Seeing in the Dark: The Future of Automation With Unstructured Data
Let's see what the future looks like for automation with unstructured data.
Join the DZone community and get the full member experience.
Join For FreeAbout 2.5 quintillion bytes of data are created every day according to Forbes magazine. And about 90 percent of that is unstructured data — video, audio, image, email, instant messaging, and other types.
This "dark data" creates a major headache for organizations. 80 percent of business processes today rely heavily on people to locate, organize, and input unstructured data before the process can even begin.
Cognitive automation solutions have stepped in to help. Using artificial intelligence techniques, these solutions learn how to handle semi-structured and unstructured data so that processes can be automated from end-to-end.
Many of the applications for cognitive automation to date have focused on semi-structured data — invoices, purchase orders, contracts, and the like — for extracting the right data to input into enterprise systems. But this is just the start for what cognitive automation can do.
At Automation Anywhere, we see a future where cognitive automation will apply to any dataset —identifying and extracting patterns and intelligently making decisions on these. This frees human workers from mundane tasks so that they can focus on more value-added activities like strategic problem-solving. Let’s look at some examples of what the future can hold for.
Video
Video is used across industries for monitoring, safety, and intelligence. For manufacturing, cognitive automation combined with computer vision and software bots can help automate the quality and safety control process. By identifying defects in real-time, human workers can be notified about potential problems in the manufacturing process and defect waste avoided.
More recently in social media, safety in content has become a major issue. Thousands of people have been hired to manually screen inappropriate content from social posts. Instead, cognitive automation solutions, combined with computer vision, can be trained to identify and remove inappropriate content on a 24x7 basis, applying standards consistently.
Audio
With cognitive automation, organizations will be able to process audio inputs at scale to identify issues early. For example, the audio inputs of large generators often carry signals of impending malfunctions like vibration even before traditional sensors can detect the problem. A cognitive automation solution can be trained to detect these and flag them to human workers.
Similarly, voice assistants such as Alexa and Google Assistant help consumers take actions like playing music and getting the latest news. By integrating cognitive automation with these, a business can better understand the tone and sentiment of the user beyond the specific request, identifying any issues with customer service early and take action.
Image
Applying cognitive intelligence to images can also help keep things on track. With construction — whether residential or commercial — customers often face significant ambiguity on progress and cost estimates. A software bot could receive an image every day from a construction site, identify the progress made through cognitive automation, and send a report along with an updated cost estimate to the general contractor or the owner.
Consumer brands also face significant ambiguity about how and where their brand images are shown and mentioned in content. Together with cognitive automation, consumer brands can monitor images of how their logos are being used online at scale, gaining more detailed insights and helping to ensure their brands are properly represented.
Other Unstructured Data
The Internet of Things generates massive data, which will only continue to grow, and cognitive automation can be used to make sense of it all. A cognitive automation solution that receives sensor data — whether from a smart home, device, or building — can identify patterns and flag issues early to monitoring or maintenance staff to take action.
If that same maintenance worker sees the failure and wants to know about similar instances where that happened, another cognitive automation solution can compare the failure to all of the documentation in a database on other incidents. It can then send a report to the human worker summarizing the cases of similar failures and recommend a course of action.
Conclusion
To summarize, we’re just at the beginning of enabling organizations to "see in the dark" with their unstructured data. Cognitive automation is well-positioned to address this challenge, exponentially improving productivity and enabling new scenarios that add customer value.
You can try out IQ Bot and see how it uses artificial intelligence techniques to handle semi-structured and unstructured data by downloading Automation Anywhere’s Community Edition.
Opinions expressed by DZone contributors are their own.
Comments