The CIO's dilemma with AI

No one doubts that AI's real use cases in shipping will have deliver significant ROI. However the road to getting there is filled with mines that CIOs will need to carefully navigate.

Innovation vs. Risk Management - CIOs are expected to lead AI innovation to enhance productivity, decision-making, and user experience. However adopting Gen AI as a stand-alone solution, will produce hallucinations, wrong answers, and if such outputs are delivered directly to the users as actionable advice, it could actually reduce the organizations KPIs.

Open AI Tools vs. Enterprise Security - Employees often experiment with tools like ChatGPT, GitHub Copilot, or others. While these tools boost efficiency, they bypass enterprise security protocols, creating vulnerabilities mainly with GDPR but also with commercial trade secrets.

Quick Wins vs. Long-Term Strategy - There’s pressure to show quick results—like automating tasks or building an AI assistant. But without a long-term data strategy, model governance, and an integration roadmap, these efforts may become siloed and costly to maintain.

The approach we see prevailing with our leading CIOs, is recognizing that the risk from the three points above as significant, and as in most cases, opting for a risk managed, enterprise approach with a long term strategy. While this approach may not win you headlines or panel seats (unless of course you sponsor the event…) it would likely to make sure you actually get things done, albeit in baby steps.

As far as our approach to the architecture of Yamba ai, our initial concept of running a combination of models in concern, to contain the shortcomings of the technology, was not wrong. Running leading tools in a closed and controlled environment within the JiBe ERP sphere along with some proprietary methodologies, will enable avoiding the risks of adverse results. The key to this approach is the single version, with a single scheme and central data libraries, used by all clients.

If you ask Perplexity or ChatGPT what percentage of global enterprises run on a cloud based single version ERP, the answers are 50-60%. Feels low doesn’t it ? In shipping that number is lower than 5% and its mostly JiBe. We find that number concerning, as the industry continues to pay the price of project companies lipstick’ed as SaaS.

From our very first version in 2013, we have kept with a single version, where data was organized and validated for the use of Big Data and in turn – AI. This has been the best practice in tech for nearly 20 years now, the fact that shipping is willing to accept anything less is an issue. Its elementary Watson ..