Artificial Intelligence and Staffing One Year In: Part 2

As we near CONNECT 2024, we recently caught up with Odell Tuttle, Chief Technology Officer at Avionté, as he is set to take the stage once again at this year’s event to discuss the state of platform staffing and how AI (Artificial Intelligence) will impact future product developments at the company.

In the first part of this interview, we revisited how his views on AI have evolved over the past year—what predictions have proven accurate and which developments have taken him by surprise.

In this second part of our interview with Odell, we explore how AI is impacting business, specifically staffing. We’ll discuss the skills recruiters will need moving forward and how Avionté is leveraging AI to create meaningful products that enhance our customers’ lives while upholding the values of our company and our clients. Join us as we delve into the future of recruitment in an AI-driven world and Avionté’s approach to navigating these changes.

Interviewer: Based on our discussion, and considering the entire AI landscape, how do you think new AI technologies will impact business, particularly in the staffing industry? If you were on stage again at CONNECT one year from now, discussing AI and staffing, where do you think you’ll see early traction, and where do you believe it will make the most significant impact in the way staffing firms run their business. 

Odell Tuttle: I don’t think that answer has changed much over the past year. It’s still about the same key points we’ve been discussing and evolving. There are two major categories in staffing software that are crucial for staffing agencies: improving recruiter efficiency and driving talent engagement.

Both are essential components for a healthy staffing business. AI helps with communication at scale, matching skills and resumes, generating content like job descriptions, and saving humans from sifting through large data sets or creating content manually.

AI is also great for customer service. We’re seeing this in our own operations, and other companies are adopting it, either by developing their own solutions or using existing tools. Initially, the main benefits are in scaling communication. Staffing is a highly collaborative business, and being able to use AI to enhance that communication is already bringing a lot of value.

Interviewer: Shifting gears a little bit. Many people in the tech space talk about AI constituting a fourth industrial revolution, particularly focused on robots replacing humans. As you mentioned before, we have empathetic models that are accelerating robotics. Can you start by explaining what tech experts mean by a fourth industrial revolution?

And, in your opinion, if that’s the case, is AI truly replacing humans, or is it just changing the way we work and approach daily activities?

Odell Tuttle: Yeah, well, let me frame it the way I understand it. The idea of a fourth industrial revolution isn’t so much about AI replacing humans—AI is just one component of this next industrial revolution.

It’s about blending human, digital, and biological worlds. For example, working in the metaverse and advancements in robotics are key parts of this evolution. These technologies fundamentally change everyday life and work, altering how we communicate, exchange information, and make decisions.

The third industrial revolution, often called the Information Age, was driven by the advent of computers and technologies like spreadsheets. We’re transitioning from that era now. While we don’t fully know what this so-called fourth industrial revolution will look like, it will likely encompass a broad range of advancements across various industries, including bio research, robotics, space travel, and more.

In manufacturing, for instance, AI and data can be integrated into the design process to create more capable, efficient, and fluid outputs. For us, it’s about driving better efficiency and engagement in the way we interact and work.

Interviewer: So how do these theories and this type of thought apply to staffing?

Odell Tuttle: For traditional staffing as it stands today, AI means better engagement, efficiency, outcomes, and more successful placements. It allows agencies to focus more on strategy and business growth rather than the mechanics and operations. AI excels at tasks like communication, data consolidation, generation, classification, and matching. These tasks can be automated, reducing the need for human intervention in daily, repetitive activities.

As a result, the administrative parts of staffing become easier and more efficient, allowing recruiters to focus on higher-level, long-term strategies while AI handles day-to-day tasks. This shift enables staffing firms to direct their attention toward where they’re taking the business, while leveraging AI to support daily operations.

Interviewer: Let’s drill down on this point a bit deeper. I’m very curious about your thoughts on the mix of skills frontline recruiters and staffing agencies will need to leverage this new technology. Sometimes, we think of technology as dumbing down a job, but that’s not always the case. Take the creation of the spreadsheet which we keep referring to. The advent of spreadsheets didn’t make us dumber; it made us smarter. How do you think this new technology will change the mix of skills needed for frontline recruiters to do their jobs going forward?

Odell Tuttle: That’s what I was getting at about elevating the types of thinking and tasks you do. Your spreadsheet example is perfect—instead of just doing arithmetic like you would have in the past, you’re now writing complex formulas and even programming. This shift created a new wave of FP&A professionals, accountants, and financial analysts who became IT wizards in some respects. The same will happen here.

You’ll need to think about how large language models handle certain prompts and how to ensure you’re putting the right guardrails around these models. As technology evolves, you might spend less time writing prompts and more time configuring or selecting options for your AI.

You’ll have to figure out how to give it proper instructions and ask the right questions to extract the data you need. This changes your role from a transactional, one-task-at-a-time approach to more of an artist, thinking about how to deploy AI components in a way that supports the business.

Interviewer: So, someone could spend a lot of time programming their prompts or thinking about new ways to input them. We’re going to have people who become prompt experts, understanding how to apply these inputs to get consistent, repeatable results from artificial intelligence.

Odell Tuttle: Yeah, I think that’s kind of like the quote from a little over a year ago by Elon Musk, I believe, who first grasped how LLMs were functioning. He looked at it and simply shrugged, saying, “Oh, we’re all going to become prompt engineers.” Which is a significant aspect of AI development.

Prompt engineering is a non-trivial part of AI development because it involves writing in modern natural language, unlike the curly braces and for loops of decades past. On the surface, it may seem easier or more straightforward because it’s in everyday language, but it’s actually quite nuanced and challenging. It’s akin to structuring instructions for humans to carry out tasks, and then framing questions to extract data from those tasks.

With AI, you’re granting it autonomy to perform these tasks, unlike traditional computing. This raises the stakes, making it crucial to implement safeguards and prevent issues like prompt injection that could adversely affect AI models operating in the real world.

Interviewer: So, when I think of someone who isn’t digitally savvy, who’s been doing things a certain way, this is a whole new world. Even though we’re acknowledging that jobs are evolving, they’re not disappearing. We’ll still need people in staffing agencies, but the nature of those roles is changing. They’re becoming prompt engineers.

I understand the resistance and fear that comes with this shift. “I’m not a prompt engineer. I’m not even an engineer at all. I’ve never had any digital background.” What would you say to traditional recruiters who are feeling this way? How can they better prepare themselves for this change?

Odell Tuttle: I don’t think traditional users will need to become prompt engineers. Rather, the idea is that traditional software engineers will increasingly work with prompts as much or more than they work with traditional programming languages in the future. AI capabilities are being integrated into platforms like ours, so users won’t have to do their own prompt engineering. They’ll focus more on tasks like selecting configurations and options built by other prompt engineers or software engineers.

This approach aligns more closely with traditional software service delivery, but it does shift how we build and think about software and the required skill sets. For staffing users, it changes their approach as well. Instead of following a strict, step-by-step process, they’ll manage AI agents—perhaps a federated group—that they’ve instructed to perform tasks like recruiting, selling, qualifying, matching, or analyzing data.

The focus will be on configuring and instructing these agents and then using the information they provide, which will become a more significant part of their role than executing the tasks themselves.

Interviewer: From the recruiter’s perspective, there’s a valid concern about the essential human element in selecting the right people, which AI should not overlook or replace with automated prescriptions. Recent reports, such as those in the New York Times, have highlighted instances where AI has negatively impacted the recruitment process by making poor talent selections or candidates using AI to inaccurately enhance their resumes.

There’s a fear that AI could potentially remove this human element and, in doing so, complicate rather than simplify matters. So, how can we ensure the development of ethical, human-centered AI that enhances the recruitment experience rather than replacing recruiters and potentially making the process impersonal and problematic?

Odell Tuttle: It’s an intriguing question, and we’re seeing it unfold in real-time. The talent marketplace, where job seekers and employers connect, has become an AI arms race. Both sides are leveraging AI as a baseline requirement. If you’re not using AI to find talent, you’re at a disadvantage. Similarly, job seekers using AI to navigate job opportunities gain a competitive edge. This raises the bar for how we operate within this evolving landscape.

However, in any arms race, there’s a tendency to develop countermeasures. I believe it will eventually settle into a healthier state where selection is based on genuine data, qualifications, abilities, and interests rather than just presentation. This shift will encourage a focus on substance over superficial enhancements, akin to how resume services once operated. Employers, too, will refine their approach, emphasizing structured data and fair comparisons facilitated by AI’s organizational capabilities.

Overall, while the path forward may not be entirely clear, these advancements are likely to lead us towards more standardized and equitable practices in talent management and recruitment.

Interviewer: So, right now, what are the most exciting uses of AI in staffing?

Odell Tuttle: Yeah, right now it’s all about chatbots and insights. Their ability to interact in real-time with people, monitoring large-scale interactions, and handling complex decisions and meaningful conversations without needing a massive workforce. They can gather, classify, and summarize vast amounts of information efficiently, enabling quick decision-making and constant connectivity. This capability is crucial for talent engagement, allowing recruiters to stay connected and make informed decisions swiftly. These are the most compelling areas of interest.

Interviewer: So, as the CTO of Avionté, how are you integrating AI into your software development process? How has it transformed your approach to software development both currently and looking forward into the future?

Odell Tuttle: For us at Avionté, our approach is similar to that of many software-as-a-service organizations. Initially, we’ve integrated AI into test automation and software development tasks like code generation and understanding. However, where it gets more interesting is how we’re embedding AI into our applications to deliver user value, enhancing efficiency, and simplifying traditional software complexities.

Currently, we’re in a phase I call ‘widgets and chatbots,’ where we’re rapidly deploying intelligent assistants into our products where data-rich interactions are crucial. Looking ahead, we envision progressing to a ‘full copilot’ stage, akin to fictional icons like R2D2, where AI provides better-than-human analysis and support alongside users. Ultimately, we aim for AI agents capable of autonomous action, strategy creation, adaptation, and return. This roadmap guides our development, aiming to evolve our tools from simple chatbots to advanced autonomous agents.

Interviewer: That’s fascinating. So, how is this going to affect the overall product design process? What key considerations do you have when you’re implementing AI in your product design for software applications?

Odell Tuttle: Well, it’s already starting to change some elements of feature design. We’re now thinking about prompt configuration and how you might delegate a task to an AI component. Some of this is going to take a while to sort out in the way we do application design and as design patterns are starting to emerge.

Concepts like retrieval, augmented generation, and other design patterns are emerging from the integration of language models into software applications. If we look back historically at how design patterns evolved in software development—such as the Gang of Four patterns that have been foundational for decades—these frameworks took years to develop and refine, guided by pioneers like Martin Fowler.

Today, the software development and user experience communities are similarly exploring AI. They’re discovering what works well, what doesn’t, and gathering insights to shape future tools and methodologies. This process, akin to what we might term ‘AI-first development,’ is likely to evolve over the next few years. It has the potential to fundamentally transform how we build modern software applications, moving beyond the graphical user interfaces of the past, which originated from Xerox in the 1970s, to embrace more innovative and contemporary approaches.

Interviewer: Ultimately, though, all companies have missions, values, and goals. So how do we make sure that these AI developments align with our company’s mission, values, and goals as we integrate them into our applications?

Can you elaborate on Avionté’s mission and goals and how they guide the integration of AI in your development process?

Odell Tuttle: That’s a big question that’s on the minds of many in the software community right now, especially with the integration of AI. When we think about AI, it’s not just about programming anymore—it’s about how it influences human interactions, makes decisions, and represents our businesses to the world. There are several critical factors to consider. Trust is paramount. Is our AI implemented ethically? Are its intentions clear and honorable? Transparency is key too. Do we know how the AI model was created? Are there biases in the model, and how are decisions being made?

These aren’t just ethical considerations; they’re increasingly mandated by legislation, especially in areas like hiring and human resources. We need to ensure our AI systems are safe, reliable, fair and unbiased. This means working closely with vendors who prioritize robust testing, transparency, and understanding of how their models operate. In sectors like human capital management, these issues are particularly critical for building trust and ensuring ethical AI deployment.

Interviewer: So, I’d like to follow up on that because it’s always been a pressing issue, especially with government regulations struggling to keep pace with technology. With past tech revolutions, we’ve seen government policies playing catch-up.

Given the current rapid pace and the inherently reactive nature of government regulation, especially in technology, how do you navigate potential future challenges as a software developer or company leader? Are there any insights or advice you can share on handling uncertainties around evolving guidelines?

Odell Tuttle: Well, I think, like in any field, choosing partners who are deeply engaged with these issues is crucial. When you’re running a business, you can’t afford to have a team of lobbyists and lawyers in every state capital and in Washington, D.C. to navigate regulations. That’s why it’s important to work with partners who prioritize compliance and rigorously test their software, ensuring it’s certified and trustworthy.

For instance, we use Anthropic and their language models extensively in our products because they’re a public benefit corporation with a clear mission focused on safety, security, and ethical implementation. They continuously adapt their practices to meet regulatory requirements and uphold trust in critical areas like staffing.

As I’ve learned through experience, it’s not just about checking boxes or flashy marketing. It’s about understanding a partner’s mission, vision, funding, and long-term goals. These factors give insight into their future direction and whether they align with your own business goals. It’s a significant commitment to integrate software, so it’s essential to align not just with their current offerings but also with their future aspirations and where they see themselves in the next five years.

If you’d like to join us at CONNECT 2024, where Odell will speak in greater depth about AI and its future implications for Avionté and the staffing industry, register here. The event will take place at the Hilton Downtown Minneapolis from July 29-31, 2024, and registration is free for Avionté customers.