FEATUREDApplication Development & Modernization
KEITH MANTHEY | Field CTO
The main takeaway from MIT’s latest research is clear: it is not just about the AI strategy but also about execution and democratization. While companies continue pouring resources into generative AI initiatives, 95% of enterprise pilots are delivering little to no measurable impact on P&L, with only 5% achieving rapid revenue acceleration. This stark divide reveals two fundamental truths that validate everything BlueAlly has been saying: flawless execution requires the right partnerships, and real AI impact comes from democratizing access to individuals who understand the problems.
Execution: Why Building Beats Buying in Unfamiliar Territory
The data reveals a critical execution insight that confirms my philosophy: companies building their own proprietary AI tools succeed only one-third as often as those purchasing from specialized vendors and building partnerships, which succeed 67% of the time. However, the key point is that this isn’t about altogether avoiding builds. It’s about understanding that in an unfamiliar AI landscape, you need someone knowledgeable about the terrain.
MIT’s research shows that enterprises often attempt to build their own tools, particularly in regulated sectors, and fail because they’re navigating unknown territory without guides. The challenge of execution lies not in the quality of AI models, but rather in executives often blaming regulation or model performance. However, research from MIT indicates that poor enterprise integration is the real issue.
The “learning gap” identified by MIT researcher Aditya Challapally exposes why going it alone fails. Generic tools like ChatGPT excel for individuals due to their flexibility, but they struggle in enterprise use because they fail to learn from or adapt to workflows. The BlueAlly approach solves this: find a partner (like BlueAlly) who is familiar with the technology landscape, who understands the delivery mechanisms, tools, and integration challenges that trip up internal teams.
When companies partner with experts in AI implementation, they gain not only technology but also the execution skills necessary to navigate complex delivery, select suitable tools, and address integration challenges that might hinder internal projects. The successful 5% aren’t avoiding builds; they’re building smart by partnering smartly with companies that use their tools and understand the execution landscape.
Democratization: Moving AI Out of Centralized Towers
The MIT data reveals a critical democratization failure that BlueAlly has been highlighting: centralized AI initiatives aren’t delivering returns because they’re disconnected from real problems. Over 50% of generative AI budgets are allocated to sales and marketing tools. However, MIT discovered that the highest return on investment (ROI) comes from back-office automation, which involves eliminating business process outsourcing, reducing external agency expenses, and optimizing operations.
This misallocation happens because centralized AI towers chase high-visibility, customer-facing use cases instead of focusing on where individuals in the trenches need automation and efficiency solutions. Sales and marketing POCs look impressive in board presentations, but they’re not where the real productivity gains live.
The research supports our belief that democratization is essential. It is crucial to empower line managers—not just central AI labs—to lead the adoption of AI. Those who work closest to operational challenges are better equipped to identify where AI can effectively address real problems. They see practical opportunities for automation that reduce manual tasks and enhance daily processes, rather than just the theoretical efficiency gains that centralized teams often discuss.
The Power of Individual Problem-Solving
Workforce disruption is already taking place, particularly in customer support and administrative roles, as companies increasingly choose not to fill vacancies when they arise. Notably, this trend is not being driven by centralized AI mandates; rather, it is a result of individuals discovering ways to automate their own tasks.
The widespread use of “shadow AI”—unsanctioned tools like ChatGPT—supports this observation about democratization. People in the trenches are already using AI to solve automation and efficiency problems, often without official approval. Rather than fighting this trend, smart organizations should embrace it by democratizing access to better tools and supporting bottom-up innovation.
A Two-Pronged Approach for 2025
Looking ahead, the most advanced organizations are already experimenting with agentic AI systems that can learn, remember, and act independently within set boundaries. This next phase requires both execution excellence and a strategy for democratization.
- On Execution: Don’t build in unfamiliar territory without expert guidance. Partner with those who understand AI delivery, tools, and integration challenges. The 67% success rate of partnerships highlights the importance of building with expertise.
- On Democratization: Stop pouring budgets into centralized AI towers focused on sales and marketing showcases. Instead, democratize AI access to the individuals who understand operational problems. The best back-office automation opportunities delivering real ROI are found by those doing the actual work, not in boardrooms.
The question isn’t whether to embrace AI, but rather how to execute it through expert partnerships while democratizing problem-solving for those who understand the real challenges. The 95% who fail stick with centralized approaches and unfamiliar builds. The 5% who succeed combine execution expertise with democratized innovation.
I’ve seen too many organizations learn this lesson the hard way. The path forward requires both expert execution and grassroots democratization. Don’t let your organization join the 95% who miss both opportunities.
