fixes an eye firmly on Data Science

Salesforce plans to be a major player in data science in 5 years time

The groundwork for this actually started five years ago. Salesforce will not become a huge innovator in a vast array of domains within artificial intelligence (AI) overall, but I believe it will become a world leader in the lucrative domain of predictive analytics.

Salesforce recently acquired RelateIQ. It has started working to integrate AI technology from that acquisition into its software, seeking to add predictive capabilities. This actually references plans to integrate deep learning to empower better predictive analytics. Deep learning is a subset of machine learning designed to move us closer to true AI — it involves using ‘crude’ simulated neurons to process data. Crude is part of the lexicon — not a pejorative reflection on the tech. Deep learning is a branch of AI made up of techniques that do not necessitate experts to program knowledge into algorithms. These techniques promote learning via data analysis.

Are we really in an AI Spring?

We are, of course, nowhere near true AI (despite the buzz-language being used). We may be chronologically in the first weeks of an AI spring, but we are technologically something more akin to the first baby steps up an AI Everest. The disconnect between the chronological and the technological is explained by the fact that at certain points progress will accelerate greatly. The term “Artificial Intelligence” was coined in 1955 by John McCarthy and described as the science and engineering of making intelligent machines.

Reflect on that term.

Our concept of intelligence is based around humans — our most complex known reference point. Humans are not machine-like (as we conceptualise machines today) at all. Emotions? Many people have published articles on the emotional capabilities of machines like the Japanese robot Pepper.

Unfortunately people are applying the term ‘humanoid’ to Pepper as a noun rather than an adjective. Misleading. Pepper responds to you with apparent emotional reactions. Pepper is as emotional as the 2D cartoon character Peppa Pig. ‘Reading’ outward signs of emotions and experiencing emotions are a galaxy apart. Self-awareness? Empathy? We’re nowhere near even knowing how to get there. If you have been hitherto self-deluded on this score, wave goodbye to the idea that humans have engineered emotional robots to any degree whatsoever. The age of quantum analog computing will already be established before we reach that phase. It’s enormously unlikely that we will successfully produce self-aware, emotional beings in the digital age.

“The sad thing about artificial intelligence is that it lacks artifice and therefore intelligence.”
Jean Baudrillard

A new form of machine intelligence that will outstrip humans in many areas is inevitable if we persist as a species

Despite the long road ahead, we will certainly get there if our biological machines persist long enough into the future. Assuming the survival of our species, self aware, emotional human engineered intelligent machines are as inevitable as the sun rising in the morning.

“If a machine is expected to be infallible, it cannot also be intelligent.”
Alan Turing

Marc Benioff rarely makes off the cuff comments — it’s all planned is new to this AI forum, but Benioff would not be talking about it publicly if he did not have a definite plan to take a seat at the AI table with Google and IBM in the coming years. However it’ll be a seat with a narrow, lucrative focus.

Am I 100% sure of that statement?
I’m never sure of anything, but I’m very often right.

Benioff always tells us what he is going to do in advance. He plants the seed. He moves our thinking in that direction and associates his company with the destination long before it even gets there. But, crucially, he makes very good bets when he’s at A. Even more crucially he makes sure his company gets to Z.

Like all master innovators in technology, Benioff is a futurist and looks to ensure his company becomes an early mover in any groundbreaking domain. He looks at technological trends and predicts what areas will be lucrative in the future. Benioff then asks himself which of these areas will become ripe opportunities for He does this roughly 5 years in advance of making noises about it and then nails his colours to the mast another 5 years out. There’s no fixed number of course. The 5 + 5 pseudo-rule happens to fairly neatly fit the agenda with data science in my view. The point I’m making is that he plans innovation further out than other entrepreneurs. And he allies his evolving product roadmap to his public communications more strategically and successfully than others.

Once he’s decided he’s interested, he starts talking with Parker Harris about the practicalities. When the two agree on something, they turn their heads to the product roadmap and start planning, implementing, hiring, acquiring etc. Further on down the road Marc starts dropping hints in the media. If innovative cycles could ever be described as ‘Groundhog Day’, this is it.

When Marc saw how scalable and useable Amazon was, he decided to bring a far more evolved B2C model (carved by realworld mass customer usage and feedback) to a B2B world. He decided that he’d build a company that would develop a virtual monopoly in CRM and at the same time create the ‘no software’ revolution — forever changing how computing power is distributed. did not change that single-handedly, but they were a major catalyst in this domain — and they came from a then ultra-conservative world of business software. He talked endlessly about software as a service and accelerated the rise of the cloud long before the term had been coined. His ambition always extended beyond CRM. The billboard messaging has always made that obvious.

It’s not just about Big Data.

It’s sometimes about the Big Message

It doesn’t matter if technically the slogans are not even true — or even ironic. Its what the message implies or elicits from the receiver (customer) that counts. As long as they get one from A to Z efficiently, and as long as Z is both a profitable and a good place to be, then it’s all good. No Software.

A few years ago in Australia he commented on how email was on the way out. The first of many comments. Then he hit the world with the statement “Email is Dead!”
This statement was incorrect then. It’s still far from true today. However it was an extremely clever, provocative, planned communication. Salesforce was one of the first investors in Hootsuite — a very good company with a great social media management platform. Benioff probably said to himself “I want one of those and I can make it more useful because of the ecosystem of apps I’m building”. I’m sure I haven’t got that word perfect as I haven’t learned how to read minds, but that would have been the gist of it. So he started preparing the ground with increasingly strong statements on the demise of email and rise of the social enterprise. When he announced the death of email he created a tremendous backlash, pushing 1,000s of naïve buttons amongst journalists and bloggers and social media feeds around the world. Some months later he released Social Studio.

So now Benioff says: “We’re in an AI Spring.”

Off the cuff? Yeah right.

He’s even playing with words. Referring to the acquisition of EdgeSpring, a company that made analytics software helping companies make business decisions using unstructured or partially structured data coming from sources like social media streams, email and mobile apps. That acquisition was the source of Wave, which powers the Salesforce Analytics Cloud. It is a first major step in the direction of their predictive analytics that includes machine learning and data mining to analyse facts and make predictions about unknown events — including the future. Predictive models exploit patterns found in historical data to identify risk and opportunity.

Email is dead.
No software.
We’re in an AI spring.

All carefully planned and orchestrated.

“The revolution in data science will fundamentally change how we run our business…. based on the simple fact that there’s just a huge amount more data than ever before, our greatest challenge is making sense of that data…..and we need a new generation of tools to be able to organize and view the data. The whole concept of data science is that the software becomes the expert, and you as the average user are able to understand what’s going on.”
Marc Benioff

These are very similar to the statements I made when I announced the concept of “Understanding as a Service” or UaaS in October last year. UaaS is quite simply the science and art of turning data into understanding.

So why is Salesforce obsessed with data science?

Because of the quantity and, especially, the quality of data at their disposal. They have platinum and gold in their mines.

They have a deep footprint in many companies and an unbelievably rich global database. 3 billion transactions go through Salesforce daily. And much of the data is hugely valuable. By becoming leaders in data science, Salesforce can empower companies to do amazing things.

In purely qualitative terms forget Big Data.
Salesforce has Vast Data

I would say it’s a no-brainer, but that term feels inappropriate in this case. Our current crude baby-step AI powers IBM’s Watson, Apple’s Siri, much of Google’s core functions, Netflix’s movie recommendations, self-driving cars and autonomous drones. Imagine what Salesforce can achieve with that technology applied to such a vastly rich database.

Salesforce will, at a minimum, look to become the king of predictive analytics.InsideSales is an impressive example of a company that already uses machine learning for predictive analytics. So it’s no surprise that Salesforce Ventures has been a major investor in InsideSales. And the fact Salesforce named them as a premier partner is even further evidence of their focus on this area. SFDC has a pattern of investing in technology companies and then either buying them out or acquiring/building a competitive alternative to them.

Benioff has been considering the data science opportunity for years. He’s joining the fray while it’s still relatively early. There are powerful software companies that sit at the top table of AI — notably IBM (the world’s most inventive company) and Google(the world’s biggest born in the cloud company). So it will be no small challenge to take a seat beside them in this domain. However Forbes has ranked Salesforce as the world’s most innovative company (in any industry) 4 years running for good reason. And Salesforce is the only major born in the cloud company with a very deep footprint inside companies.

Benioff recently said: “Why would Uber have all these data scientists?” Salesforce has poached almost the entire data science team from LinkedIn. This is very obviously all about predictive analytics. But Benioff is also watching even smarter entities like MetaMind.

MetaMind is impressive enterprise AI startup focussing on deep learning, natural language processing (NLP) and image prediction. They are applying new deep learning research in language and vision to easy-to-use applications. Marc Benioffsits on the board. He says MetaMind’s “deep learning technology is going to have enormous impact in multiple industries, and has the potential to provide a generalised mathematical model for building machine intelligence that could be adapted for almost any discipline.”

Benioff has also been a lead investor in machine translation company Cloudwordssince Series A funding days. Machine translation is a problem we’ve been trying to solve since the 1950s. It’s a problem that frightens away a lot of people. An aura of perceived difficulty creates enormous opportunity for those who understand that the timing is right to solve a particular problem. We’ve slowly come some way since the heady, unrealised statements that followed the Georgetown-IBM experiment in 1954. We’ll need analog quantum computing before we can break a sweat, but that’s another article.



Well to conclude let’s return to what Benioff is saying. Remember — his statements are a very reliable weather vane pointing us in the direction of where Salesforce is moving in the medium term.

“We don’t need more cloud. We don’t need more mobile. We don’t need more social. We need more data science.”

Actually that’s not true at all.

We need more cloud.
We need more mobile.
We need more social.

Benioff’s first three statements are incorrect.

But, as per normal, he knows he’s ‘wrong’ of course.

Just like “Email is dead”, this is another deliberately inaccurate statement. It’s also an extremely astute statement. It serves a purpose. Benioff is already riding the cloud, mobile and social waves successfully. He has people executing phenomenally well in these domains. He did more to create the cloud wave than anyone else in B2B. Those machines are up and running and he has every intention of continuing to innovate in them.

What he’s really saying is:
“Data science is becoming huge and Salesforce will become a major player in this domain.” But he’d never be so boring as to state things so plainly.

The extent of Salesforce’s ambition regarding AI

Will Salesforce be minding a quantum analog computer in a giant freezer, producing it’s own Asimo or joining Amazon and Google in drone wars? I can’t see that. Salesforce has always been razor focussed on becoming the dominant cloud provider of business software. Difficult to imagine that changing in the next 5 years. So Salesforce will now focus relentlessly on being the King Kong in predictive analytics powered by machine learning and data mining. They have made a strong start with Wave and RelateIQ, but they’ve miles more to go before they sleep.

This is no small ambition, but past experience has shown us that Benioff’s statements are far more substantial than bravado. Benioff once more is threading his own tried and trusted path to success. Even if the process and the noises are familiar, it’s a brave new adventure nevertheless. An adventure that only an arch entrepreneur with an insatiable hunger for success and a genuine interest in progress would embark on.

“We can only see a short distance ahead, but we can see plenty there that needs to be done.”
Alan Turing

p.s. As I write this I’m sitting in the NDRC (National Digital Research Centre) in Dublin, Ireland — an accelerator that occupies the Digital Exchange building (an historic Guinness storehouse). I’m conscious that MIT Media lab occupied this building for several years. MIT is, of course, a major player in data science and AI. So it feels appropriate to use a gorgeous MIT generated image for this article.NDRC is an investor in AppSelekt — a marketplace under construction that makes it easy to select the right business apps. It’s built on the principles of Understanding as a Service (UaaS) and powered by unique machine learning and natural language processing.

Live long and prosper!
Stephen Cummins
March 7th, 2015 — — — — — — — — — — — — — — —

If you found this interesting, it’s always great to get feedback with a comment and/or ‘like’.

If you are interested in knowing more about AppSelekt, you can follow my writing here: Understanding as a Service (UaaS).
Alternatively connect with me on social networks you use:
AppSelekt on LinkedIn

Header image © MIT Media Lab
‘Steinunn the silent Puppetmaster’ © Stephen Cummins Photography
(Sculpture by Steinunn Thorarinsdottir)
The Dream (Dreamscape of Survival #1) © Stephen Cummins Photography
The Silver Hunter’s Break © Stephen Cummins Photography
Pepper the Robot © Aldebaran Robotics and Softbank Mobile
Asimo © Honda
Stature of Turing © Bletchley Park (sculpture by Stephen Kettle)