Boyd Davis, CEO Payfederate
Just over 3 years after the launch of ChatGPT, a very clear narrative has emerged that AI will kill the software industry as we know it. Developers are at peril because Claude can create applications easily. Software firms will go out of business or dramatically downsize. Business professionals will become vibe coders, creating on the fly agents at will to automate. I’ve been at the forefront of technology for over 35 years, and this narrative hasn’t resonated with me. Something big is clearly going on, but the hype has definitely outpaced the reality. As the CEO of a SaaS startup, the question is existential for me, and I’d like to share what I’ve concluded.
As a bit of background, I started my career in 1990 at Intel. The start of my career was the pinnacle of the personal computing revolution, which coincided with the shift in application architecture from mainframe/mini to client/server. I watched the emergence of the internet/cloud, followed by the rise of mobile computing, followed by the emergence of big data. These are the major shifts of our lifetime. Each (client/server, internet, mobile, and big data) were transformative. Each was also a critical precursor to modern machine learning and large language models.
In each transition, there has been massive displacement. New companies thrived and traditional players faltered. However, the overall market for and impact of technology has been steadily growing. Through it all, the demand for advancement has outstripped any capacity of technology. Let me give you a small example from my time leading data center chip marketing. Prior to the emergence of cloud computing, most servers in the world were utilized less than 10%. Cloud computing offered a discontinuity, where server utilization could increase 10X. Combined with Moore’s Law doubling performance every 2 years, we were genuinely worried that we’d deliver too much compute capacity for the demand. What actually happened? Intel was disrupted by Nvidia, who ended up delivering even more compute capacity at an even lower price. Fast forward and Nvidia is worth $4T+, and hardware still can’t keep up.
Will AI impact jobs? That is inevitable. In any transition, there are winners and losers. In a transition as big as this, there will be more and bigger winners and losers. However, there is not a shred of doubt in my mind that the opportunities made possible by AI will be positive, – on employment, on the growth of the economy, and on the impact to human lives.
Is software as an industry one of the big losers in the transition? It could still be true that AI is accretive to employment, but that software companies are a casualty. Anthropic, OpenAI, xAI and others can simply provide an agent foundation, and applications of any type (SaaS or otherwise) will become less common as people spin their own. Most of us using AI day to day have lots of little epiphanies about how tools can impact our personal productivity. We can all read anecdotes about comp professionals vibe-coding a solution for a merit cycle recommendation tool or a survey job matching tool. I must conclude it’s possible that DIY in the age of AI will impact software revenue.
Having said that, make vs. buy has always been a consideration in enterprise software, and most software vendors are way more advanced in their understanding and use of technology than enterprises are. We have all the tools available to end users, with a whole lot more. We were also born in the age of AI – our team is tiny, and we have zero sales development reps. We won’t raise a ton of money just to spend it on marketing fluff. AI enables people with the problem we solve to find us, and us to serve those customers with extreme efficiency. In short, we’re designed as a company for the software market of the future.
So what’s my conclusion? AI will put employees out of work and companies out of business. AI will simultaneously drive employment and create new companies. Those that thrive will have a few characteristics:
1) Companies will be extremely lean, focused on their core differentiation
2) Software will be AI-native, meaning designed from the ground up for AI
3) Solutions will be more complete – governance, security, permissions, integrations, workflow orchestration
My biggest suggestion for users: stay agile. Avoid 3-year software contracts. AI should make switching costs lower and you want to avoid lock-in. For rewards professionals, separate the decisions on market data (which is an ingredient) from the tools you use to manage pay.
I encourage experimentation and vibe-coding for total rewards professionals. Our profession has been leaders in HR analytics for decades using Excel, which in many cases gets way more complex than Claude or the alternatives. Once a rewards professional sees the individual productivity gains, they’ll be in a better position to adopt an enterprise solution like Payfederate.