We’re often asked about our tech stack and our approach to innovation. Here’s a look under the hood at how we’re leveraging AI to drive efficiency and growth.
As a COO, my job is to filter out the hype and focus on what actually works. The noise around Artificial Intelligence is deafening, but tangible results are what matter. We didn’t adopt AI because it’s fashionable; we integrated it surgically to solve specific business problems.
Our journey with AI wasn’t a single giant leap. It was a series of deliberate, measured steps. For any founders or operators looking to do the same, I’ll share our approach.
1. The Foundation: Automating the Grunt Work
Before we could even think about using AI for complex forecasting, we had to win back time. The biggest and most immediate ROI comes from automating the repetitive tasks that consume dozens of hours a week. We automated our financial and administrative operations using AI-powered accounting software and transcription services like Otter.ai. We consider these tools to be non-negotiable operational basics.
2. Using AI to Predict Sales and Cut Costs
Next, we moved from defense to offense. We built an internal AI model that analyzes thousands of data points to generate a “conversion probability score” for every new lead, dramatically lowering our cost per acquisition. We then extended this model to our financial operations. As a global business, we use it to analyze currency markets and optimize the timing of our foreign exchange conversions. This system now saves us thousands of pounds every month in fees and poor exchange rates.
3. From Reactive to Proactive: Mastering Customer Retention
Once our acquisition model was humming, we turned our attention to the other side of the equation: keeping the customers we win. We developed an internal AI system that monitors our customer base, generating a “health score” for every account based on their usage patterns. This alerts our team to intervene with support before a customer becomes frustrated. It’s also led to more intelligent conversations at renewal time, which has had a fantastic impact on our customer lifetime value.
4. The Strategic Leap: Owning Our Own AI Infrastructure
As our use of AI matured, we hit a predictable wall. For a high-volume platform like a global job board, the cost of millions of API calls to third-party services like OpenAI becomes a major operational expense. We calculated that at our scale, the investment in self-hosting a powerful open-source model would pay for itself.
After extensive research, we have now successfully deployed a self-hosted instance of Falcon 180B. This was a serious undertaking, and to get it right, we brought in new talent, including a Lead MLOps Engineer and a Distributed Computing Architect, to build out the high-performance cluster required.
The results have been well worth the effort. By running our own model, we are no longer paying per-token costs for our high-volume internal tasks. We are on track for savings in the hundreds of thousands of dollars on API and licensing fees in 2025 alone. For any business operating at significant scale, this is a path that offers complete control over data security and the ability to build a truly unique competitive moat.
Our journey with AI has been about building one layer upon another. Start with the foundation, prove the ROI, and then reinvest those gains into more sophisticated systems. It’s a marathon, not a sprint.
All the best,
Adil Rehman
COO, WhatJobs
Follow Us: WhatJobs
Hey folks, Beau Ireland here. The article mentions automating grunt work as the first step to AI implementation. I’m curious, what repetitive task do you think would provide the biggest ROI if automated in *your* current role? Let me know your thoughts.
Hey everyone, Beau Ireland here. The COO mentioned automating “grunt work” with AI as the first step. Do you think that’s the best place for companies to start with AI, or should they aim higher right away?
Okay, here’s my take:
This company seems pretty level-headed about AI adoption. I appreciate that they started with practical applications before jumping into complex stuff – good foundation. Focusing on automation first makes sense, rather than getting lost in the hype.
Honestly, I’m glad they started with automating the boring stuff. So many companies jump straight to the “sexy” AI applications and forget about the quick wins that actually free up resources. Making the basics work is key.
Okay, here’s my take:
Sounds like this COO gets it. Too many companies jump on the AI bandwagon without a clear plan, so I’m glad they started with the basics like automating grunt work. I’m curious to see how they’re actually using AI to predict sales, though.
Okay, so they automated basic tasks first. Seems like a smart, practical way to get started with AI rather than jumping into the deep end. Good for them for focusing on ROI and not just buzzwords.