Braintrust AIR is revolutionizing hiring, leveraging cutting-edge AI, trained with humans and years of proprietary data, to match the right talent with the right opportunities, significantly reducing the time and cost of hiring, and improving outcomes for customers and talent alike.
Check out our Frequently Asked Questions:
- What is unique about Braintrust AIR? How is it differentiated from other things on the market?
- Comprehensive: No other platform offers AI-driven job description creation, precision candidate matching, automated screening, and onboarding, all under one roof. It’s an end-to-end solution that the market DESPERATELY needs,
- Data-Driven: 4 years of proprietary data + HUMANS did the training; 750K talent, thousands of completed jobs, for Fortune 500 companies. We’re not just using ChatGPT to write a generic JD and calling it AI-powered. Every aspect of the platform relies on purpose-built models to achieve the specific goals that recruiters care about.
- Unbiased: While humans are inherently biased, our AI is not. It ranks and matches talent for jobs based on discrete skills, not personally identifiable information.
- Ok, but how is it different from the jobs LinkedIn or Indeed recommends?
- These sites don't do any vetting of the talent. They only allow you to post jobs and get applications. Braintrust is a specialized AI-driven hiring platform designed to automate and streamline the entire talent acquisition process, end to end.
- LinkedIn is primarily a professional networking platform with some recruitment features. It caters to a broad range of users across industries and job levels. Indeed is a large job aggregator with a vast listing of open positions, attracting an even wider audience of job seekers and employers.
- How did you train the AI?
We utilized a range of AI techniques like Machine Learning, Deep Learning and Generative AI, and leveraged rich proprietary datasets like client data, job specifications, talent profiles, and client and talent feedback from the thousands of jobs completed and hundreds of talent profiles on the platform.
Our training workflow involved rigorous model validation and selection, during which we assessed performance metrics such as precision and recall to identify the most effective models. Inference accuracy is paramount, and we dedicated significant effort to ensuring that our models deliver precise results.
In addition to quantitative evaluations, we place great emphasis on qualitative feedback obtained through human evaluation. We actively engage internal team members, clients, and talent community members to assess our models' performance. Their insights and feedback played a crucial role in identifying improvement areas and validating our AI solutions' effectiveness.
Ultimately, our overarching objective is to develop AI models that are not only accurate but also robust enough to handle real-world scenarios effectively. Through rigorous training methodologies, performance evaluation metrics, and human feedback, we aim to enhance the quality and reliability of our AI-driven solutions continuously.
- Q: Is company data used in training the AI just for them? Or for everyone?
- First and foremost, all the AI models we develop are created in-house, and all data remains within the Braintrust infrastructure, ensuring that no client data is shared with any third party.
We use data such as company size and historical hiring records for our AI models. This approach allows us to identify patterns and understand our clients' specific recruiting needs.
- Do talent give you permission to use their data? Again just for them, or to make assumptions about future candidates?
- First and foremost, all the AI models we develop are created in-house, and all data remains within the Braintrust infrastructure, ensuring that no client data is shared with any third party. We leverage data including talent availability and past work experiences to inform our AI models. This approach helps us detect trends and grasp our talent's unique requirements, enabling us to speed up the process of matching them with the most suitable job opportunities.
- How do you mitigate bias?
- Braintrust AIR removes bias from hiring. When recruiters get applications, they look through the stack until they come up with a pool of “good enough” candidates. Our AI looks at every single applicant, and our async video screening enables companies to video screen a lot more applicants than they would with a traditional screen - potentially giving a shot to a variety of candidates that would not typically be invited to a screen (if there's only 5 spots available)
- We're not using any PII, such as name, email or phone number, for model training and inference. Key attributes such as the candidate’s resume, profile, and hiring history factor into the candidate’s job match. The Matching Engine provides a summary for every candidate that clearly states why someone is a match or not and what criteria were used in the decision-making. We define the criteria that the model is allowed to use, and we do not use any candidate details that might lead to biased scores, such as the PII mentioned above or anything that may hint at a candidate’s gender, race, ethnicity or other protected identities. Leaving the hiring manager in the decision position.
- OK, but how are you preventing bias in the video interview stage (facial expressions, language barriers, etc?)
- Our approach is to prevent bias at multiple stages. By supporting async interviews, we allow the client to quickly and with minimal cost interview every candidate that meets their criteria on paper, rather than a human picking and choosing who moves to the interview stage. The AI analyzes interview transcripts, focusing on concrete skills and work experience, reducing reliance on subjective interpretation of appearance or communication style. The AI does not consider factors such as name, gender, age, race or ethnicity. We provide clear scoring criteria and are transparent in the feedback to maintain objectivity. Importantly, the video interview is just one piece of information considered alongside the candidate's resume, work samples, and other assessments.
- How does Braintrust AIR support DEI in recruiting?
- Employers come to Braintrust specifically to find diverse candidates, as we enable organizations to screen more candidates at scale. Our network of over 700,000 hails from over 150 countries around the globe. Recruiters use their personal networks, which are inherently biased and closed, in part because they are looking to verify quality. Braintrust does that for you without creating siloed/ homogenous talent pools.
- Can I hire anyone with Braintrust AIR? Or is it limited to tech roles?
- Braintrust’s roots are in the technical and design space, but very quickly we’ve expanded to adjacent fields - namely marketing and program management - so we have a significant advantage in finding and screening top technical talent in these categories. However, the platform's matching and assessment capabilities are designed to be adaptable to other industries, thus, the more roles we fill through the platform, the faster our AI learns and expands into new areas, thus we expect to broaden our industries even more rapidly with the introduction of AI.
- Will Braintrust AIR eliminate the need for recruiters?
- It will make them dramatically more productive, enabling them to focus on higher value tasks; i.e. connecting with the top candidates personally, vs. wading through a sea of unqualified resumes and conducting a series of 15 minute phone screens. Braintrust AIR saves human recruiters hundreds of hours of tedious and repetitive screening work, freeing them up to improve culture and add value to their organization that only a human could. Braintrust AI isn’t built to replace human resourcing specialists; the technology is designed to accelerate the hiring process, connecting customers with the right talent, faster than ever before.
- Is Braintrust AIR going to take the “human” out of human resources?
- Again, one could argue it’s putting the humans back in human resources. Recruiters were never intended to handle the deluge of applicants they face today. There is no way any human can respond to every applicant with the time and attention it deserves. Implementing AI to do all of the mundane administrative tasks required in an HR group, it frees up HR professionals’ time to be more attentive to actual human resource needs / spending time with employees and helping to improve culture.
- Is Braintrust AIR going to eliminate jobs?
- Focus: Our goal is to make recruiters 10-100X more productive; to fix recruiting and help (tech) talent find better jobs, faster than the existing/current drawn out recruiting processes out there
- Big Picture: I still believe AI will create more jobs in the long run than it destroys, but some job roles are going to go away in the next 12-24 months. New jobs that may be hard to imagine right now will replace them. The analogy I use here is when you add lanes to a freeway you actually increase traffic over time. It simply draws out more demand and makes people use those freeways who may not have before. AI is more lanes.
- What led Braintrust to lay off 40% of its workforce in June ‘23? Are you expecting more cuts over the coming months?
- Early last year, we saw around 75% of our enterprise clients freeze all hiring (including using the Braintrust platform). Months prior, we had actually ramped up hiring in anticipation of more growth, not less. So, we were forced to reverse those bets to preserve cash and keep the business financially sustainable. That reversal involved laying off 40% of the Freelance Labs staff (~28 FTEs). Since then, we’ve remained focused on automation & GSV efficiency as we wait for demand to return, which it has over recent months. Now, as we look forward, we’re beginning to expand back to hyper-growth.
- Automation efficiency? Are you saying AI will replace more Braintrust employees?
- AI will definitely pick off the low-hanging fruit. This means we can invest our capital and resources in jobs that truly only humans can do.
- In your FAQ, you say, “Braintrust isn’t taking any fees from talent for each project,” but in fact clients pay a 15% markup, which would otherwise be compensation for the talent, no?
- No. Talent set their market rate. So if a talent says, “my hourly rate is $100/ hr” they earn $100/ hr, and the customer pays an additional 15% customer success fee to Braintrust. What you find with other recruiting platforms and agencies is talent price themselves at $100/ hr, and the agency takes 30-50% away from the talent in fees, or marks them up 2-3x for clients; either way a raw deal for both sides. Braintrust prides itself on transparency, and the lowest fees in the industry bar none.
- The BTRST Token was down significantly from June through December. What’s driving the recent uptick in price? Do you expect the price to continue to climb?
- I won’t comment on our token price. What I can tell you is… [segue back to key messaging].
- Why did you create Braintrust AIR? What do you hope to achieve?
- The Problem You Saw: Briefly outline the following:
- Inefficiency: Traditional hiring is slow, expensive, and frustrating for both companies and talent.
- Lack of Insight: Hiring decisions are often based on gut feelings rather than real-time data.
- Poor Candidate Experience: The drawn-out process, lack of transparency, and generic rejections take a toll on candidates.
- The Solution - What Braintrust AI Does:
- 'Kayak.com' Moment: Just as Kayak revolutionized travel, Braintrust AI brings speed, efficiency, and transparency to hiring.
- End-to-End Automation: Cover the entire hiring journey, from matching to onboarding.
- Empowering Both Sides: Help companies find the best talent faster, while giving skilled professionals a better experience and faster access to great opportunities.
- The Problem You Saw: Briefly outline the following:
- Plans for the future?
- We're incredibly excited about Braintrust AIR's future. Our vision is to transform how organizations find and hire top talent completely, and we're constantly exploring innovative new features and refinements. As our AI learns with every interaction, we expand our matching and assessment capabilities into more industries and roles. Ultimately, we aim to create a global hiring platform that breaks down barriers, connects exceptional talent with organizations worldwide, and empowers candidates to have greater agency over their career trajectory.
- Does the AI consider soft skills? How?
- Yes. In the case of the asynchronous interview feature, AI models recommend soft skills such as communication and storytelling abilities as part of the grading criteria, which clients can then adjust according to their preferences. Following this, we use generative AI to evaluate the responses provided by the talent and generate scorecards for clients.
- Will information about successfully completed jobs be fed back into the model?
- Yes, we're taking into account the actual duration of contracts and clients' ratings of talents' qualifications as factors for determining eligibility for future opportunities.
- For the async interview feature, how do we mitigate bias with non native english speakers, people with disabilities, etc. if at all?
- We guarantee that our AI communicates in clear, concise language devoid of jargon, enhancing accessibility for non-native speakers. Additionally, we display text versions of interview questions on the screen to ensure candidates can read them clearly. Our AI model undergoes training on a diverse array of interview responses, encompassing those from non-native English speakers and individuals with disabilities, aiming to mitigate bias in its comprehension and decision-making processes. Furthermore, we conduct routine audits using human graders to review the AI’s performance, identifying and rectifying any emerging biases. Regular solicitation of user feedback is also integral to our process, allowing us to iteratively enhance the chatbot's functionality and accessibility features, thereby fostering fairness and inclusivity in the recruitment journey.
- How is data stored?
- Data is stored securely on AWS.
- Is data shared outside of the Braintrust environment?
- No
- Outside of the Client’s environment?
- Client data that relates to recruiting is used for AI training