You may not know so much about the value of APIs, or what a video interview exactly entails. An API (Application Program Interface) allows the capabilities of one computer program be easily used by another. Two different computers are able to quickly communicate with each other, allowing easy business growth and rapid development. Companies spend $590 million per year trying to cobble together different types of systems. APIs are one way to centralize this kind of information easily, which inevitably improves the efficiency of hiring.
With an API, a business can rapidly introduce new products or improve on existing features already, helping them remain competitive. It allows companies to exist and thrive on multiple platforms, from mobile to browser extensions (like Google Chrome or Safari). It’s usually pretty easy to integrate into a company’s existing system.
For the human resources industry, a fast-moving API creates an opportunity to create a faster hiring process via video. They can move into areas where they may not have considered—like using AI video interview. Companies hardly use in-house created programs, but rather outsource them for ease-of-use and speed. Then, the cobble them together for a product or experience that works.
Twitter, for example, did not always have it’s straight-forward user interface. They used one API, TweetDeck, as the user interface, placing it on top of Twitter technology. Later on they integrated other APIs like Google Maps for an even better user experience. Similarly, HR professionals have the chance to sculpt a system that works for them—a system that saves them time, curates the best applicant pool, and efficiently develops their internal team members.
And yes, an API can help HR
APIs possess the power move ventures forward into the future as it makes it relatively simply to scale across departments and countries. Companies that use “best-of-breed” APIs to structure a bigger system have a special ability to cherry-pick what works for their team and own business goals. APIs are another kind of human capital strategy: it can help talent professionals to broaden the candidate funnel, rapidly identity potential talent, aid in succession planning, and more.
But only if you’ve got the right API.
Incredible strides have been made with APIs, especially machine learning and/or artificial intelligence capabilities. This kind of tool, machine learning as a service (MLaaS), offers companies cloud-based platforms that perform data pre-processing, model training, and model evaluation. It can be incredibly useful for predictive measures when paired with a company’s own data. Four common MLaaS APIs are being rolled out: Amazon Machine Learning services, Azure Machine Learning, Google Cloud AI, and IBM Watson. A popular API for MLaas is IBM Watson. According to Enlyft, 2,939 companies use IBM Watson. The reason for its popularity is its comparative ease in set-up. Companies can build their own data models from scratch or use the global company’s own APIs and pre-trained business solutions. It’s not a bad deal for a company’s data science team that needs a bit a help.
However there are drawbacks for companies who chose off-the-shelf solutions.
Requires an in-house data science professional or team
Services like IBM Watson help companies scale quickly by offering technology that can act as another “team member”. With IBM Watson, a dedicated data scientist professional will be needed during set-up and for constant maintenance. Typically it’s used for companies and departments that are already familiar with using off-the-shelf solutions; but for departments without a dedicated IT team member, integrating IBM Watson may be a challenge. HR departments don’t usually come equipped with someone with these skills, so this could be a huge obstacle in quickly integrating and updating it.
Doesn’t process structured data directly
Organizations are constantly swimming in an ocean of data, whether they realize it or not. Unstructured and structured data are the two data paradigms. Unstructured data is the the pile of content that is unrelated with each other and in no way organized with the other types of data in mind. Think emails, blog posts, video, employee lists, or social media Tweets. It’s a load of information that ’s typically difficult to analyze. IBM Watson could be useful for this kind of information—if a business knows exactly what’s looking for. Structured data on the other hands is the kind of data that organizations tend to deal in. This looks like relational databases and spreadsheets and the types of information that is already and immediately useful to companies. IBM Watson doesn’t specialize in analyzing structured data, which may make it unhelpful for companies who already have spreadsheets and ideas about what they’re hoping to do.
More data, limited resources
Roughly 2.5 billion GBs of data are created per day, and it’s only set to burgeon. Even now organizations across the world struggle to keep up with the demand of ever-increasing data and comparatively little insight. IBM Watson suffers from these types of limitations. Technically, it offers two approaches for experienced and new data scientists: automated and manual. The automated part can solve a problem through three tasks: binary classification, multiclass classification, and regression. For the second approach, a data scientists would be required.
Limited image and video analysis
While IBM Watson does feature an image analysis function API, it definitely lacks video interview analysis offered by other vendors like Microsoft and Amazon. It’s pretty basic the offering: the ability to detect a face, food (because that’s always useful for interviews or any other HR function), text recognition, and a few other features. A bit lackluster one might say. If hiring managers are looking to integrate video as part of their candidate experience, then IBM Watson simply can’t meet that need.
Services like Google, Amazon, and Microsoft offer analysis, but it’s generalized and not user-specific. Microsoft’s Azure—arguably the most advanced—offers a few features on image analysis. It recognizes objects (bravo), detects inappropriate content in images or videos (thanks), has an API that can recognize basic emotion and facial expressions, and is able to build custom image recognition (though a company will need their own data scientist for this). Their Face API is the closest thing in finding the human within the video, as it detects age, gender, a smile, and facial hair. Though the technology is advanced, what matters is context. Their analysis is simply not HR-specific. Additionally it doesn’t specialize in analysis API, but rather it’s simply a feature.
They emphasize their use of MLaaS or AI, but it’s still in the infant stages of video analysis. For recruiters, an AI video interview by them would hardly be worth integrating. So what do they actually do if they don’t analyze the video images itself? Tech giants like YouTube, Facebook, and TikTok leverage the metadata within for their machine learning algorithms. Features like a video's description, tags, and information about where and what time it was uploaded are the main methods of how AI video analysis works. Yes the power to actually the analyze the content of the clip itself is there. However companies tend not to specialize in this as it’s incredibly time and energy intensive.
Son Han, an assistant professor at Massachusetts Institute of Technology (MIT), tells Wired Magazine, “video understanding is so important. But the amount of computation is prohibitive”. These kind of AI algorithms demands plenty of energy—these service giants don’t necessarily want to pour all their time and financial resources into improving one feature of their AI-on-demand products. According to Professor Han, it takes 50 times as much data, and 8 times as much processing power to interpret a video rather than an image. To improve one feature, it simply doesn’t make sense for big tech companies to improve. Rather smaller AI companies are the ones leading the charge in improving video analysis—like for interviews.
Voxel51, a startup out of the University of Michigan, extracts information from video. It’s slogan “we transform video into value” emphasizes how much it can offer companies, particularly for their talent strategies. Video content has become a valuable marketing channels. Companies that use videos on their websites have 41% more web traffic from searches than sites that don’t use video at all. Incorporating video on the candidate experience and employee lifecycle should be immediately considered by management.
Why Video for HR
The average consumer watches roughly 206 videos a month, which means the average person is pretty familiar with and comfortable with the medium of video. Pair comfort with a clear gain for a candidate and a company will quickly outpace the competition for talent. Talent professionals are constantly on the hunt for the most qualified candidates. Not only do they have to compete to get candidates to send in applications, they have to have a smart process that delights a candidate and makes sorting efficient.
Often efficiency and delight do not go hand-in-hand. Video can change that. With an AI video interview, for example, candidates can quickly click and upload responses to a hiring manager’s question. With this kind of ease, the recruitment process works for both candidate and company.
How does the candidate benefit?
If a video interview API is simple to implement, companies gain a clear advantage. To delight a candidate, companies need to choose their API partner carefully. The software solution needs to offer a clear benefit to the employee. It can’t be another hoop for an applicant to jump through. We’ve all been there: clicking through and typing in nine different steps of an online application process---yuck. Video is the new resume. The video interview benefit is multifold:
It’s easy to complete
More than 2 billion people have access to smartphones; the world is mobile and quickly becoming mobile-first. An AI video interview’s software, like Retorio’s, offers candidates the simplicity and convenience of submitting a video whenever and wherever they are. No more need to schedule between time zones or worrying that an interviewer will be thinking about lunch.
Candidates learn about themselves
Retorio offers a best-in-class personality assessment based on advanced psychological research. Companies far and wide invest in developing employees’ strengths and areas of improvement. A candidate will be hungry for this same kind of information; a video interview is longer is singularly a job seeking step, but becomes a win-win situation. Companies learn about a candidate while an applicant receives self-development insight.
Give candidates rapid feedback
An interview is kind of like dating. No one is exactly sure how the relationship is going to progress and any type of follow-up could take a long while. It’s frustrating for both parties. Talent professionals are set with the task of wading through resumes and working with an applicant tracking system (ATS); candidates have to play the waiting game, with usually no feedback on why they’re being asked for an interview or not. With an AI video interview API, Retorio generates a candidate profile within minutes and quickly ranks candidates in terms of skill, experience, or any other parameters set by HR staff. Managers can then quickly see which candidates look most interesting and relative to their needs; candidates receive a fast invite or rejection--the sooner, the better helps them “move on from the relationship", right?.
AI video interview analysis is set to project companies exponentially forward. In the war on talent, talent professionals can amp up their own recruiting strategies via a mean and lean video API. Just as short clips accelerate trust building, companies can leap-frog in building a quality candidate experience and broaden their talent funnel through video. Companies will be able to stand out by seeking these kind of AI solutions, gaining the ability to aggregate and find the talent gems at an extremely fast pace.