How Proshore brought AI assistance into interviewstream
When interviewstream saw a chance to use AI to enhance the interview process in a positive and ethical way, they turned to our remote dev team.
Talk to Jeroen, our Accounts director, to see if our ready-to-code dev teams are a fit for you.
Recruitment platform, interviewstream, has a strong suite of tools to simplify the hiring process and optimize the candidate experience. They also have a pre-existing partnership with Proshore.
So when interviewstream’s forward-thinking CTO, Ryan Royal, saw an opportunity to use AI to enhance the interview process in a positive and ethical way, he turned to Proshore to help ideate, iterate, and launch new functionality – in a matter of weeks. Here’s how it happened.
How to introduce AI in recruitment without enforcing bias?
Right now, almost everyone in the tech world and beyond is talking about ChatGPT and the use of AI in general. There’s no question that it’s going to be a game-changer. The question is: how are companies going to use it ethically, and to their advantage?
This dilemma is especially true in the recruitment industry, where legislation introduced in the state of New York in 2023 prohibits the use of automated employment decision tools (AEDTs) unless a bias audit has been conducted. That’s to avoid unintended bias as a result of machine learning (ML).
At the same time, recruitment is first and foremost a people business. Technology needs to support and enhance human expertise, and not diminish or try to replace it. For these reasons, introducing AI into the recruitment process requires careful consideration and execution.
Ryan Royal and the team at interviewstream had been looking at the potential of AI for some time. However, due to their product’s focus on the screening and evaluation of candidates, they wanted to ensure that any new AI feature did not introduce or reinforce bias in that process.
What they needed was AI-driven functionality that could add value for recruiters but avoid some of the pitfalls associated with its use in recruitment. That’s when Ryan pitched the idea of AI-generated interview questions.
Playing around with ChatGPT, Ryan and Principal Product Manager, Faith Peterson, sketched out an idea for the new feature. All they need was a highly-skilled and experienced development team to build it.
AI interview question assist feature
It was important for interviewstream that when they brought in AI functionality it had an obvious benefit to their customers, it was aligned with their company values, and also met the highest ethical standards. And this use case fitted the bill.
At the time, Proshore’s development team as a service was the only one working on enhancements to interviewstream’s core platform. It’s a role that gave them total responsibility for the new capabilities of interviewstream’s customers. That naturally grew to encompass the new AI Interview Question Assist feature.
When the partnership first began, the Proshore team was used to quickly adapting to changing priorities and business needs. And they were genuinely excited to get started on the new AI feature. Right away, they conducted an initial evaluation and started experimenting with the OpenAI API.
To make innovative use of AI within interviewstream the Proshore development team needed to learn new skills and knowledge – which they did. In fact, every person on the development team – Santosh, Anand, Madhusudhan and Shyam – put time into the research, going through the documentation to discover the best ways to build out the functionality. This gave everyone on the team their own perspective, so they could help each other and quickly generate solutions to any issues that arose.
During the development phase, there was a strong collaboration between interviewstream’s Product Owner and the Proshore team. Within a couple of days, Proshore had a proof of concept. From there, it was a case of working on how to best express the functionality. Working alongside interviewstream, the Proshore team iterated on the fundamental idea of giving users guidance on prompt construction for the AI in order to generate usable, targeted interview questions. The output was ten targeted questions for recruiters to add to a question bank, or include in their next interview.
Thanks to Proshore, it took just three weeks to go from ideation to production. That’s largely because the team were able to quickly adapt, quickly understand the tech, quickly implement the proof of concept, then quickly establish and iterate on user interface (UI) ideas.
Continually improving the AI feature
As a result of their partnership with Proshore, interviewstream has a new AI-driven feature that’s available to customers. As interviewstream continues to iterate on the functionality, Proshore team members, including Web Developer, Santosh Pandey, bring new ideas to the table.
One of those ideas was to add ‘request history’ so that recruiters can make use of previous requests – without having to save or download AI-generated content. The team has also added download and clipboard functionality, making it easier for recruiters to incorporate the content into their interview processes.
As the partnership continues, there’s strength in collaboration. Together, interviewstream and Proshore are taking the business problem and combining ideas from the product management, and the developers to create the best possible solution.
Off the back of the success of AI Interview Question Assist, interviewstream is looking for further opportunities to ethically incorporate other AI functions into its offering, with the expertise of a Proshore development team as a service.
Looking for a highly skilled and cost-effective offshore software development team? Book a call.
We can help you get started in 15 minutes.