Michael Yamnitsky, Work-Bench: On enterprise machine learning and why ‘it’s a good time to be a mega cloud’
The future of enterprise software will be in some part automated, with machine learning (ML) and artificial intelligence (AI) technologies really starting to come to the fore. For all the actors cast in this fascinating drama – from the largest cloud vendors to startups, and from business analysts to data scientists – it’s time to either start learning their lines or, in some cases, rip up the script altogether.
The script in question? The Empire Strikes Back.
Work-Bench, a New York-based venture capital firm focusing on enterprise technologies, released its 2018 Enterprise Almanac report last month with that very title. The reason relates to the culmination of a long-standing trend. 10 years ago, it was a clear fight between the on-prem empire and the ‘cloud rebel alliance’, as the report puts it. Today’s rebel alliances have to fight not just the on-prem overlords, but the cloud hypervendors – Amazon, Microsoft, Google et al.
This is a trend that is not going away any time soon. Michael Yamnitsky, venture partner at Work-Bench and author of the report (left), jokes that next year’s report will most likely be titled Return of the Jedi. Yet as the report asserts, large technology companies are ‘#winning’ – the report’s hashtag – at AI. Not only are the largest cloud vendors releasing various toolkits – Amazon with SageMaker and Lex, Azure with Machine Learning Studio – they’re also hoovering up the best AI talent.
Work-Bench’s vision is ‘hoping that new talent gets excited about the enterprise’ – and as this publication put it when covering the original report, the promises of AI and ML will give plenty of reason to get excited in the coming years.
In an email conversation with CloudTech, Yamnitsky gives his verdict on what has changed in the industry over the past 12 months, the rise of Salesforce as an AI force, and what the biggest cloud players and BI vendors will do from here.
CloudTech: How much has changed in the enterprise software industry between this year’s and last year’s reports?
Michael Yamnitsky: A lot! The industry is constantly evolving. That’s what makes early stage venture so much fun. Building a new company in a highly dynamic, competitive market means you always need to play mental chess to figure out the right moats and pockets of value you can monetise.
CT: The report touches on the shift of moving natural language processing to business reports. There are companies looking to do this, but is this ‘democratisation of data’ really going to change things at the executive level?
MY: It will – but it will take time. The promise of products like Salesforce Einstein are to allow anyone to find insights in data without prior knowledge of the underlying data structures. Executives are certainly not precluded from this shift.
CT: Is it wanted from all sides – and what does this mean for data scientists? Is it similar to citizen developer initiatives from a few years ago, or will this take food off their table?
MY: That’s an interesting question and it comes down to culture. Some data scientists embrace democratisation, while others want to keep the lid shut so their work – and position of power – in the company remains stable.
CT: You focus on Salesforce as someone to keep an eye on for AI with its Einstein suite – could you elaborate a bit more on why that is, compared with other companies?
MY: Salesforce Einstein is based on a product built by BeyondCore, a startup Salesforce bought a few years ago. The product is very impressive. Salesforce just doesn’t have mindshare in the BI space. People do not know much about it. Salesforce has a good eye for marketing and I’m [sure] will have no problem catching up.
There are some stealth early-stage companies trying to emulate Salesforce Einstein functionality with standalone products and Tableau seems eager to compete in this area – but otherwise Salesforce has a highly differentiated product in the market.
CT: If you are a more traditional BI vendor reading this report, what do you have to do?
MY: Traditional BI vendors certainly understand this shift and seem to know what to do about it given the recent developments and M&A we see in the market.
CT: What do the next 18 months or so hold for the ‘mega clouds’, as you call them in the report? Market share remains stellar and capex continues to climb – and they seem to be leaning on their huge growing shares in infrastructure to particularly explore ML tools. Will that last?
MY: The mega clouds continue to surprise us. We assumed last year they would stick to building developer tools. That’s certainly the case for Microsoft and Amazon, but Google seems eager to build vertical AI applications starting with customer service.
I would not disqualify the other two from pursuing a similar strategy, or from pursuing any other product-market extension for that matter. It’s a good time to be a mega cloud.