IA leaders in the best advisory companies on how they use AI, the best guests

• Consulting companies have become a destination for certain companies that seek to make the most of the AI.
• AI managers of these companies use tools like GPT Enterprise and internal chatbots like McKinsey’s Lilli.
• Bi has asked AI managers of several large companies to share advice for the use of technology.

Working with artificial intelligence can sometimes look more like an art than a science.

This is why many companies turn to consulting companies to obtain advice on how to maximize technology.

Large companies help not only companies to develop AI tools, increase their workforce and identify potential safety weaknesses, but they also create chatbots and agents to organize their business knowledge and rationalize routine tasks. Consequently, AI managers in consulting companies tend to master AI strategies that can work for a wide range of tasks.

Business Insider asked AI managers in five best advisory companies – Deloitte, EY, KPMG, McKinsey and PwC – to share their best advice to use AI in daily work.

AI managers said they regularly use various AI tools, including OpenAi, Google, Microsoft and Anthropic models, as well as internal built tools, such as McKinsey’s Lilli, Ey’s Eyq and Chatpwc, The internal version of Chatgpt PWC.

Here’s how they use AI and some of their advice to get the most out of it. The responses are modified by conciseness.

How do you use AI in your work?

Dan Priest, head of the American AI at Pwc: I do a lot of research with. For example, I was doing an analysis on labor productivity and how AI will improve labor productivity. Typical research would produce labor statistics. Well, AI, the major powerful foundation models, he will seize these facts and working statistics, he will make analyzes, this will show you trends, discontinuities or will cause analyzes. It is much more robust. In terms of research and analysis, it appears as a partner of thought against a single search engine.

You discover the dead angles in your reflection. I wrote a policy and I thought it was quite complete, and I ran it in GPT Enterprise, and he found two other points in the policy that we should add.

Todd Lohr, chief of ecosystems in KPMG: Part of my work as a leader is to be able to synthesize information. The AI ​​is very useful in that this allowed me to understand the trends and the market and allowed me to have a broader vision as a leader and to synthesize and to ingest much more information.

It was also useful for communications in terms of preparing for meetings, monitoring of meetings, as well as for correspondence.

Rodney Zemmel, world leader of McKinsey Digital and Firmway throughout AI: I found that it was excellent to creativity “level one” and to find things that you will not generally have thought of. It is an excellent help to think about our teams. I have not yet seen it as having real unlimited creativity, that is to say a new way of looking at the world. It will not be far behind, however.

Related stories

What are your invites or advice of sharp intestance to write good prompts?

Dan Priest, Pwc: I will give a context on what I try to do, a short and impactful question, then ask follow -up that makes them more and more specific, then you can adapt according to what you see.

During the week, I travel a lot, and if I get 100 or 200 emails per day, it is really difficult to follow each of them. I go to Microsoft teams, activate the co -pilot and I ask him to review all the messages in teams and email and find the actions for me. I’m just going to spend 15, 20 minutes at the end of the day, invite “Identify the emails that are sent to me directly or who have an action for me”, and he produces the list. It’s not perfect, but it’s good in this area.

I like to cook and I don’t like to waste food in the refrigerator. So I’m going to invite, “Create a recipe with these ingredients”, and I will simply list the things I want to get rid of the refrigerator.

Rodney Zemmel, McKinsey: Too many people always use it to search for something. The trick is to have a dialogue with it and make construction agents who can perform simple tasks. Let AI manage the 80% of the tasks in which we are mediocre, so that we can excel at 20% exciting, as one of my colleagues likes to say it.

Matt Barrington, Director of Americas Technology at EY: Context management is essential. I keep “workspaces” separated for different areas of intervention – such as technical questions and answers or the writing of customer communications.

I also give the clear instructions of AI on the style and depth of the answer I want, such as “providing a concise summary and bubbles” or “act as a financial expert” or “quote sources or credible references and provide links ”.

What challenges is there with the use of AI?

Dan Priest, Pwc: It changes muscle memory.

I have spent a lot of years developing a certain writing style, a certain research technique, and I had to change this. And I am better for changing it, but that does not happen overnight.

It was like everything you learn, you must be disciplined to learn it, then he sticks.

Todd Lohr, KPMG: The biggest challenge is to connect all my disparate individual data sources. If I want to build my own personal AI, the challenge is to have access to good information and knowledge.

I was deliberate to take up this challenge when I took over in my current role. I put everything in a file and I personally organized the content with which I agree and that I liked.

Matt Barrington, Ey: The main challenge is to follow the pace of innovation. There is a constant flow of new models, tools and capacities, and it can be difficult to determine the option that best suits a given task. I am the newsletters, I participate in AI -oriented events and I learn AI practitioners – but in my opinion, practical experimentation is the most effective way to remain informed and find what works Really.

In addition, it is important to remember that if these models are confident and impressive, they can be false. Always validate the information and the output before using it.

What do you want to know your customers on the integration of AI into their business?

Dan Priest, Pwc: The questions have somehow changed. A year ago, they asked: “What is the case of use of the killer?” “What is the most industrialized use?” “What is the case of use that will produce the most savings or the biggest gaps?” Now, the questions we obtain relate to these technical use cases less and they are much more on “How do you evolve the commercial strategy to take advantage of the AI ​​capabilities?”

Rodney Zemmel, McKinsey: They want to understand how AI agents can integrate into their workforce, acting as talented trainees who need appropriate training to be effective. We have also seen the conversation go from simple productivity to growth and productivity, and to find ways to do things better and faster than humans to do things that no human could do.

Do you have something to share on what you see in consultation? Business Insider would like to hear from you. Send an email to our advice team from a non -work apparatus deliverable@businessinsider.com With your story, or ask for one of our journalist’s signal numbers.