AI’s agent’s accreditation: maximizing gains while moving in challenges

While Agency AI definitely represents a turning point in the interaction of human computers, and the transition from the use of the tool to cooperation, the next step is to actually integrate these factors and derive value. In Venturebeat Transfer 2025Matthew Krob, the administrative director and a senior partner in BCG, presented a game plan for the development of workflow, employee accreditation and organizational change.

“The companies in the upper part of this curve – what we call in the future, and the most mature companies – you see large results: 1.5 times the growth of revenues, and it was 1.8 times value for the shareholders.” “There is value here, but we are early.”

https://www.youtube.com/watch?

Publishing, reshape, invented

To benefit and create a value with artificial intelligence and with agents, the company needs to determine the place of focus, using a publication frame, reshaping, and invent. Artificial intelligence is already published in each institution, and they will have agents during the next few years. But if you are given a Chatbot employee, it will not change the way the work is done. You must rethink the work, reshape the functions of the departments and the workflow by determining the place of human work.

“We advise companies currently to focus on your three or four large rocks. If you have a large institution to support customers, you must apply artificial intelligence in customer support. It has a great impact. If you have a great engineering institution, you must use tools like Windsurf to reshape the method you do in engineering, and develop software.”

The invention is still very early, but institutions should consider how to use the ability of artificial intelligence to be creative, reason and planning. Look at services and products, and how do you interact with customers: Can you re -invent those capabilities? “

For example, the L’Oreal makeup company launched a virtual beauty consultant to expand the scope of exclusive service outside their retail sites, re -invented the way they think about interacting with its customers on a large scale.

Thinking behind basic use

It is important to think about how to change the artificial intelligence of your work. There has been a lot of focus in the past two years on reducing costs by replacing workers, but this is not a big thinking in the picture. AI currently pumps the employees you have, which greatly increases their productivity.

“This is what we see in developing software,” he said. ))

In a study conducted by BCG with Harvard, Wharton and MIT, they asked 750 knowledge factor to write a commercial and marketing plan, with and without the Toulidi AI. Participants who use GPT4 performed 25 % faster, completed 15 % of the tasks, and their production quality was 40 % better. When LLM was given, the lower performers in the foundation line did and the best performance.

He said: “Everyone’s performance, which is very strong, has raised it, because in most organizations, the new carpenters are less effective than the most experienced people.” “It has the ability to increase time to efficiency.”

Artificial intelligence can go beyond human size, even opening new applications that were not possible before. For example, in the medical space, the results of patients are greatly improved through follow-up before surgery and post-surgery from a nurse, but the implementation of this was expensive-until the appearance of artificial intelligence nurses that could take this task for large patients for the patient.

Overcoming: adoption

While these tools are great, people do not use them. BCG tracked the adoption of GitHub Copilot scales and the production standards of the institution that contains about 10,000 software engineers. The best 5 % of the productivity of productivity doubled in four months, while 60 % showed a zero improvement, because they did not adopt the tool at all.

Why do not people adopt? There are three reasons. The first is the ability to ignorance. Second, stagnation usually. The third is the threat of identity, and this is the most difficult to overcome. The developers ask, “If this artificial intelligence can write a symbol for me, then who am I? What is my value?”

“This will be the real work during the three years to the next five,” said Krup. “It makes people use agents.”

Strategies to overcome frequency

There are some valuable ways to overcome these challenges. Of course, getting the correct tool is the first step, and merged it in the way that people work explicitly. It is important to measure and celebrate adopting for these employees who actively use tools so that everyone begins to see that they need to obtain this vehicle.

Another important step is to increase the scarcity – this means that the resources take away until the employees need to do more at the lowest. At the same time, it is necessary to redesign work operations along with these employees on the front lines. Not only does the hard operations as manual work can be a conference – select the parts in which humans bring value.

“We reduce toil and increase joy.” “We have left a more efficient process, a more efficient company, a more productive labor force, and the functions in which people like to be.”