When IBM and Salesforce announced their major partnership around Cognitive, Artificial Intelligence (AI), and Machine Learning many commentators, myself included, were surprised and maybe even a little confused. Even after IBM’s Interconnect conference keynote, where the warmth between IBM CEO Ginni Rometty and Salesforce Founder was palpable, warm and genuine it wasn’t clear how the deal would work. Shouldn’t Einstein and Watson compete?
Surely both companies needed their own strong AI/ML plays. How would a partnership work in practice.
It is probably best to start with the somewhat prosaic – classic multivendor services. With its acquisition of Bluewolf, IBM is now one of the biggest Salesforce implementation shops on the planet. There is plenty of growth potential in the Salesforce enterprise customer base, and IBM wants to tap into that. Data is a big part of that upside, according to Bluewolf’s fifth annual The State of Salesforce report.
Most companies however are just not set up to realise the value of their assets. Data transformation is the new digital transformation. Jerry Chen over at Greylock has started talking about Systems of Intelligence as the new moats driven by aggregation and AI
“the battle is moving from the old moats, the sources of the data, to the new moats, what you do with the data. Using a company’s data, you can upsell customers, automatically respond to support tickets, prevent employee attrition, and identify security anomalies. Products that use data specific to an industry (i.e. healthcare, financial services), or unique to a company (customer data, machine logs, etc.) to solve a strategic problem begin to look like a pretty deep moat, especially if you can replace or automate an entire enterprise workflow or create a new value-added workflow that was made possible by this intelligence.”
Much of the potential data out there though is not in the places where we need it. For systems that were built before cloud native was a thing there is a treasure trove of data out there in legacy logs, formats and documents. Consider all of the domain specific data held in PDFs for example, in any vertical industry. Turns out Watson is pretty good at parsing this stuff. Never mind the higher order Cognitive story IBM is selling, this is about automated extraction and classification, to make amenable for aggregation.
How do you build digital twins of products across the entire lifecycle? Consider that Daimler is using Watson to read specifications and manuals in natural language written in Word and Powerpoint documents in order to start turning them into working digital models, prototyping from design to manufacture and production. We also have language translation issues to overcome, another area where IBM is doing the heavy lifting with Watson. Another industry data with data in a bunch of hinky formats is Insurance – IBM will make this data available for aggregation with Salesforce. Could be transcriptions of accident reports for example. Healthcare – Watson has shown stellar results in diagnosis that are very relevant to patient outcomes.
Or take weather – IBM bought the Weather Company, has a data moat and it makes no sense for Salesforce to do anything but partner there. Weather changes human and consumer behaviour and affects corporate and personal assets. And there is a lot of weather happening these days.
The above are examples of work that Salesforce doesn’t want do with Einstein. Salesforce is all about structured data and customer experiences and will focus its attentions there. Of course there are going to be overlaps – see for example Einstein Vision. Image Recognition is the new Hello World. Overall though the structured unstructured distinction works pretty well. And while the new ML/AI APIs are very easy to use, there is a lot of heavy lifting in the background to deliver a simple experience.
Now it’s worth taking some time to consider story sequencing and the customer journey in understanding IBM and Salesforce positioning.
IBM has an over-arching narrative around Cognitive that frankly requires a lot of explanation, a lot of upfront and presales work and customer hand-holding. Cognitive is not a SKU. IBM has PhDs that will talk about how the brain works, and how we need to make companies more like that. Well Ok.
But Salesforce doesn’t take that kind of approach. Benioff prefers a simple story that he can tell 180,000 people at Dreamforce with a call to action. “Hey look, admin, you’re awesome but we can make you even more awesome – flick this switch and you’ll get vastly improved lead scoring because of machine learning.”
Salesforce wants to improve its core CRM apps quickly with machine learning and AI techniques in an easily consumable way. Of course once the admin flicks the switch, the process needs improving – oh look we better call Bluewolf for some help with that.
I talked to Phil Cooper, VP, Product GTM, Einstein at Salesforce this week. He explained
“Einstein is not a cloud in its own right. It’s capabilities are all baked into the platform. With the Intelligent Customer Success platform – any function should have intelligence baked into it.”
The bulb really lit up for me though when he said:
“Einstein is to Salesforce as Cognitive is to IBM”.
Einstein is about capabilities not product, and and as such Salesforce is of course going to partner with companies that can make its apps more intuitive and helpful in the short term. Add to this IBM being able to offer implementation and consulting services to help enterprises integrate their own legacy data and processes with Salesforce. IBM gets some high dollar services contracts, rather than Watson API micropayments, as well as great partner bragging rights. Yeah I can see why Watson and Einstein could get along as well as Rometty and Benioff.
IBM and Salesforce are both RedMonk clients.
(Read this and other great posts @ RedMonk)