seems to be the most viable for me, because
it is a step-change towards the “instrumented retail
environment”
@danfaggella
Retail Robotics
18. Retail - Trends (1)
What Does This Mean?
• In the future, isles and products may be arranged so that they
are accessible (visually) to robots and more advanced camera
systems installed store-wide, not just to humans
• Until some BIG stores have profitable applications of this
technology (Lowe’s,Wal-Mart, etc), it is highly unlikely that it
needs to stay on the mind of small or mid-size retailers
@danfaggella
Retail Robotics
19. Retail - Trends (2)
Marketing to an Audience of One
@danfaggella
20. Retail - Trends (2)
Marketing to an Audience of One
@danfaggella
• AI marketing vendor companies see LOTS of promise in retail,
although eCommerce players will likely be quickest to use these
technologies (due to their instrumented environments)
• Technology and service providers will develop in order to
“upgrade” brick-and-mortar retail business marketing
• Store setups will need to accommodate for more testing /
moving of products and arrangements
• Checkout technology will inevitably lead to individual customer
profiles
21. Retail - Trends (2)
Marketing to an Audience of One
@danfaggella
What Does This Mean?
• Transaction-related vendors companies (POS systems / etc)
should be considering how they’re evolving to accommodate
the “audience of one” future
• Some in-store vendors / installers may find themselves selling
the benefits of reconfiguration / mobility / “iteration” of their
products, not merely it’s sturdiness and professional look after
a first installation
22. • Think about AI adoption the same way you would think of
adoption of any other emerging technology that you consider
to be essential to the future of your industry.You might test
and try with a little most gusto because AI is indeed
inevitable, but don’t be rash for “FoMo” sake!
• AI adoption should involve an informed, forward-looking
industry / competitive analysis (like the kind that you’d do at a
quarterly off-site), nothing less.
Where AI “Should” Be Used
@danfaggella
23. • Common cardinal sin:“Toy” applications
• “Toy” applications are technologies or projects taken on because
they use AI, not because they solve a business problem.Vendors
play into this because they need guinea-pigs to “pilot” products,
and they’ll sometimes encourage closing deals even if they aren’t
well organized
• They almost all end the same way: Lacking resources to back
them, lacking gusto to carry them through, and negatively
impacting the funds and human resources of the company (and
making the “toy” initiator into a fool).
Cardinal Sin of AI Applications
@danfaggella
24. Adopt or Wait?
• For any given AI application area (marketing, business
intelligence, procurement, etc), determine where you want to
be on the adoption curve (FEW established firms must be or
should be “innovators” or even “early adopters”)
@danfaggella
25. Concluding Thoughts
Priorities for executives:
1. Understand what kinds of problems are solve-able with AI
2. Consider the domains within your business that can most
benefit from the use of AI
3. Talk to companies who have developed and implemented
similar applications and get a realistic understanding of what it
would take to implement them yourself
4. Then you can make your decision on whether or not to invest
in AI, to develop AI, to acquire AI, etc…
@danfaggella
26. That’s All, Folks
Email info@techemergence.com and I’ll reply in the coming
week. Feel free to include any relevant questions from this
presentation.
@danfaggella
27. Resources (1)
• TechEmergence articles about retail AI:
• https://www.techemergence.com/artificial-intelligence-retail/
• https://www.techemergence.com/machine-learning-retail-applications/
• https://www.techemergence.com/robots-in-retail-examples/
• Machine learning in marketing:
• https://www.techemergence.com/machine-learning-marketing/
• Applying AI to Business Problems (General Understanding):
• https://www.techemergence.com/how-to-apply-machine-learning-to-business-
problems/
28. Resources (2)
• https://hbr.org/2017/04/the-first-wave-of-corporate-ai-is-doomed-to-fail
^ Quote from this article:“We believe AI will indeed transform industries. But the companies that
will succeed with AI are the ones that focus on creating organizational learning and changing
organizational DNA”
• https://hbr.org/2017/04/how-companies-are-already-using-ai
^ Good article, but author is downplaying the job automation concerns of AI.All big, bloated
consulting companies do this, be wary of people-heavy companies assuring everyone that AI won’t
replace people.
• http://www.gartner.com/smarterwithgartner/artificial-intelligence-and-the-enterprise/
^ Most relevant part of this article is the third question “How will AI impact the talent needs of an
organization?”
• https://hbr.org/2017/06/if-your-company-isnt-good-at-analytics-its-not-ready-for-ai
^ Extremely useful perspective on the “baby steps” needed to begin working with AI seriously.