A Journey to the Starting Line

Magicpill Labs is a project focused on helping SMBs build successful AI Solutions faster and with higher quality and ROI. After ChatGPT launched in late 2022, I started thinking about how AI could be used for synthetic data generation for demos, side projects, a way to get around PII issues with testing, etc. From there I also started building my own tools for evaluating new models that were coming out as a better way to analyze different models’ performance. That resulted in the groundwork for Magicpill Labs’ eventual first app: SparkEval.

After all these projects and following the LLM model topic for a while, around 2024 I started spending time with individuals who were interested in using AI to improve their daily workflows. During this time, I noticed a clear divide forming between two groups: those leveraging AI effectively and those struggling to get projects “across the last mile” or who were disregarding LLM’s value potential entriely. What I found was helping people grow with AI came down to a few elements:

• Isolated Use Cases that Provided Value. Teams that were getting value from AI were starting with specific problems and creatively exploring how AI was able to accomplish that goal. This is a slower rollout strategy than just buying lots of ai licenses and letting it loose everywhere, but does build momentum with through every success.
• Developers that understood the buisness use case. AI is making software a commodity, but a “solution” is more then just code that meets a requirement list. Deeply understanding the business and users for projects was a key factor in addition to software engineering skills to delivering projects that actually worked and provided value.
• Accessible tools and platforms for SMBs. Theres a growing list of products and platforms for building with AI, but many tools are focused on developers. While important, we are entering a phase where more non-technical users are becoming developers and building tools, automations, apps and more on their own. This raises the need for more Agent Operations type products that are focused on making it easier for non-developers to build and maintain AI solutions, and for developers to collaborate with non-developers on those projects.

You can’t just throw technology at the problem. You need clear use cases, capable teams who can execute, and platforms that make complex development tasks manageable. With AI assisting more in development, code is easier then ever to create, and thats putting more emphasis on the design, business analysis, adoption and human factors of software solutions; and thats when Magicpill Labs was born.

AI is accelerating. The skills gap is widening.

Now is the perfect time to focus more on leveraging AI for better solutions. Every few months the capabilities take another leap. Longer context, better reasoning, more reliable tool use, emerging multi-agent frameworks. Andrej Karpathy described 2025 as the year reinforcement learning from verifiable rewards fundamentally changed what LLMs can do, and coding agents like Claude Code emerged as the first convincing demonstrations of real AI agency. What required a team of engineers six months ago can now be prototyped by one person in a weekend.

That sounds like it should be great news for everyone. But in practice, the benefits are concentrating. The people and companies who are exploring how to leverage these tools are pulling ahead, while those who don’t are falling further behind. We’re watching an AI skills gap form in real time.

Don’t take our word for it though. According to IDC’s 2025 AI Workforce Readiness Report, over 90% of global enterprises will face critical skills shortages by 2026, with sustained gaps risking $5.5 trillion in losses from global market performance. 94% of CEOs identify AI as their top in-demand skill, yet only 35% feel they’ve actually prepared their workforce. PwC’s 2025 AI Jobs Barometer found that AI-exposed roles are evolving 66% faster than others and command a 56% wage premium. The World Economic Forum reports that 94% of leaders face AI-critical skill shortages today, with one in three reporting gaps of 40% or more.

We’ve heard stories (and you probably have too) of organizations that rolled out AI initiatives across thousands of users, spent significant budgets, and watched the whole thing flop. Not because the technology didn’t work, but because nobody started with a clear, valuable use case. Nobody asked “what problem are we actually solving?” before jumping to “let’s deploy AI everywhere.”

The research confirms this pattern. MIT’s GenAI Divide report found that only 5% of enterprise AI pilots achieve measurable P&L impact, despite $30-40 billion in enterprise investment. S&P Global reported that 42% of companies abandoned most of their AI initiatives in 2025, up from just 17% in 2024. The common thread isn’t that AI doesn’t work. It’s that organizations are deploying it without clear strategies, workflow redesign, or skills to bridge the gap between a demo and a production system. McKinsey’s 2025 AI survey confirmed that organizations reporting significant financial returns are twice as likely to have redesigned end-to-end workflows before selecting modeling techniques.

Every one of those failed rollouts makes the macroeconomic problem potentially worse. The organization writes off AI as overhyped. The team that championed it loses credibility. And the gap between the companies that are successfully leveraging AI and the ones that tried and gave up gets wider.

That’s good for the companies successfully leveraging AI, but it’s bad for the economy and society as a whole as this phenomenon scales.

Why the skills gap matters to us

This is where our mission comes in. We believe the healthiest future is one where more people and more businesses can leverage AI effectively, not fewer. More competition means a more balanced economy. More people with AI skills means the benefits get distributed more broadly rather than pooling at the top. The IMF’s recent research on AI skills and labor markets reinforces this: while AI skills are boosting wages and employment for those who have them, the diffusion is uneven, and the polarization is contributing to a shrinking middle class.

That’s why we focus on small and mid-sized businesses. These are the companies that don’t have a dedicated AI team or a seven-figure R&D budget, but they have real problems that AI agents can solve today. Salesforce’s SMB Trends Report found that 91% of SMBs already using AI say it boosts their revenue, and growing businesses are nearly twice as likely to be investing in AI as those that are struggling. But there’s a meaningful adoption gap: while 58% of SMBs are now experimenting with generative AI, only a fraction have fully integrated it into their operations. The biggest barriers are the same ones that plague larger organizations: skills gaps and lack of in-house expertise, cited by 50% of SMBs as their top challenge.

These companies just need the right partner to help them identify the right use cases and build solutions that actually work.

Looking forward

As AI continues to accelerate, we think the most valuable skills won’t be writing code. Building software is quickly becoming a commodity. What will matter more is the ability to identify the right problems, define them clearly, and design solutions that create real value. Ideas, critical thinking, domain knowledge, and architecture will be the differentiators. Deloitte’s 2026 State of AI report found that only 34% of organizations are truly reimagining their business with AI, and that insufficient worker skills remain the biggest barrier to integration. The opportunity isn’t just in deploying AI. It’s in thinking clearly about where and how it should be deployed.

That’s where we’re headed. MagicPill Labs is here to build AI & software solutions for customers. But the long term goal is to create tools for builders that help them scope, design, build and manage their AI solutions so they can focus on their magical ideas they want to bring to life. The real work is just beginning, and we’re excited about what comes next.

---

If you’re thinking about building an AI agent for your business, or if you’ve tried before and want a better outcome, we’d love to talk. Reach out to us and let’s figure out the right approach together.