JBS Dev Executive Says Companies Don’t Need Perfect Data to Start Using AI
JBS Dev Executive Says Companies Don’t Need Perfect Data to Start Using AI
Joe Rose, president of technology provider JBS Dev, says many companies are delaying AI adoption because they mistakenly believe their data must be perfectly organized before they can use generative or agentic AI systems effectively.
Speaking ahead of the AI & Big Data Expo, Rose argued that modern AI tools are already capable of handling messy, incomplete, and poorly structured information far better than many businesses realize. He said vendors often push organizations toward massive data lake projects and expensive multi-year transformation programs, leaving executives overwhelmed before AI projects even begin.
According to Rose, today’s large language models can already understand poorly written prompts and work through imperfect datasets, making it possible for companies to start experimenting with AI much sooner than expected. However, he emphasized that human oversight remains essential because AI systems can still produce unreliable or inaccurate outputs.
Rose explained that businesses should approach AI adoption gradually, improving automation over time instead of expecting immediate perfection. He described a healthcare client that used generative AI to migrate data into a new billing reconciliation system despite dealing with inconsistent records, PDFs, scanned images, and misplaced patient information. AI tools were able to extract and organize the information, while more advanced agentic systems later compared billing records against insurance contracts to verify pricing accuracy.
Rather than aiming for full automation immediately, Rose said companies should focus on progressively increasing efficiency—from 20 percent automation to 40 percent, then eventually 60 or 80 percent as systems improve and employees gain confidence in the technology.
Looking ahead, Rose believes future discussions around AI will focus less on massive leaps in model capability and more on sustainability, portability, and infrastructure costs. He argued that the industry cannot continue building data centers at its current pace indefinitely and expects more attention to shift toward running advanced AI systems locally on laptops, phones, and smaller devices instead of relying entirely on cloud infrastructure.
Rose also suggested that many organizations already possess the tools needed to begin implementing AI projects through existing cloud services without purchasing expensive new software platforms. He encouraged businesses to rely more on the built-in AI capabilities offered by major cloud providers rather than immediately turning to third-party SaaS vendors.
According to Rose, the growing maturity of cloud-based AI tooling means businesses can start deploying agentic AI workloads much faster and with fewer barriers than many executives currently believe.
