Beyond Cost Cutting: Finding Real ROI in AI
The following blog post is a summary of our discussion on the podcast produced with the assistance of AI
Prefer to watch?
Fraser Paine, Head of AI at HSO, joins Nathan Bregmen to share how executives can get real ROI from AI:
Filtering hype vs real use cases
The ROI equation made simple
Quick wins vs long-term transformation
Why most pilots fail—and how to succeed
Every boardroom is buzzing about AI. But while executives know they need to “do something with AI,” many organisations approach it with the wrong lens. Instead of solving specific problems, they apply AI like a hammer looking for a nail—or worse, a bazooka trying to kill a fly.
The result? Bloated costs, unreliable outcomes, and pilots that fail to deliver business value. A recent report shows that 95% of AI pilots fail. But there’s good news: success rates jump to 33% with a clear business plan, and to 67% when external expertise is involved.
So how do you cut through the hype and find where AI actually drives return on investment (ROI)? This conversation with Fraser Paine, Head of AI at HSO, offers a framework for executives who want to move beyond experiments and build sustainable impact.
The Problem with AI for Everything
It’s tempting to throw AI at every process, but that’s rarely the smartest path. Many companies are paying thousands per month to have AI models screen for profanity or restructure simple data. The irony? The same models can write scripts that replace themselves—running faster and more accurately, at virtually no cost.
This is where leaders must ask: What truly needs AI, and what can be solved with traditional automation?
Don’t use AI where simple code works. Regex or if-statements are cheaper, faster, and more reliable for well-defined rules.
Do use AI where complexity is high. Tasks like parsing contracts, extracting structured data from invoices, or summarising call transcripts are exactly where AI earns its keep.
By filtering opportunities through both lenses—what shouldn’t be AI, and what AI isn’t ready to do—you’re left with the sweet spot of genuine value.
Cost Cutting vs Revenue Growth
When executives talk about ROI, the default lens is cost cutting. And yes, AI can reduce labour hours, shrink call times, and automate repetitive tasks. But Fraser makes an important point: the bigger opportunity is expanding revenue.
Cost Cutting: Automate post-call notes in a 500-seat contact centre and you might save millions per year. These projects often pay for themselves in 6–8 months.
Revenue Growth: Free up sales teams from admin, double their customer-facing time, and suddenly your best reps can outperform two new hires. That’s a multiplier effect.
AI isn’t just about saving hours. It’s about creating capacity to grow.
The ROI Equation
So what does ROI in AI actually look like?
Lets break it down simply:
ROI = (Value Generated – Operating Cost) ÷ CapEx
Value is the time saved × labour cost.
Operating Cost is trending down as models get cheaper.
CapEx is shrinking as off-the-shelf tools and smarter coding assistants accelerate delivery.
All the arrows point in the right direction: smarter models, lower costs, faster development. But waiting isn’t the answer. Companies that start now build the infrastructure, governance, and change management muscles they’ll need later. Those who don’t will play expensive catch-up.
Short-Term Wins vs Long-Term Transformation
There are two paths to ROI:
Short Term: Automating efficiency in major cost centres like customer service. These projects are measurable, low-risk, and deliver quick payback.
Long Term: Rethinking entire business models. For example, Tekton transformed animal health testing from a six-day lab process into same-day results using AI-powered microscopy. That shift didn’t just save costs; it unlocked precision agriculture and opened up entirely new markets.
Both matter. Quick wins build confidence and free resources. Big bets create lasting competitive advantage.
Why Pilots Fail (and How to Avoid It)
With 95% of pilots failing, the lesson is clear: dabbling isn’t enough. Success comes from structure.
Have a business plan. Move beyond “let’s try AI” to clear use cases, metrics, and ROI targets.
Engage leadership. Shadow IT projects rarely survive. Executive buy-in is critical.
Bring in expertise. External partners can double success rates.
Build foundations. Data platforms, governance, and change management make or break scaling.
AI isn’t plug-and-play software. It’s more like hiring: you recruit, onboard, train, and iterate. Treating AI like talent—not a tool—changes how you budget and plan for impact.
Wrapping Up: Growing the Pie
If there’s one message to leave with, it’s this: AI should help you grow, not shrink. Cutting costs is useful, but sustainable ROI comes when organisations also increase revenue, expand capacity, and design entirely new processes.
Executives who invest now—building foundations, piloting wisely, and scaling strategically—won’t just save money. They’ll unlock competitive advantages that define the next decade.
Start Building ROI with AI
Whether you’re exploring automation in your call centre or rethinking your entire business model, clarity is critical. Let’s talk about your goals, map opportunities to ROI, and chart the roadmap that ensures AI delivers more than hype—it delivers growth.