The success of artificial intelligence implementation in companies may depend on employee involvement in training specialized models, according to Financial Times columnist Sarah O’Connor in a report.
This involves tacit knowledge—practical skills and decision-making processes developed through experience but rarely documented in manuals and corporate records. FT linked this concept to philosopher Michael Polanyi, who articulated the principle: “we know more than we can tell.”
The author suggests that general AI models perform better in areas with large volumes of open data and clear verification criteria. In tasks where a company’s internal context and expert judgment are crucial, written instructions alone may not suffice.
The publication compared this to the work of American engineer Frederick Winslow Taylor, who in the early 20th century attempted to systematize factory workers’ experience through observation, recording, and time measurement.
Bridgewater Enhances Model with Expert Evaluations
One example is a joint study by Bridgewater AIA Labs and Thinking Machines Lab, published on June 30. The teams tested AI on six tasks from the investment company’s workflows: models analyzed financial news, central bank documents, and other materials potentially significant for decision-making.
Using simple instructions, Claude Opus 4.6 and 4.8, Gemini 3.1 Pro, GPT-5.4, and GPT-5.5 showed accuracy from 45.6% to 50.1%. Prompts prepared by Bridgewater experts increased accuracy to 74.3–78.2%. Automatic optimization of instructions did not provide additional improvement.
“An explicit prompt can only convey the part of intuition that an expert can express in words. The most important judgments are often the hardest to articulate,” the authors noted.
Researchers then further trained Qwen3-235B on data labeled by Bridgewater specialists. The system’s average accuracy reached 84.7%. According to the developers, the number of errors decreased by 29.8% compared to the best-tested general model.
The cost of task processing was 13.8 times lower. The authors called the approach “differentiated intelligence,” which involves tailoring individual models to the processes and requirements of a specific organization.
Ford Expands Team of Experienced Specialists
FT also cited Ford as an example. Over the past three years, the automaker has hired, promoted, or rehired about 350 experienced technical specialists. They review designs, identify potential defects, train young engineers, and participate in improving automated quality control tools.
Ford’s Vice President of Hardware Development Charles Poon stated that the effectiveness of AI depends on the quality of the information used for training.
“Artificial intelligence is a great tool, but it is only as good as the information used to train it,” he noted.
According to Poon, the company has not systematically preserved the knowledge of experienced engineers. Some specialists left Ford before their expertise could be fully integrated into internal processes.
In October 2025, ForkLog explored how developers adapt large language models for office tasks.
