How AI Will Transform Marketing In The Next Five Years · How AI Will Transform Marketing In The...
Transcript of How AI Will Transform Marketing In The Next Five Years · How AI Will Transform Marketing In The...
How AI Will Transform Marketing In The Next Five YearsRob Gregory
What do we want?
AI
When do we want it?
Now!
Coming upHistory of Artificial Intelligence
What is AI (and what it isn't)
The current state of AI in marketing
Privacy and Ethics
Predictions for the future
Your business and AI
AI: A Human Obsession…
Humanoids – Automata - AIGreek mythology | Robots and artificial beings such as Talos, Galatea and Pandora
384 BC | Aristotle described the syllogism, a method of formal, mechanical thought
1206 | Al-Jazari created a programmable orchestra of mechanical human beings
1495 | Leonardo da Vinci sketch – if built could move its arms, twist its head, and sit up
1642 | Blaise Pascal invented the mechanical calculator
1738 | Jacques de Vaucanson - mechanical figures including a gold-plated duck
1818 | Frankenstein by Mary Shelley - ethical questions around creating sentient beings
1863 | Samuel Butler suggested that Darwinian evolution also applies to machines
1941 | Konrad Zuse built the first working program-controlled computers
1950 | Alan Turing proposes the Turing Test as a measure of machine intelligence
1950 | First working AI programs written (chess and draughts)
1956 | Dartmouth College, New Hampshire holds first ever Summer AI Conference
1961 | First industrial robot worked on GM production line
1966 | Negative report on machine translation kills work on NLP for many years
70’s + 80’s | Academic and scientific work around NLP, neural networks, data modelling
1986 | First robot cars, driving up to 55mph on empty streets
1997 | IBM “The Deep Blue” defeats the then world chess champion, Garry Kasparov
2001 | NASA's Spirit and Opportunity autonomously navigate the surface of Mars
2006 | The Dartmouth Artificial Intelligence Conference: The Next 50 Years
2011 | Siri, Google Now (just part of Search today) and Cortana NLP to answer queries
2015
Google DeepMind's AlphaGo defeated 3-time European Go champion by 5 games to 0
2015
Stephen Hawking, Elon Musk, Steve Wozniak + 3,000 researchers in AI and robotics sign
open letter to ban development and use of autonomous weapons
2018
Alibaba language processing AI outscores humans at a Stanford University reading and
comprehension test
2018
Google Duplex books appointments over the phone (with a human)
What is Artificial IntelligenceWhat is Machine Learning
AI vs ML
Artificial Intelligence
A broad term describing the concept of machines being able to carry out tasks in ways
that we as humans consider to be smart or intelligent.
Machine Learning
A current application of Artificial Intelligence that centers on providing machines with
enough instructions, data and processing power for them to learn for themselves.
Machine Learning
Arthur Lee Samuel - 1959
• Teaching machines everything about the world would be very difficult
• Teaching them to learn themselves was a better approach
Internet and Web
• Connectivity
• Vast amounts of data
• Processing power
Define Problem
Collect Data
Prepare Data
Train Model
Evaluate
Deploy &
Improve
How ML works
AutocompleteDefine a problemComplete a word for a user based on their keyboard input.
Collect dataRecord keystrokes from hundreds of thousands or even millions users.
Prepare dataSegment into languages, remove spaces, standardise cases etc.
Train the modelTeach the model what inputs result in which correct word using spelling, proximity to other words, context etc.
EvaluateConfirm outcome of training yields an agreeable success rate.
Deploy and improveContinue to train the model using more data to refine the outputs.
2019: AI in Marketing
What marketers want
To offer better customer experiences
Design personalised approaches
Provide most relevant and timely content
Understand needs in near real-time
Provide micro-moments that delight
Curate meaningful engagements
Productivity
Efficiency
Profitability
What customers want
Better customer experiences
Personalised offers, content and experiences
Content that engages and entertains
Their needs easily met
Enjoyable interactions
Meaningful engagements
AI you probably use daily• Search
• Predictive searches / search algorithms
• Voice search
• Filters
• Smart replies / autocomplete
• Ecommerce
• Product recommendations
• Personalisation
• Social media feeds
• Smart assistants
• Voice control
• Sensor integration
• Entertainment
• Music / TV recommendations
• Upscaling (8K)
• Security
• Fingerprint
• Facial recognition
How your marketing fits in• Search
• Search Engine Optimisation
• Can we show up for voice searches
• Provider choice
• Email marketing tools
• Ecommerce
• Platform selection
• 3rd party AI integrations
• Social Media
• Smart assistants
• Apps and skills
• Custom sensors and integrations
• Entertainment
• Advertising (personalisation)
• Content creation
• Security
• How can our transactions use the latest
security features
AI: Coming next
AI for Robotic Process Automation• Data entry
• Workflows
• Data translation
• Data transportation (routing)
• Reporting and analysis
• Integrations
Chatbots become CaaS• Chatbots mature
• Move from procedural to more conversational bots
• Decision trees and NLP – Machine learning
• Integrations with other AI and workflows
• Conversation as a Service
• Coordinate dialogue between a machine and a human
• Coordinate microservices for processing events, tasks or processes
• Suggest handover for human-to-human interaction
Voice search• 50% of all searches will be voice searches by 2020 (comScore)
• About 30% of all searches will be done without a screen by 2020 (Gartner)
• VoiceCommerce predicted to reach $40 billion by 2022 (OC&C Strategy Consultants)
• Voice searches on mobile 3 x more likely to be location specific (SearchEngineWatch)
• Voice search and AI/ML
• Natural Language Processing
• Voice Synthesis
• Accents, dialects, regional colloquialisms?
Content generation / curation• Auto-generate and suggest content
• Write reports and news based on data
• Highlight content opportunities
• Spot gaps in content
• Auto-generate marketing emails
• Personalised content / language
• Interests
• Interactions
Predictive analytics• Use data to predict outcomes
• Predictive modelling for customer segmentation
• Lead scoring – use data to rate a sales lead
• Campaign nurturing
• Upselling and cross-selling
• Personalised product recommendation
User Experience• Page layouts
• Messaging and language
• User journeys
• CTA’s
• Fulfilment messages
• A/B testing at the per customer level
• Personalised landing pages
• Auto-generated for email marketing campaign
• PPC landing pages
Live AI Demo
Privacy, Ethics and The End of the World
1818 | Frankenstein by Mary Shelley - ethical questions around creating sentient beings
1950’s to present day | Massive technical advancement and commercialisation of AI
April 2018 | Microsoft introduces guidelines for developing responsible conversational AI
January 2019 | Facebook backs an independent AI ethics research center
28 March 2019 | Google announce AI ethics board
04 April 2019 | Google cancel AI ethics board
08 April 2019 | EU releases guidelines for ethical AI
….. What comes next?
AI Predictions
Artificial intelligence will become a positive net job motivator,
creating 2.3 million jobs while only eliminating 1.8 million jobs
by the year 2020.Gartner - June 2018
Impact on jobs
AI will intelligently orchestrate the full customer journey
powering the full lifecycle, from customer acquisition to
engagement to retention.Forbes – Jan 2019
Customer journey
Artificial intelligence-driven creation of ‘counterfeit reality,’ or
fake content, will outpace AI's ability to detect it, fomenting
digital distrust by the year 2020. Gartner - June 2018
The potential down-side
40 percent of customer-facing employees and government
workers will consult artificial intelligence virtual support
agents on a daily basis by the year 2022.Gartner – June 2018
Supporting human workloads
By 2024 Artificial Intelligence will no longer be a buzzword.
Machine learning will be baked into all marketing tools and
services. Customers will expect personalisation, precision and
speed as standard. Rob Gregory – April 2019
AI in 5 years time
A large-scale privacy breach powered by Artificial Intelligence
will impact consumers. Nation states will look to impose
regulations on AI and its providers. Large tech companies
begin signing-up to an AI treaty.Rob Gregory – April 2019
Mistakes and ethics
Adopting AI into your business
Data is key• Make sure you collect data
• GDPR / PECR
• What data do you already have that can be used
• Programmatic advertising
• Business data for training custom ML models
• 3rd party datasets
• Public data sets (Google, Microsoft, Amazon)
• Public models
• Public AI / ML API’s
What to consider
• Does AI make your customer offer better?
• Does AI make your business more efficient?
• Dose AI make your business more profitable?
• Can you achieve ROI?
• Can we use a product or service
• Is there value in creating a custom AI solution
• What could the negatives be?
Mind the gapMicrosoft Tay 2016: twitter bot
Bot began releasing racist and sexually-charged
messages in response to other Twitter users.
Facebook AI Research Lab 2017
Chatbots created their own language, which the
human researchers were not able to understand.
Not surprisingly, the experiment was shut down.
You might not even need AI
• Don’t believe the hype unless it can be proven
• Emperor's new clothes are on spring sale at the moment!!
• Other options:
• Better reporting / analytics
• Improved integrations
• Training on current tools and services
In summary
Key takeaways
• AI isn’t all that new - the Ancient Greeks predicted its rise!
• Machine Learning is the current practical application of Artificial Intelligence
• Don’t be scared or apprehensive, embrace the benefits
• You're already using AI and ML anyway
• Your customers are already using AI and ML anyway
• Tell your children to work hard in maths class and become data scientists
Remember• AI/ML needs data – make sure you collect as much as you can
• You are still smarter than a machine (for now)
• AI will not solve your business problems, although it may shine a light on them
• The mere presence of AI does not ensure success
• Like any tool you must use AI correctly and for the right reasons
Never implicitly trust a machine….