Franki Chamaki » Uncategorized
AI and You = You 2.0
  • Screen Shot 2017-06-19 at 6.48.29 am

“The real problem”, as BF Skinner the social scientist rightly pointed out “is not whether machines think. It is whether men do.”


When it comes to technological advancements it would seem Skinner was right on the money. We tend to take an alarmist view, every time we are confronted with something new. Take the arrival of the textile loom in 12th century Europe. Today we recognize this as the harbinger of the industrial revolution. Back then however, it caused riots and the Dutch even destroyed the loom as it caused job losses among sheep farmers.


The problem is that while the science has advanced tremendously from the basic loom, we have not managed to advance the argument. As the COO of a company that is using AI to unlock hidden profits in the retail environment, I find much of my time is spent dealing with a variation of a nine-century old question. How will technology (in this case AI) impact jobs? The answer, unfortunately, is also nine centuries old. Like every great advancement in technology, AI is massively disruptive. And like every great technology AI will propel mankind to greater heights of achievement in all spheres of activity be it science or art, physical or mental. There is not a single aspect of our lives that AI will not touch – if it is not already – and greatly enhance. Let me share a couple of examples.


A common story in the medical profession is that AI will cause massive job losses among doctors as AI driven apps like Babylon can blitz through databases of hundreds of thousands of patients to quickly diagnose symptoms. However, it took a combination of both human and artificial intelligence to work out the 3D structure of an HIV enzyme in just three weeks. Alone neither could get the job done and together we are greater than the sum of the parts. AI enhances humans as much as human intelligence enhances AI’s abilities.


Similarly there is a fair bit of hullabaloo in the legal community about AI’s potential to replace lawyers, or even judges, and while AI is threatening these kinds of white collar jobs, it is going to replace many jobs that require routine mental work, such as writing up contracts. AI is also an excellent assistant to the trial lawyer looking for the one loophole that will exonerate their client. Think back to that classic 1992 Hollywood courtroom comedy My Cousin Vinny. The fate of the case hinged not on Vinny’s knowledge of the law, but on his girlfriend’s expertise of automobiles. Now, you wouldn’t expect most lawyers or their girlfriends to know that much about cars. But given an AI assistant Vinny would still have solved the case. All he would do is feed the critical photographs to his assistant and ask it the question he asked his girlfriend. Bingo, he would have got the exact same answer, without the attitude!


That’s the whole point with AI. It will help us as the human race move on to the next level of achievement. Our personal virtual assistants will be constantly by our sides, monitoring us, advising, informing and helping us get better at whatever we do. We’re seeing the initial burst of products in the virtual personal AI space. From Apple’s Siri, Google’s Assistant, Microsoft’s Cortana to Amazon’s Alexa and a few interesting ones from Smartphone manufacturers you can get a teaser of what to expect when you collaborate with AI.


But these are early days. Moving forward your personal assistant will become capable of a lot more, will understand you better and tailor itself to creating a better you. The soldier will have her assistant, the musician his. Eventually, she will become a better soldier and he a better musician. We are not there yet, but I believe Tony Stark and Jarvis is the vision of an AI-enabled future. There is no Ironman without Jarvis and no Jarvis without Tony Stark. It is an integrated future full of amazing possibilities So let’s junk the old argument and embrace the brave new world of AI.


P.S. One of my business fellow who is a father of 3 already think about how AI can help his kids, and maybe yours, learn smarter.

Interesting presentation by  Tom Davenport at 2015 Salesforce keynote.

Tom Davenport is the President’s Distinguished Professor of Information Technology and Management at Babson College, the co-founder of the International Institute for Analytics, a Fellow of the MIT Center for Digital Business, and a Senior Advisor to Deloitte Analytics. He teaches analytics and big data in executive programs at Babson, Harvard Business School, MIT Sloan School, and Boston University. He pioneered the concept of “competing on analytics” with his best-selling 2006 Harvard Business Review article (and his 2007 book by the same name). His most recent book is Big Data@Work, from Harvard Business Review Press.


He views are consistent to want we do at HIVERY, we are pioneering what we call “Prescriptive Science™” – what he calls “Analytics 4.0”


Below I’ve summarized his talk along with few examples:



Screen Shot 2016-03-27 at 10.19.18 am

Here is a summary.  He talks about 4 types of analytics:


  • Analytics 1.0 was the era of business intelligence –  descriptive, which reports on the past.
  • Analytics 2.0 was about Big Data – which uses models based on past data to predict the future;
  • Analytics 3.0 is the era of data-enriched offerings – which is about being prescriptive. It uses models to specify optimal behaviors and actions.
  • Analytics 4.0  — the idea of automated analytics. These come to full fruition in a new era. Machines talk to machines to carry out decisions within human input.
    • Moving towards “Automated decision from interconnected smart machines”.
    • Connected sensors and the “Analytics of Things”.
    • Things will be “augmented” rather “automated”


Screen Shot 2016-03-27 at 10.32.31 am

Examples of established (old) companies leveraging data-driven approaches to model business innovation



GE 3.0 a 123 years old company

  • Built a new business – $2b initiative in software, analytics and what they call the “Industrial internet” called Predix.
  • Producing “data-based products and services”
  • Add sensors to their industrial products (e.g. gas turbine, jet engines, and trains etc) to understand how they are preforming.
  • Key objective: to revolution services it offers.
  • Why? 75% of its profit comes from industrial services. If its able to estimate when I jet engine is going to break and service it before it breaks, they can make a lot of money. Ability to sell a unit of “jet thrust” instead of jet engine and produce that “jet thrust” for less, they can make more profits. So GE puts sensors on their jet engines to allow them to service them better (i.e. forecast when likely its going to break down) but also create new business model of charging for jet engines


Monsanto 3.0 – 114 years old company

  • Creating “frankenfood,” but Monsanto is not the only company that produces genetically modified organisms
  • If we are going to feed 9 billion people, I think we going to need some innovation in agriculture. Its a agricultural biotechnology company
  • It has “Precision planting” or “prescriptive planning” offering to sold to farmers. That is not just selling seeds and pesticides, but sell “advice” to farmers when to plant, what to plant, how much to water, when to put pesticides, when to harvest this year. These are new “data-products”
  • In 2013, it acquired is a huge agritech startup: Climate Corporation for approximately $1.1 billion. The company uses machine learning to predict the weather and other essential elements for agribusiness.
  • It provides “field-level” highly granular weather data – called “FieldScripts” That is provide insight to when its going to rain and when pest going to start, to prescribe when to apply pesticides to farms.
    – Yield with this advice for farmers, increase by 10% to 20%.


Ford 3.0 – 112 years old company

  • Bill Ford “The car is really becoming a rolling group of sensors”
  • Ford Creates New Chief Data, Analytics Officer Position
  • Fords Digital analytics and optimisation team – defined Ford’s digital web analytics strategy & standards for all B2C properties providing Ford a competitive advantage of integrated business intelligence and targeted marketing opportunities.  Example of talent include.
  • Targeted marketing – Digital In-Market Manager is responsible for all activities related to targeting and messaging in-market auto shoppers in order to convert them to Ford customers. This includes establishing an “always on” approach to targeting shoppers online. Position responsibilities will include strategy, creative, production, optimization and media planning for all in-market activities
  • Business intelligence – help dealers become more successful by providing smart inventory system: instead of sending new cars to all dealers, now they can figure out what type of car is most likely to sell in particular dealer lot and increase revenue by $100m

LinkedIn – 14 years old company

  • Has a lot “data-products” including “People You may know”, “Jobs You May be Interested in”. “Groups you May Like”
  • Uses it data to determine who is most likely going to buy Linkedin services


View it here or

The process of turning guesses into fact –  the hammer, nail and wood story


So you have an idea that its been crowd sourced and people love it!   You have a business sponsor willing to take your idea but now your asking yourself – how on earth am I going to  implement it?  Would it work really?


I am a big believer of using Start up thinking in a corporate environment to move an idea into reality. At my work I often use frameworks advocated by Eric Ries in Lean Start up  –  which I often call the “nail”Steve Blank on customer discovery  –  which I like to call the “wood” and incorporate a little Alberto Savoia‘s pretotyping – which I call the “hammer“.  I know its stretched out analogy – but play along with me!  Does the “nail” (ie MVP) pass through the wool easily (ie be accepted by customers) when you are using the hammer (ie the process of conducting the test)? If so, keep progress, else pivot your approach and just another nail!


A minimal viable product (MVP) can be used against an agree set of hypotheses to validate whether an idea will work.  Now there are few ways to do this and depends on where you are at in your hypotheses.


Pretotyping is about testing whether an idea will actually solve a customer problem (ie will it be used) while Prototyping  tests whether one can actually can build the solution in viable way.




By testing I mean turning these those assumptions or “guesses” (i.e. your hypotheses) into real solid fact.


I like to talk about  three methods of test  – and this was posted on which provides a good practical summary:


1. Exploration Method


Based on ethnographic research techniques, this method is basically exploring the solution through conversation.  You need to first  ensure that the person you are interviewing fits your customer target  –  ideally early adopters of your solution who will help you validate  your assumptions.  Here are few statements to start with:


  • “Tell me about the last time you…”
  • “How are you currently solving [insert problem]?”
  • “How/where did you search for a solution?”
  • “If you could wave a magic wand and have anything you wanted, what would it look like?”

2. Pitch Method


This method is basically used to engage with your target customer  and try to see if they are willing “invest” in your solution by getting them to exchange some form of “currency” (eg an email, a tweet, an verbal agreement or even money).  Its also known as smoke testing.


Forms of Pitch
  • Landing Page –  I’ve used this and actually set up one using
  • Cold Email
  • Cold Call
  • Gimmick Sign
  • Fake Ad (ie gumtree, ebay)
  • Walking into a store and pitching the owner
  • Kickstarter
All of these forms can be considered “pretotyping” and you can use the following types of currency
  • Cash $$$
  • Letter of Intent
  • Email Addresses
  • Pay With a Tweet
  • Taking a Meeting
  • Time

3. Concierge Method


I love this method and I have personally used it to test some of my “start up” ideas. It is often referred as Wizard of Oz ie. faking the product with as little technology as possible.


If you go to Alberto Savoia’s around 16.42 minutes you see a classic example of this in action.
The goal of concierge is have your target customers engage with your product (ie use it) and see if would they buy it right now.
In the video example with IBM, initially people thought “speech-to-text” was a great idea, but after the people had the chance to actually “use it” , people didn’t like. Some said, their throat became sore or the office became be very noisy, others said it would be very difficult to dictate confidential memos.   Killing the idea completely but savings millions of dollars in the process!
Summary:  when you have an idea, you really have assumptions.  You need to turn those assumptions about your idea (ie features, target segment, price etc)  into fact.  These method help with this process.  Always ask yourself what is the most simplest possible way you can test your idea and turn any assumptions into fact and determine how and if your would actually idea work.  Use the Learn startup method, create a MVP, apply pretotyping methods and get out of the office!