Uber is an app that was created to match drivers with people looking for rides. People can open the app, and request a ride. Drivers (registered Uber drivers, that is) located nearby can then pick up the person requesting a ride. After a ride, the driver is paid through the app.

To ensure a more secure experience, a photo of the driver is sent to the passenger. This way, passengers can feel safe that the driver is who they say they are. This may cause problems, as drivers may not always look like their photo. They may have grown a beard, or shaved off a beard, or similar changes may have occurred.

To account for this, Uber decided to add a new feature. Each driver needs to sign in when they are using the app. Doing so will periodically request them to take a selfie. This image is then sent to the Face API for verification. If the verification fails, due to glare from glasses, or something similar, the driver is requested to remove such items.

According to Uber, they spent around 3 weeks implementing the Face API into their systems.

DutchCrafters is an American company that sells handmade furniture. They do have a physical store, but more importantly, they have an e-commerce website. This site contains more than 10,000 products, where each product can be customized.

They had a low conversion rate on their site, and as an attempt to improve this, they used manual recommendations. Manually adding recommended products on each product is rather time-consuming. Looking into their options, they discovered the Recommendations API from Microsoft Cognitive Services.

They were already relying on REST APIs, and as such implementing the Recommendations API was quick. DutchCrafters have stated that they spent 5 days in total implementing the functionality needed.

As their site was already built with ASP.NET and running on IIS, they decided to move everything to the cloud. Doing so has improved their site, and with the addition of the Recommendations API, their foundation has improved.

At the time of writing, they are utilizing the You might like this feature, recommending 10 items per product. They are also looking into adding real-time recommendations, based on users' history, which we have seen is possible using the Recommendations API.

A direct result of implementing the Recommendations API is an improvement of the conversion rate. They have seen a three times increase in the conversion rate, with about 15% of the sales coming from recommended products.

CelebsLike.me is a web application from Microsoft. It was primarily created to show off some of the features of Microsoft Cognitive Services.

The purpose of the application is to find your celebrity doppelganger. You can upload a photo, or use one found online, and the app will match faces found with similar celebrities.

The app takes advantage of the Bing Image Search API, the Computer Vision API, and the Face API. It recognizes celebrity faces in web images. When someone uploads a photo of themselves, facial features will be used to find matching celebrities.

Pivothead is a company working with wearable technology. They have combined eyeglasses with high-quality cameras, providing still images and videos. These glasses allow people to capture vivid point-of-view content of what they see. Pivothead currently has customers in the consumer market, but also in the business market.

Over time, Pivothead had seen growing success, but could not seem to create a device to help visually impaired and/or blind people. They struggled with the technology, as machine learning itself can be quite complex. When they learned of Microsoft Cognitive Services, they were able to reach a breakthrough.

If a person is wearing the glasses, they can slide a finger along an earpiece. This will capture an image of what is in front of the person. The glasses utilize five APIs from Microsoft Cognitive Services. These are Computer Vision, Emotion, Face, Speech, and LUIS.

With the image of whatever is in front of a person, the image is analyzed. The person wearing the glasses will then get the image described through an earpiece. If a person is detected, the gender, how they look, what they are doing, their age, and their emotion is detected and described. If text is detected, it will be read back to the person.

According to Pivothead, they spent around three months months developing prototypes of these glasses. They also stated that they could have done it in three weeks, had they been working with it full-time.

The Zero Keyboard app was created by a Finnish company called Blucup. The company had discovered a common problem for salespeople. They wanted a way for salespeople to capture customer data and generate leads while on the go.

They started developing an app for iOS, Android, and Windows Phone to help solve this problem. The idea behind the app is to record customer information, which is then automatically stored in the Customer Relationship Management (CRM) system.

At the time of development, Microsoft Cognitive Services emerged, and Blucup decided to give it a go. Earlier, they had tried a few types of open source speech recognition software and image analysis software. None provided the quality and features needed.

Using the Computer Vision API, the app can take pictures of business cards or identification badges, and identify text. This data is directly uploaded to their CRM system. By using the Speech API, sales representatives can also record voice memos for each contact.

Blucup states that Microsoft Cognitive Services delivers very accurate data. In addition, they have been able to implement the needed APIs rapidly, as the APIs are a good match from a developer standpoint.

As you can see from all these examples, Microsoft Cognitive Services provides good quality. It is also quick to implement, which is important when considering new APIs.

Another great thing about the APIs is that you do not need to be a data scientist to use them. Even though the technology powering the APIs is complex, we, as developers, do not need to think about it. We can focus on what we do best.