Big data has been a buzzword for over a decade, and people are often still just as mystified about what to do with their data now as they were back in the 2010s. Trying to capitalize on their newfound commodity, companies producing large quantities of data needed people who knew how to work with it, leading to data science becoming recognized as an emerging discipline.
As one author explains it, data scientists “combine the skills of software programmer, statistician, and storyteller/artist to extract the nuggets of gold hidden under mountains of data.” Companies pay top dollar, sometimes to their detriment, for this valuable set of skills that will help companies leverage existing data to create business growth. Businesses can end up hiring data scientists before they are ready to dive into the world of AI and machine learning, resulting in a highly paid team doing simple data integration and reporting.
While it is important for any company to build a proper foundation that can provide quality data to data scientists, that’s not our topic today. (Read about those building blocks here.) Instead, let’s look at available AI tools that can give companies a quick return on investment. Technology has become so advanced that we’ve reached a point where companies don’t always need a team of data scientists to be able to utilize AI capabilities. There is an appropriate time for a top-notch data science team to join a project. Their specialized skillset is unmatched and required for advanced AI and machine learning projects, but bringing in data scientists no longer needs to be the knee-jerk reaction.
Historically, if a company started asking questions about prediction, anomaly detection or identifying things in pictures, they were entering the domain of data science. However, Microsoft has put those capabilities into the hands of any developer with the launch of Cognitive Services. An application developer or business analyst, any technical person with no machine learning expertise, really, can now tap into the world of AI solutions.
Cognitive Services are broken into five categories:
Analyze images and videos for desired information; image recognition; identify people and emotions
Transcribe audible speech into readable, searchable text; convert text to speech; real-time speech translation; speaker recognition
Build natural language understanding into apps, detect sentiment, key phrases and named entities, translation
Anomaly detection, content moderation, personalization
Crawl PDFs, scanned documents, etc., and search those files
This is a whole new set of data-science-like capabilities that allow companies to take advantage of the complexities and capabilities of artificial intelligence without needing a team of data scientists.
Blueprint, for example, designed and built an end-to-end reporting and invoicing process for a client that eliminated time-consuming manual processes and the need for an expensive third-party vendor. Azure Cognitive Services enabled the solution, which Blueprint planned, built and launched for the cost of one year’s license with their vendor, to scan physical field tickets and create metadata to match each ticket to the appropriate data reps entered in the field. This then triggered an invoice to be sent to a customer. (Read the whole story here.)
Going even further, Uber used Cognitive Services’ face recognition function to help ensure the driver using the app matches the account on file with the company. It’s important to Uber that the driver that shows up for a ride matches the driver account on file. The company says that’s both to protect riders and help ensure the driver’s account hasn’t been compromised. So, Uber added photo-matching technology to its driver screening methods, requiring a driver to take a selfie periodically to compare the actual driver with the driver’s photo on file. The company considered developing the technology but realized that would require a lot of time. Uber was able to integrate the Cognitive Services API to its platform in just three weeks (it was still in preview at that time).
The opportunities that come from these capabilities are nearly endless. An oil production company could monitor gauge readings from images and set alerts for abnormal readings. A manufacturer could set an audio alert to notify someone if something in the factory is not running correctly because the audio in the facility changed or set up image alerts to notify someone if people on the production floor are not wearing hard hats. A quick-service restaurant could use images to collect visitor sentiment and demographics about their in-store customers to better understand their customer makeup during different dayparts.
Whether you are wondering how best to utilize these tools or want to focus on more advanced machine learning, Blueprint’s expertise in application development and data science will help you find the best way to solve your critical business problems. Let’s start a conversation about your business needs and how we can help you identify and develop the best solution.