The IRS's AI Playbook: Lessons in Turning Tech Dreams into Profits
As businesses pour billions into AI, only 15% of execs see profits. The IRS's approach shows how targeted AI use can close this gap. Here's what crypto can learn.
JUST IN: AI investment isn't always adding up. With companies spending billions, it turns out only 15% of executives see profits from their AI efforts by 2025. That's a wild mismatch between expectation and reality. But there's a surprising role model for getting it right: the IRS.
How the IRS Found Its AI Groove
The IRS, not typically seen as a tech trailblazer, embarked on one of its largest modernization efforts in 2023. The goal? Use AI to overhaul taxpayer services. Each AI dollar spent needed to show a tangible return, given that this cash came straight from taxpayers. It's like managing a giant crypto portfolio where every trade's outcome is scrutinized.
The IRS pinpointed pain points first, applied AI practically, and kept measuring impact. This wasn't just about flashy tech. It was about results. They started with their hotline, notorious for its frustrating backlogs and wait times. Deploying domain-specific AI cut response times from 28 minutes to just three. Millions more calls got answered live too. Traders are watching closely.
The Triple Shot: AI Value in Three Lanes
For businesses wanting to see AI's potential, the IRS's approach is illuminating. Their strategy unfolded in three lanes. General-purpose AI aimed at boosting daily productivity, domain-specific AI focused on precision tasks, and custom AI tackled unique challenges.
General-purpose AI, think large language models and agentic workflows, can help free up employee time for more strategic work. But the IRS found you can't just throw AI at a problem. Especially when errors are costly, as in tax admin. Businesses need to shift from broad tools to those built for precision.
Domain-specific AI is important where accuracy is king. Whether it's for legal research or tax analysis, these tools are built around reliable data and come with safeguards that boost reliability. They also tend to pay off faster because they're designed with specific workflow needs in mind.
Custom AI: The High Stakes Gambit
It's tempting to jump straight to custom AI solutions. But the IRS discovered these come with hefty price tags and complexities. Custom AI should be the last stop on the AI journey, after exploring other lanes. They used custom AI to manage cases across millions of transactions, preventing billions in fraud and mistakes in 2024. But it only made sense because they had the backing of general and domain-specific wins first.
In the crypto world, this is like building a bespoke trading bot. It could pay off massively, but it's risky and resource-heavy. Starting with adaptable, off-the-shelf solutions can generate the insights and ROI needed to justify such a leap.
Innovate, Iterate, and Repeat
To make AI pay off, you can't rest on early wins. The IRS learned this as they explored using AI to maintain, not just replace, outdated code. It's less risky and more scalable. A smarter move. Companies need to pressure-test their AI strategies constantly. Are you just rushing to integrate AI, or are you doing it purposefully?
The real winners? They'll be the ones who harness AI with intention across all three lanes. It's not about having the biggest budget. It's about knowing which AI tools fit where.
Crypto's Takeaway
For the crypto market, the IRS's story offers a roadmap. AI's not just a buzzword. It's a tool that needs careful, strategic application. Are crypto companies applying AI where it matters, or are they dazzled by tech for tech's sake?
As AI continues to evolve, the challenge for the crypto industry will be to follow the IRS's lead. Identify the right problems, measure impact, and don't rush to custom solutions. And just like that, AI could become the ally that turns tech hype into reliable profits.
