Earl Harvey, our resident Competitive Intelligence guru, sent over a brilliant article from Harbor Research (Smart Systems and IoT Growth Themes and Technology Trends to Watch in 2018), our neighbor here in Boulder CO. As an AI vendor, we’ve touched on IoT and Condition Based Maintenance issues before. Our role in the project was to integrate unstructured and structured data, teaching the machine to read text. But if you think of unstructured text as sensor output, the relationship between IoT and Smart Systems to our CI solutions becomes clear as a core element to digital transformation. In this post, I review the Harbor Research report from our perspective here at BEA but I encourage you to read the full report.
Our mission is to build the open intelligence knowledgebases needed by companies and their stakeholders to provide context for analytics and to accelerate machine learning initiatives. In order to approach human decision making with machine intelligence, data transformation must be addressed from a broad perspective that includes the external (messy & unstructured) environment, the machine (IoT) data and everything in between. Without it, your machine (AI) will never have the contextual understanding needed to power decisions outside of a very narrow scope.
The Challenges to Smart System Potential
From the Harbor report, the excerpts below highlight issues that resonate with us as we integrate open intelligence (again, think of people as another type of sensor that produce messy data). First, on the limits of connectivity,
Connectivity alone ….. does not allow the end user and customer to leverage very much intelligence across myriad brands, suppliers and diverse systems.
The challenges of gathering machine data and integrating diverse data types have been big end customer adoption hurdles, particularly for industries where the range of brands and equipment types number in the hundreds.
…all of this adds up to a huge collection of information-islands.
Compound that with the challenges of different data types, formats and frequency,
The “Achilles Heel” …. i.e. the weakness lies with basic data management technologies—in particular, data transformation and modeling tools—and the restrictions they place upon organizing and utilizing sensor and device data to conduct analytics.
Historically, computing systems have stored information in one of two basic ways: utterly unstructured, or completely structured. At the unstructured end of the spectrum are static Web pages, blog postings, emails, etc., which are free-form and lack any fundamental identity. At the other end of the spectrum are very structured relational databases that are not at all flexible and make rigid assumptions about the meaning and context of the data they store. Between these opposite extremes, intelligent machines on networks are now producing a vast array of semi-structured data types, including machine logs, data streams, sensor values, control signals and more.
Data management and transformation is a critical step in the data analysis value chain
The Solution – New Software Platforms
Enabled by developments across a range of technologies, new platforms are needed. Again, from Harbor, these platforms must
“anticipate developers’ and users’ toughest challenges… in a single, unified, scalable software solution. We believe there are four critical requirements for platforms:
- A fully configurable software platform architecture that enables both peer-to-peer and client-server distribution of services;
- A platform that can simultaneously and asynchronously act on any type of information from any device, storage or streaming source;
- Enabling real-time temporal, spatial and state-based contextual processing; and,
- A platform that provides tools for development of real-time, state-based applications.
The Opportunity – Big Markets
The opportunities and the impact of future developments in this field will drive progress across many of the biggest sectors of the economy. This chart by Harbor highlights the sectors and the contribution of different technological innovations.
In our BEA areas of interest, the Automation & Analytics segment within System Applications grows from $1.17b to $3.2b, from 2017 to 2022. Additionally, Database & Analytics segment of the Value-Added Applications Market grows from $6.24b to $25b, from 2017 to 2022. Plenty of market for new companies and lots of innovation.
Digital Transformation – Smart Factory to Smart Enterprise
While these markets are plenty big, the discussion was focused on the IoT and Smart Systems of the smart enterprise. The challenges and opportunities are even larger when adding the open intel/external sources required to achieve the full potential of integration, digitally transforming the enterprise from islands of Smart Systems to a Smart Organization.