Views from the Marketplace are paid for by advertisers and select partners of MIT Technology Review.
AI Gets More Real, Thanks to Contextual Deep Learning
No longer is artificial intelligence the stuff of books and movies. It’s already part of our everyday lives—in search engines and weather forecasts, for instance. Now, with machine learning and contextual technology, AI is evolving to the next level.
When you first think of artificial intelligence (AI), you might envision a fully automated society in which robots serve our every need. Or you might imagine the disembodied voice of HAL, the recalcitrant computer that wouldn’t open the pod-bay doors in “2001: A Space Odyssey,” the renowned 1968 science-fiction film.
But science fact is more practical, more affordable, and more widespread than Hollywood portrays.
RAGE Frameworks, a provider of knowledge-based automation technology and services, epitomizes that reality. The Dedham, Massachusetts–based company has developed artificial intelligence technology that scans mountains of structured or unstructured documents. The engine of RAGE AI, as it’s called, sorts through the data it ingests and provides analysis or interpretation of any patterns it discovers. “RAGE works in both assisted and unassisted mode—assisted by human experts or on its own,” says Venkat Srinivasan, CEO of RAGE Frameworks. Moreover, RAGE AI configures end-to-end enterprise-grade applications rapidly—that is, in a matter of days or weeks.
Patterns and Comparisons
RAGE AI can quickly identify patterns and comparisons that might otherwise take an inordinate amount of time or be missed altogether. “One example is to think about a merger-and-acquisition transaction,” Srinivasan explains. “If you’re doing a cost analysis for two companies merging, you’d need to review thousands of contracts to identify any synergies. Those contracts come from suppliers, customers, and so on. Those contracts are in many different shapes and forms. We are able to rapidly digest all these contracts and extract the right terms to show the delta.”
What makes RAGE AI stand out among similar frameworks is the manner in which it interprets documents. It uses linguistics-based machine learning developed by RAGE Frameworks to provide a greater understanding of the context of the documents and articles being examined. The framework can extract contextually relevant information to identify any meaning more accurately and comprehensively. “We take the document and understand it linguistically,” Srinivasan says. “We understand and decompose it like you or I would.”
The RAGE AI platform contains a set of 20 configurable engines, including real-time content integration, data access, decision tree, workflow, and an intelligent document builder. They all work together in a configurable architecture to automate knowledge-based business processes. Speed-to-market with RAGE AI is rapid because no coding is required, and applications are assembled using these configurable engines. “Enterprises love it because they can bring their ideas to life and seize the opportunities that are here and now,” says Srinivasan.
In addition to being an enterprise application development platform, RAGE AI also excels at advanced analytics. For instance, a business-to-business (B2B) company used the platform to analyze all its customer data to locate inefficiencies in internal processes and to understand where the firm could improve the customer experience. RAGE ingested all the customer-facing structured and unstructured data for multiple quarters to deliver insights on the underlying reasons for inefficiencies and customer contacts. Consequently, the company was able to identify half a dozen processes for improvement and automation.
“It understands in nuanced ways,” Srinivasan explains. “You can calibrate the degree of impact by having it read all the text in a document. It can interpret superlative words significantly and amplify its understanding of the impact.”
When using the system, users have full control of and full end-to-end process visibility across its multiple components. They can set warning and alert thresholds, audit trails, and access controls. If something is puzzling or warrants further analysis, users can delve more deeply to understand what’s driving any particular outcome. “It’s not a black box,” says Srinivasan. “The user is able to see why the machine is interpreting what it interprets. If users want deeper interpretations, those are completely traceable and auditable.”
When using RAGE AI in assisted mode, business analysts can enter their own rules in the system and edit them to help it recognize various patterns and relationships. RAGE AI will continue to refine those observations and analyses as it works through more data. And since the system’s computational logic is transparent, human analysts can gain a full understanding of its reasoning.
It doesn’t stop there. Applications of the RAGE AI framework include analyzing companies for stock investments, competitive intelligence, and credit-risk assessment and monitoring. Companies can set up the RAGE AI system to manage process orchestration and business process automation. It will automatically aggregate, cleanse, and parse data before applying linguistic learning and analysis. So a B2B company, for example, could have an early-warning system that analyzes its customers’ or suppliers’ business risk every day.
This intelligent machine scans the Internet around the clock to identify, interpret, and analyze the impact of market developments on specific businesses across a variety of industries, and then it generates a dashboard and alerts to monitor enterprise risks. The machine processes millions of documents every day and assesses risks across dozens of business drivers for each company, all in near–real time.
RAGE AI can also be integrated with other systems, such as SAP or an enterprise resource planning (ERP) package, and even use them as steady data sources. “One way to look at it: in the world of big data, there’s volume, variety, veracity, and velocity,” Srinivasan says. “Handling volume and velocity in a scalable manner is part of our core architecture. We process millions of documents each day as they’re published. We handle the variety really well in context, and the veracity in terms of interpretation.”
With the RAGE AI system doing much of the business data analysis, Srinivasan predicts a significant impact on how workers will spend their time in the future. “Intelligent machines or systems can take over the execution part,” he says. “The role of the human will change with all this automation: it will require new skills. Existing jobs will disappear, and new jobs will appear.”
If you have an idea about transforming your business using AI, talk to RAGE Frameworks. RAGE will help you test and rapidly bring your ideas to life.
Keep up with the latest in artificial intelligence at EmTech MIT.
Discover where tech, business, and culture converge.
September 11-14, 2018
MIT Media Lab