The AI ROI Crisis: Why 95% of Companies Fail (And How One CEO Cracked the Code)

One Las Vegas CEO grew from 30 to 100+ employees while implementing AI. His people-first approach reveals why most companies struggle to achieve AI ROI.

McKinsey reports that 95% of companies see little to no ROI from AI investments. Meanwhile, tech headlines trumpet mass layoffs blamed on AI automation.

But Nathan Strum's story tells a different narrative.

The CEO of Las Vegas-based Abby Connect has grown his workforce from 30 to over 100 employees while successfully integrating AI into customer service operations. His approach challenges everything we hear about AI destroying jobs.

The Fundamental Mistake Companies Make

"They're trying to do too much," Strum explains. "Before that McKinsey survey came out, I went to AI conferences and heard from big companies saying the exact same thing—we're not getting ROI. We're having to roll these systems back."

The difference between small business success and enterprise failure comes down to scope. While Fortune 500 companies attempt enterprise-wide AI transformations, Strum focused on one specific problem: professional phone answering.

This mirrors what Praveen Rangnath, CMO of AI company Deepgram, observes: "Real ROI comes when you pick a single painful step in a real process, define the success metric in dollars, ship a scrappy version into production traffic, and iterate weekly."

The People-First Strategy That Works

Strum's success stems from an unconventional approach: bringing employees along for the AI transformation rather than replacing them.

"The first thing we did was start talking about AI and how they can use it in their personal lives," he says. "As they became comfortable with it personally, we could start the conversation about how it impacts them at the business level."

The results speak for themselves. Eighty percent of Abby Connect's staff has completed AI training, with 270 total course completions across the workforce. Rather than downsizing, the company plans to add another 100 AI-integrated roles over the next two years.

What Human-AI Partnership Actually Looks Like

The daily work has undergone practical changes. Instead of spending eight hours on customer calls, AI supervisors now:

• Monitor AI conversations in real-time

• Grade AI performance and provide feedback

• File detailed bug reports for developers

• Train clients on platform usage

• Step in when situations require human judgment

"They're still focused on the core competency of customer service," Strum notes. "Now their job is to give feedback to developers instead of talking to customers directly."

This creates a natural career progression: receptionist to AI technician to AI setup specialist to AI engineer. Pay scales follow this progression, with workers earning 30-40% more in AI-augmented roles.

The Hidden Problem: AI Washing

Strum reveals an industry secret that explains many AI failures: competitors are faking their capabilities.

"A lot of voice providers are using edited audio calls in their demos. They have zero delay in recorded calls they're advertising, but when you sign up for their service, there's the delay."

This points to a broader issue. Companies invest in AI solutions based on misleading demos, then face production reality. The gap between demo and performance kills ROI.

Industries Ripe for AI Collaboration

Beyond customer service, Strum sees opportunities across creative industries. "I was blown away by things I saw at AI conferences. Creative directors explained how AI lets them quickly iterate and take feedback from every corner to create better products."

Rangnath from Deepgram agrees, noting AI's impact across sectors: "In healthcare, doctors leverage AI for transcription and paperwork, spending more time caring for patients. In finance, AI catches fraud and speeds claims processing. In field services, AI provides hands-free reporting and real-time troubleshooting."

The pattern is consistent: AI handles routine tasks while humans focus on strategic, relationship-driven work.

The Skills That Pay Off

For workers concerned about AI's impact, both executives emphasize that success requires embracing the technology rather than avoiding it.

"The same soft skills that were useful pre-AI are useful post-AI," Strum explains. "It's the mindset of being able to use that new tool."

Rangnath reinforces this: "The most important skill is learning how to work alongside AI to get things done faster and smarter. That comes down to being curious, asking good questions, and figuring out how to let AI handle repetitive stuff so you can focus on strategic decisions."

Action Steps for Companies and Individuals

For companies struggling with AI ROI, the advice is clear:

• Start small: Pick one specific workflow problem

• Keep humans involved: Design AI to augment, not replace 

• Focus on production: Build for real traffic, not demos

• Measure business impact: Tie AI performance to dollar metrics

• Iterate quickly: Use feedback loops to improve performance

For individuals, the path forward involves:

• Personal experimentation: Use AI tools in daily life first

• Skill development: Focus on AI collaboration rather than AI avoidance

• Industry awareness: Identify where AI enhances rather than replaces human skills

• Continuous learning: Take advantage of free training resources

The Real Future of Work

Strum's message counters the doom narrative: "We're not just replacing customer service jobs—we're replacing them with higher-paying jobs. We've got teams who are more motivated, making more money, focusing on higher-value work because AI has taken the routine tasks they don't like away from them."

Companies winning with AI aren't those avoiding the technology; they're those leveraging it effectively. They're the ones implementing it thoughtfully, using AI to eliminate tedious work while empowering people to do more strategic, meaningful tasks.

The opportunity exists. The tools are available. Success comes down to focusing on people alongside technology, starting small, and measuring what matters.

The question isn't whether AI will change your business. It's whether you'll shape that change or watch it happen to you.