Summary
Many business leaders believe they cannot use artificial intelligence until their company data is perfectly organized. Joe Rose, the president of JBS Dev, recently explained that this idea is a common mistake. Modern AI tools are now smart enough to handle messy or incomplete information effectively. By using these tools correctly and keeping humans involved in the process, companies can start using AI immediately rather than waiting years for data cleanup projects to finish.
Main Impact
The biggest impact of this shift is that it lowers the barrier for companies to adopt new technology. In the past, businesses were told they needed massive "data lakes" and expensive transformation programs before they could try AI. Now, the focus is shifting toward using AI to fix the data as it works. This allows businesses to save money and see results much faster. Instead of spending years preparing, they can begin automating tasks right away, which helps them stay competitive in a fast-moving market.
Key Details
What Happened
Joe Rose shared insights on how generative AI and "agentic" systems—AI that can take action on its own—are changing the workplace. He pointed out that Large Language Models (LLMs) are surprisingly good at understanding what a person wants, even if the instructions or the data provided are not perfect. He argued that the tools available today are better than ever at dealing with low-quality information. However, he also stressed that these systems are not "set it and forget it." They require constant attention and human oversight to ensure the results stay accurate.
Important Numbers and Facts
During his discussion, Rose used a medical sector example to show how this works in the real world. A client needed to move records to a new billing system. The data was a mess: some files were images, some were PDFs, and names were often in the wrong boxes. The AI was able to extract the correct text and organize it. Rose suggested a step-by-step approach to automation. A company might start by automating 20% of a task, then move to 40%, and eventually reach 80% or more as the system learns and the human workers provide feedback.
Background and Context
For a long time, the rule in computing was "garbage in, garbage out." This meant that if you put bad data into a computer, you would get a bad result. Because of this, companies spent millions of dollars trying to make their data perfect. They hired consultants and bought complex software to clean up their records. While clean data is still helpful, modern AI works differently. It can "read" and "reason" through errors much like a human does. This change means that the "garbage in" rule is no longer as strict as it used to be, allowing for more flexibility in how technology is used.
Public or Industry Reaction
The tech industry is currently divided on this issue. Many large software vendors and consulting firms still push for big, multi-year data projects because those projects are very profitable for them. However, experts like Rose are encouraging businesses to be more independent. He suggests that companies should stop buying expensive new software licenses from outside vendors. Instead, he advises them to use the tools already available in their existing cloud accounts. Most businesses already use major cloud providers that offer AI tools as part of their standard service.
What This Means Going Forward
In the future, the focus of AI will likely move away from making models bigger and "smarter." Instead, the goal will be making them cheaper and more efficient. Rose calls this the "last mile" of AI. This involves getting AI to run on smaller devices like laptops or smartphones instead of requiring massive, power-hungry data centers. As the cost of running these models becomes a bigger concern, companies will look for ways to make the technology sustainable. We may see fewer "breakthroughs" in what AI can do and more improvements in how easily and cheaply we can use it every day.
Final Take
The era of waiting for perfect data is over. Businesses that wait for every record to be in the right place will fall behind those that start using AI now to sort through the mess. By combining the power of modern cloud tools with human intelligence, companies can build systems that grow more accurate over time. The key is to start small, use the tools you already have, and focus on making the technology work for your specific needs without overspending on unnecessary software.
Frequently Asked Questions
Do I need to clean all my company data before using AI?
No. Modern AI models are very good at understanding messy or imperfect data. You can start using AI to help organize and clean your data while it performs other tasks.
What is a "human in the loop"?
This means that a person stays involved in the AI process. The human checks the AI's work, fixes mistakes, and helps the system learn to be more accurate over time.
Can AI run on a regular laptop?
Currently, most powerful AI runs in large data centers. However, the industry is working on the "last mile" to make these models small and efficient enough to run directly on laptops and phones in the near future.
