Accelerating AI: Enterprise Simplification and Deployment on the Horizon
Business use of AI grew 270 percent over the past four years, according to Gartner, while Deloitte says 62 percent of respondents to its corporate October 2018 report deployed some form of AI. That’s up 53 percent from a year ago, but what we’ve learned is that adoption doesn’t equal success, and success is an evolving model in this phase of our digital revolution.
Unfortunately, roughly 25 percent of companies have seen half of their AI projects fail. Failure, in heavily technical deployments, like AI projects, is incredibly expensive when data scientist and other team time, technical cost of computation, and resources wasted is accounted for. Statistics like these have generated tremendous buzz around the end results: success or failure, but we’ve reached a pivot point where we must widen our lens and shift our attention.
To see widespread adoption over pieced together strategies, it’s time to focus on the enterprise-wide processes, the implementation, and the potential impact of successful adoption. The participation in simplifying the path to adoption, in understanding the benefits and use cases, and in aligning this vision with how the world does, and will do, business is where we can see the most innovation and success.
“AI is really impacting the way the world does business. We see 75 percent of commercial enterprises are doing more with AI in the next several years… As companies are driving this forward, developers are on the front lines, trying to figure out how to move their companies forward, how to build these models and how to build these applications, and help scale with all the changes that are moving through this.” -Eric Boyd, Microsoft CVP AI Platform
Simplifying to Scale
The tools, techniques, and methods of analyzing data, predicting, and learning from it presents the process but it needs simplification to reach the scale necessary to begin to see that failure rate level and eventually drop off. The problem with scale is the adoption rate that the current models support.
The rate at which businesses currently have capacity to initiate and fully step into AI is not where it needs to be for success. This is why we are seeing the tech giants of the valley deploying top talent and resources, to look at the current programs, to build new- more effective programs, and to show businesses how they can utilize and scale, with what they have, in a more simple way.
Architecting A Successful Implementation of AI Technology
Fivetran is one of the most recent companies building out tools that can be leveraged to make AI and big data deployments easy, cost effective and quick- across an entire enterprise. This type of pre-built, zero maintenance solution is the first step in architecting and simplifying the implementation of enterprise-wide AI processes.
“Our pre-built data connectors deploy in minutes, and are fully managed which means data engineers no longer have to waste their energy on pipeline maintenance. Automatic schema migrations and continuous ingestion of change data into a central warehouse provide data professional query-ready access and the ability to combine data sources in a systematic way.”
Part of the problem Fivetran, and plenty of others competing in this space have recognized, is the under-performance and complicated configurations of cloud-based software and storage. Streamlining and accelerating analytical and data-based projects allows the Fivetran backend to carry the autonomous work seamlessly, and with much less risk. This level of simplification and systematizing is also what will allow mass scale adoption, rather than the “pieces and parts” approach we are currently seeing.
H20.ai is another machine learning solutions firm experiencing massive growth and investment, after just completing a $72.5 million round led by Goldman Sachs and the Ping An Global Voyager Fund. Google’s BigQuery, a serverless data warehouse that scales up and down automatically, is making big strides towards simplification. By analyzing big data in less time, with easy access and massive capacity, BigQuery is showing us the future of applicable AI. Again, reiterating the need for simplification so we can continue deploying more complex use-cases in every sector.
“We still remain far from general AI that can wholly take over complex tasks, but we have now entered the realm of AI-augmented work and decision science — what we call ‘augmented intelligence. If you are a CIO and your organization doesn’t use AI, chances are high that your competitors do and this should be a concern.” Chris Howard, Gartner, Chief of Research
McKinsey Global Institute predicts a labor market shift that will result in a 1.2 percent increase in GDP growth over the next decade. This would help capture 20-25 percent additional net economic benefits. This statistic, above all, identifies another factor in this digital revolution that must be met with tools and resources: the talent pool.
AI Talent Development
We’ve identified the seemingly common problem of enterprises with disjointed processes that often do not work together for large scale functional models. While we have to accept that, with any new digital mass-scale adoption, growing pains are normal and to be expected; it’s important to also look to the innovators like Fivetran, H20.ai, and others for solutions. This shift to solutions allows us to put focus on another pain point, a lack of incoming talent where AI could be developed or scaled. We seem to be ripe with startups and firms, but the enterprises are where the talent needs distribution, so internally there is support and expertise to implement.
Ironically, where AI will continue removing the need for human-based repetitive strategies or processes, it’s creating a need for more human engagement on the solutions side of growth. For data scientists, developers, and analytics-based career paths, we need tools for control, and to track and manage assets so the experience is simplified. Enterprise positions are going to be in heavy competition with the learning firms and startups being funded now, because essentially they are up against a developer or data scientists chance at building a legacy. The stakes are high, and this makes it even more important for enterprises to have a clear plan in place. Utilizing programming and tools available through some of the heavyweights listed above is one of the top industry secrets- giving those enterprises a leg up so to speak, because they are actively mitigating risk through already proven tech.
Building New Capabilities
With a clear path forward, or at least a jumpstart, tech in place to support that path, and recruitment of talent- businesses will begin to see how many possibilities and use cases are on the horizon. From there, the process that makes the most sense is to utilize the systems built for simplification, so adoption happens swiftly and laterally, rather than here and there. That clear path forward, an enterprises starting point, needs to have a process defining the entire path.
- Identify the solutions you seek out with AI. What problems are you attempting to solve? What use cases are the most appealing to your business? Even if items on your list are (or seem) out of reach current day, leave them as part of your potential future planning.
- Identify your internal capacity. The failure rate we see right now is being driven partially by attempts at implementation lacking the proper internal support. Creating job shares or investing in training is a starting point. From a tech standpoint, and from a talent standpoint, coupled with the buy-in from executives and trained teams- understand your capacity and potential expansion. This is an integral part of the planning process.
- Identify external resources that can automate your resource gaps. This balancing act of teaming the inner with the outer, in the past, has been more difficult, but we are seeing this gap begin to slowly close as the gaps become more clear and the tools become more useful.
Machine learning, deep learning, and natural language processing are allowing businesses to integrate AI into their current processes and platforms. This is what we could refer to as phase one. This is the most basic use, while still incredibly enchanting, of how AI can make processes better. As machine learning drives AI, the shift into more complex phases of use-cases will pick up speed and we will see rapid change in all adopting sectors. Whether we’re talking about repetitive processes, CRM, social data, insights and intelligence, logistics, efficacy in current market holdings, or the growth of existing services and products- we can see that the next swell is just on the horizon.
“Right now, AI is being driven by all the recent progress in machine learning. There’s no one single breakthrough you can point to, but the business value we can extract from machine learning now is off the charts. From the enterprise point of view, what’s happening right now could disrupt some core corporate business processes around coordination and control: scheduling, resource allocation and reporting.” -Luke Tang, TechCode, GM of Global AI+ Accelerator
Deloitte’s second State of the Enterprise compendium hit heavy with a handful of other category statistics we should consider:
- 42 percent of executives believe AI will be of critical importance over the next 2 years;
- 62 percent of companies have adopted NLP, up 53 percent from a year ago;
- 58 percent of companies have adopted machine learning, up 5 percent year-over-year;
- 54 percent of respondents identified the skills gap as their organizations biggest challenge.
Dr. Mark Esposito, co-founder of Nexus Frontier Tech and the instructor of Harvard’s AI in Business program, has pointed out that there is still a lot of misunderstanding surrounding AI and the current capabilities, with the biggest one being that AI can do everything a person can do and then some. Addressing the misconceptions will also allow us to shrink the failure rate.
“The more you learn about the technology, the more you understand that AI is very powerful. But it needs to be very narrowly defined. If you don’t have a narrow scope, it doesn’t work.The low-hanging fruit is recognizing where in the value chain they can improve operations. AI doesn’t start with AI. It starts at the company level.”
Democratizing AI, removing barriers, and taking a wider perspective is where we are now. Our business ecosystem, nationally and globally, has raised the bar and where we must meet it is in a place of simplified sophistication. Creating enterprise-wide impact, in every sector and individual marketplace, is the ecosystem of the future. That next swell on the horizon, is the one that will wash over us technological use-case capacity, with internal expansions, dedicated talent, bringing a new phase of our digital revolution. We are close enough to feel the buzz in the lines but far enough to be experiencing a 25 percent failure rate among adoption. During this in-between, the growing pains should serve as lessons learned, to enterprises everywhere, that they are capable of lateral expansion- but how much is really up to them.
About the author: Ben Bloch is a Los Angeles-based entrepreneur and founder of Bloch Strategy, a consulting firm that helps startups and Fortune 1000 companies raise additional rounds of funding.
This article originally appeared in sister publication Datanami.