How Splunk Built a Transparent AI Philosophy for Hiring

This week's episode was originally broadcast in January 2025 during our Social Talent Live webinar. Visit socialtalent.com live to make sure you catch our next event.

Ali as our last panelist, it's a pleasure to welcome. Ali is the Senior Manager of Talent Assessment at Splunk. If you haven't heard about Splunk, maybe Ali might kick off by telling a little bit who they are. And we want to talk about how do you embrace AI in hiring? So Ali, so good of you to join us from the east coast of the U.S. i wonder if you might introduce Splunk to those of our listeners who may not be familiar with the organization and why particularly AI is important to you and your organization.

Absolutely. So Splunk helps to make organizations more resilient. We build and deploy AI and machine learning solutions for our customers to help provide visibility and context across their most complex systems.

So Ali, Splunk has developed a clear AI philosophy for hiring aligning HR legal, DI teams. I first had the pleasure of hearing about how you implemented this and what you did live when we were in Nashville over the summer. But I'd love to understand what were the driving factors behind this philosophy and how did you ensure alignment across the different groups that were involved?

Absolutely. So this really started as a need. When I was hearing from multiple hiring teams, I was sitting in on interview debriefs with a number of teams and picking up on common questions, common concerns about how they were handling candidates, who they were suspecting that were using AI, whether it was resume enhancements or in the interview itself. A lot of the same things that have been discussed in chat and through the two other panelists earlier today. And the hiring teams were handling these same situations in a variety of different ways. And while there are a number of things that we can do in the moment, I wanted to take a step back and say really, the best way that we can address this problem was to identify the root and really try in earnest to prevent it from happening, which was the impetus for this initiative. So we wanted to get out into the front and at the point of application, let our candidates know what is our philosophy on how we expect them to or how we expect them not to use and how we permit them to use AI.

So that transparency piece is what really struck me, struck me about your approach because certainly summer 24, when I, when I heard you share this, you know, you were the first company I'd ever heard of to take such a, I'll call it proactive approach to this Rather than a defensive, what is this? We need to be fearful of it. Let's drop a risk charter. You really seem to get out ahead of it. And that transparency piece, I think is the cornerstone of this. Can you walk us through how you communicate your expectations and let's call them policies or guidelines to recruiters, hiring teams, but also candidates, to ensure we're all on the same page and everyone understands what's going to happen?

Yeah, you're absolutely right. And last year when we were thinking about this, there really were three approaches. You can keep your head down and see what everyone else is going to do and not say anything to your candidates. And then that was the situation we were in where hiring teams were saying, well, what do we do now? You can take the approach of fully banning it, but then you run into the issues that again, we've talked about in chat of, well, how do you reliably detect that? And then you can take the approach we did where we said, okay, we're going to embrace and we're going to coach. And so we include the philosophy right on our career site. So as our applicants are looking for their next opportunity with Splunk, there are multiple opportunities, there are multiple places on our career site where it is linked. We created sort of a cheeky video that goes over the do's and don'ts and that is available on our LinkedIn pages, on our socials, and it is available in our candidate communications as well. And we trained our recruiters to become comfortable with these guidelines so they feel really empowered to have the conversation not only with candidates, but also with their hiring team. So when our recruiters are meeting with their hiring leaders in that rec strategy meeting, this is really a core part of the conversation when they're talking about the overall hiring screening process.

So you emphasize as an organization that AI should enhance candid capabilities rather than replace them. That's been a subject in the chat for the last 45 minutes or so. Ali, you've been joining in to that conversation as well. Folks pushing back, saying, hey, why can't we embrace this? Let's be less fearful. Can you share maybe some specific examples of how exactly you guide candidates to use AI appropriately during the hiring process? What do you encourage them to do?

Yeah, so we've talked about resumes, some. So I don't want to get in too much. I mean, I think we all agree spell check, grammar, innocuous tasks like that, that's a good use. Misrepresentation, making up experiences, that's a bad use. One use that I really like is preparing for interviews. So if you think about the idea of a career coach, no one is going to say, if a candidate takes a job description and brings it to a career coach and says, what sort of job, what sort of interview question should I expect to be asked during my interview? No one is going to consider that to be treaty, to be cheating. But access to a career coach is not universally accessible. So with AI, with ChatGPT or any other AI tool that you're using, you can take that same job description, you can put it into the tool and you can say, what sort of interview questions should I expect to be asked? That is a great way to democratize access to a resource that otherwise probably would have been pretty limiting. I think that's a really powerful tool. It's a really great way to increase DEI initiatives and it's really a use case that was at the forefront of my mind when I was thinking about where I why I don't want to ban AI just wholesale.

I love that we saw this kind of challenge with candidates, folks from marginalized backgrounds, people who are doing, who are, you know, internal hires who haven't done an interview in seven or eight years. We saw this in Social Talent as a challenge and we developed in our platform, for example, a way for you to practice, literally voice practice. You can go do an interview as a candidate a hundred times if you want to, which I know as a recruiter, I used to work in staffing agencies, Ali, and it was the big thing you did with folks who were nervous or inexperienced was you, you just did mock interviews. Right? Very hard to scale that, to give everyone that approach before AI. But I agree with you, it's a wonderful approach. Actually, I'm doing a session in Dublin tomorrow for staffing recruiters on business development and how to do cold calling. And actually we're going to be demonstrating a tool where you can practice a cold call, where AI will be the potential person you're cold calling and they'll respond dynamically to what you say. And this ability to practice is brilliant. It's not cheating, it's practice. It's scale practice. And your point about democratizing it is wonderful. Everyone can access this with even a free ChatGPT account. You can, you can get that up and running.

That's exactly right. And that is enhancing, that is not replacing.

I love that. And it's come up in the chat, by the way, where we've, you know, a lot of people are saying, well, you know, how do you basically make sure candidates don't do the things we don't want them to do with AI, the blatant cheating are our fabrication. So talk to me about the policies on the other side that you, you say we, you know, candidates should not use AI for. And how do you communicate those?

We communicate them through the same mechanisms, through the, through our policies that we put on our career site and through the same communications that we have in the same behavioral interviews. We absolutely tell our candidates that reading from or referencing AI produced content is not a true representation of their skills, their qualifications, potentially even their experiences in technical interviews. Using AI to check for syntax errors, perfectly acceptable. But copying and pasting full code segments and presenting that work as your own, that's not acceptable. When you are in a technical interview and you are collaborating with an engineer from Splunk, being transparent about how you are or are not using an LLM is absolutely an expectation. And similar to what I believe Paul mentioned, we empower our interviewers to be transparent with our candidates, asking them how they are using it and setting the stage. If that is reaching our expectations, if it is going against how we are expecting our candidates to use the LLM, setting that stage, giving them benefit of the doubt, giving them a chance, and then reinforcing it.

Liam, who's listening and on the chat, has probably best encapsulated what I think is the audience feedback on what you're saying here, Ali, which is, you know, genius of declaring expectations to candidates before the interview levels the playing field, builds trust, and shows the employer acknowledge the impact of Gen A technology. And I think this is particularly pertinent for an organization that, like yourselves, that are trying to hire folks who ideally embrace AI and understand AI as well, which is really, really important. So kudos for that. And I think that's what struck me as well, Ali, is that, you know, this is a real positive step. Like I think anybody can take this step. It's still surprising how few organizations do take this step because it's, it's, you know, they're still fearful, they're still in those two other phases, so to speak, or levels that you mentioned. I'm wondering, by the way, you mentioned, obviously, the role of the recruiters and the hiring manager. We talked about how you set expectations for the candidates, but what kind of guidance or tools have you provided to the recruiters in your team and the hiring managers to help them have the confidence in navigating AI in the hiring process and being comfortable having those conversations with candidates?

Sure. So AI, of course, is a new way for Candidates to get their foot in the door again goes back to the conversations we've had live on this webinar and also in chat. It's but it also, it highlights the need for interviewers to be fully confident in their ability to deliver a high quality interview. So they need to know how to align job expectations with questions and most importantly, they need to know how to probe for detail if they are getting questions that are well structured but generic. As another panelist said earlier, they need to know how to probe and get the actual details to validate the that the candidate has the knowledge, skills or abilities that align with the expectations on the role. And so that's really what we are focusing on for our interviewers. We're not trying to have them be vigilantes and and experts in detecting AI. We are trying to have them be experts in detecting skill validation.

I love that. I know. If you don't mind me saying, Splunk and Cisco, that you're part of the Cisco family have been really successful in augmenting our program for your hiring managers and recognizing years ago that it's important enable folks to be able to probe. And I think that to me has always been the best way you know how to interview. Well, if you're a good interviewer, then you need not to worry about AI cheating because you have a structured process that gets to the truth. And I've never seen a candidate using all the AI tools available get past that if you have a good interviewer. So it still comes down to that human skill. Speaking of that human centered practice, I know Splunk embraces AI as a technology forward organization, but you also strive to balance this with very human centered practices. You know, you still do in person interviews when, when necessary or as appropriate. So how do you try and strike that balance in a way that still aligns with your hiring goals?

Sure. So AI is the future of Splunk as a company. And no one can deny that AI is changing the talent acquisition landscape, but we will always have a human in the loop. That is a core philosophy of our talent acquisition program at Splunk and at Cisco when it comes to in person interviews. When feasible, they're a great solution for a final round interview. But it's not always possible for all teams in today's virtual world. And additionally I would caution teams from over indexing out of fear of AI. You know, we are not going back to the days of paper resumes. So I don't want people to think because they fear AI. We must go back to all in person interviews. That obviously has a huge cost to the business, but it also has a very big cost in terms of dei when we are shrinking the applicant pool by saying we are only going to look at applicants who live geographically near the near the office.

So we have a customer who shared with us that their CEO in a panic said we need to go back to in person interviewing for everything. And the talent acquisition team said, no problem, we'll cost that for you. And they went back with a cost of an extra $8 million a year. The CEO quickly changed their mind and decided we'll try and find a technological solution. So I'd 100% agree with you, Ali, but I see it as a knee jerk and lots of leaders may do that and we need to be ready for them. I really, really appreciate your insights. The audience loves it. And ask one final question. Maybe a final takeaway for leaders listening to this who are looking to embrace AI and their hiring process who like what you've done. Team in Splunk, what steps would you recommend they take to establish those clear policies and ensure transparency? What lessons can you share with them?

First, I would look internally at your own organization's AI philosophy. If you don't have one, you're going to need to start there. Your philosophy should align with your organization's stated values and your lived employee experience. And lastly, don't try to take this on in a silo. You are going to need partnership from a number of teams internally. This should not be a TA owned and driven initiative that does not bring in your legal teams, your security teams, your partner organizations from dei, et cetera.

Ali, thank you so much. I think that's been a hugely helpful way to end our session. Tons of advice there. I can't see a reason why anyone here wouldn't take your advice. Use that as the template. I love what your team is doing, Splunk. I love that you're leading in this as well and coming out as an organization hopefully will set the scene for the rest of ta. I thank you so much for joining us this morning from the East Coast. I hope you have a great day and we'll hopefully have you back on soon.

Thank you so much.