Our Take on Google Cloud Next 2017

Google has been providing effective services for so long that we don’t remember how we lived without them. For me, no Google search or Gmail is as good as no power.

But Google isn’t resting on its laurels. One of Google’s newest platforms – Google Cloud – is poised to make a huge difference in enterprise IT. At its 2017 Google Next conference – held March 8th through 10th in San Francisco – Google unveiled its new plans for its Cloud platform and its plans for how Cloud will affect enterprises.

SpringML was at the conference. Our focus is data analytics; specifically, we work to solve hard data-intensive predictive use cases. We went to the conference to get our questions answered. Like, for example, will enterprises really adopt Google Cloud? And, if so, what will be the impact of the platform on data analytics and machine learning? Further, is Google investing beyond engineering to areas of channel enablement, enterprise sales teams, and partners?

Over the last 18 months, we’ve been able work closely with Google team. We’ve seen them add features to the platform and invest some serious money in Google Cloud. Things that got our attention 1) Google’s deep expertise in AI and ML. 2) Building the  Google Cloud Machine Learning, a managed service that lets customers build machine learning models and work with data and 3) TensorFlow, an open source program for Machine Intelligence that is getting wide spread adoption. We were impressed, so we built connectors and were included as launch partners.

We know we’re not the only ones wanting to learn more about Spanner, Video Intelligence API etc. Listed below are some of our observations while attending the keynote, going through demo zone and walking the partner floor.


One of the biggest hurdles of machine learning is building a whole pipeline. When enterprises put together all the pieces (data prep, data storage, data integrations, workflow, job scheduling and machine learning), the result can often be a hodgepodge of SDKs, ML APIs, open source libraries, and use of languages like Python, Scala, R. If that weren’t tricky enough, adding the infrastructure to a cloud provider requires DevOps maintenance. Ultimately, these elements make AI even more out of reach for enterprises. We know that enterprises adopt platforms, like CRM, Marketing, more easily than standalone services that require they spend a fortune on integration.

Google Cloud’s genius is its simplicity. One of the first differences we saw that sets Google Cloud apart from its competitors is its clear messaging. The platform promises to provide a one-stop shop for computing, storage, data analytics, and machine learning – and it delivers. Sure, Amazon and Microsoft Azure are ahead of Google Cloud on platform maturity, but the integrated Google Cloud is far easier to use than its competitors’ products.

Platform vs APIs

AI and machine learning are in their early stages. We think customers will adopt these new technologies in two approaches: starting with APIs followed by platform.


Specific ML functions exposed as a service. For example, Salesforce uses Salesforce Einstein, an easy-to-use APIs that provides AI capabilities with little or no development. One of the service is a Predictive Vision Service created by MetaMind that helps developers build AI-powered apps simply and quickly. Other examples of APIs are Google Speech, a speech-to-text translation service, Vision API, a platform that helps developers understand an image’s content using machine learning models and Video intelligence API that was released during the conference.

Similar APIs have the potential to proliferate in the short-term, but we don’t think they’ll last in large numbers. Instead, we believe, they’ll go the way of the search engine. Remember 1999 when there were hundreds – if not thousands – of search engines? We think ML APIs will go the way of the search engine: in the long-term, there will be just a few for each specific function.


Very few enterprises have the resources to build a full stack of Machine and analysis Data platform. In the past, large players like IBM, SAP, Microsoft dominated this space but more lately its between Amazon, Microsoft and Google.

Next generation platform developers are less interested in DevOps and more inclined towards server-less services. Eric Schmidt’s quote on “Leave the infrastructure to us, we put in $30B building it” is great example of investment required to build such platforms.


We already knew that Google had good engineers who could code well. But in the enterprise space, that’s never been enough. To enter the market, companies need sales people with enterprise backgrounds, marketing, channel partners, alliances team, field enablement teams, training, and field readiness. Does Google have all these things to get the attention of enterprise IT?

The answer is yes. We learned at the Google Next conference that Google has a dynamic team – and we got to meet a good number of their team members. We met Google’s hardcore product team members, developer advocates, sales teams, partner alliances team, Office of the CTO members, Customer Success engineers and Professional services team, It’s nice to see a mix of google veterans, enterprise IT executives and those who joined recently from enterprise software companies like Microsoft, AWS, Oracle etc.


While it’s hard to predict whether Google Cloud will be a winner at this point, the Google Next conference proved to us that Google has the money, means and commitment. Certainly, Google still has tweaks make the platform more easier to use and build a solid enterprise business unit but judging from the great start at Next and the killer team they’ve assembled, we are excited about Google Cloud. In any event, customers are starting to take a serious look at Google Cloud – for good reason.