It wasn’t long ago that a consultant at McKinsey & Company looked closely at the report to make sure it fit the company’s writing style.
Now, an AI agent called “sound tone” does it.
The Boston Consulting Group uses a tool called Deckaster to reduce the time spent grinding PowerPoint slides. In Ernst & Young, instead of contacting a payroll, consultants can ask the chatbot to explain the payroll slip.
Consulting companies are early leaders in the generative AI trend. They help other companies train their employees, develop new tools and regulate their technologies.
We’ve also internally tested our generator AI, and have announced a new suite of chatbots, agents and applications that have quickly and quietly changed the way consultants work in the past two years.
McKinsey has embraced the genai conversational approach
At McKinsey, consultants use an in-house generated AI chatbot called Lilli. The company told BI.
The user enters the request into Lilli. This aggregates key points, identifies five to seven related internal content pieces, and points users to the right experts within the company. Users can choose to answer queries via the company’s internal knowledge repository or external sources.
Lilli’s use in the company has exploded since its first rollout in 2023. Over 70% of their 45,000 employees currently use the tool. Those who use it rely on it about 17 times a week, McKinsey’s senior partner, Delphine Zurkiya, told BI.
When McKinsey first launched Lily, employees experienced uncertainty about what the company calls “quick anxiety” or what to ask the bot. However, we found that just an hour of training improves employee engagement. Zurkiya said the tool has also evolved since its launch. Initially, most of the company’s knowledge was not designed to analyze the PowerPoints, where it existed.
A consultant at McKinsey told BI he is using it for research and is being used for document summary, data analysis and brainstorming. In a case study published on its website, the company reported that workers used Lilli to save 30% of their time.
Zulkiya, who describes herself as “one of Lily’s heavy users,” said she often uses it with her team to identify the right approach to solving client problems. “We have almost all AI in our rooms, because we often say it.
McKinsey’s partner told BI that the company has been developing AI products for many years. In 2015, it acquired data analytics and design company Quantumblack. This now serves as McKinsey’s AI consulting arm. It employs 7,000 technical experts in 50 countries.
“Around 40% of the work we do is analytics-related, AI-related, and many of them are moving to Gen AI,” senior partner Ben Ellencweig told BI last year. McKinsey has built client-generated AI solutions through an “ecosystem” of alliances with 19 AI companies, including Microsoft, Google, Anthropic and Nvidia, and has completed over 400 Genai projects for clients.
However, ChatGpt’s popularity has crystallized the value of conversation tools, Zurkiya said. “There hasn’t been a major change in strategy in the sense that we already have a lot of tools developed internally. These tools have made them faster because of their natural user interfaces,” she said.
McKinsey Consultants are unable to access ChatGpt.
Lilli is just one of several AI tools that change the work within the enterprise. Zurkiya said that AI technology is being deployed at three levels. At the individual level, the platform allows consultants to build their own AI agents. This is a technology that can solve problems, among other things, and perform tasks autonomously. Next, there are more domain-specific tools. The Life Sciences agents where Zurkiya works can help consultants speed up certain companies in the sector. There are also solid tools, such as new tools for meeting and travel bookings.
The company also applies lessons from building Lily to new client projects, developing similar tools to suit their needs.
Despite the hype of Genai tools, consultants don’t seem to be worried that their work will be threatened as a result. A commentator for Fishbowl, an anonymous professional networking app that works for McKinsey, described the tool as “functional enough” and perfect for “very low stakes issues.”
BCG wants AI to “increase the joy” of work
For the past two years, BCG has been pushing to train employees with AI.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
In 2023, the company announced ChatGPT Enterprise to all employees under the provision that all data is under its control. Since then, the company’s 33,000 employees have built over 18,000 custom GPTs (tailored versions of ChatGPT) from document summaries to generating automated email responses from answering HR-specific questions.
BCG has also developed eight or nine internally generated AI tools, developed by partner and managing director Scott Wilder.
Wilder said Deckaster, a slideshow editor, is one of the tools that invest heavily. It is trained with 800-900 slide templates to help consultants create presentations quickly. Wilder said one of Dedkster’s most popular features is the “This Review This” button. This helps junior consultants by grading their slides based on best practices used by mid-level managers and leaders. About 40% of associates use Deckaster every week, Wilder said.
The tool is becoming so popular that some consultants are worried about job safety. “The BCG people who tried Decker: How much should we worry about our work? Does it already create groundbreaking productivity where more junior people are not needed so much?” One consultant wrote about Fishbowl last year.
One of the more experimental tools announced by BCG is Gene, a conversation chatbot. The bot is built on top of the 11 of the 11 on top of the GPT-4O and has a purposefully robotic voice.
“It’s a deliberate choice, a subtle reminder that I’m not a human, I’m AI. I’m suppressing my expectations,” Gene said of the voice during a December 2023 BCG podcast.
Gene also explained that its knowledge base is “built from an extensive collection of BCG’s best ideas about Genai, shaped by conversations with industry experts, articles and research research.”
The bot is designed to be a “conversation partner,” the company said. Consultants have adopted it for brainstorming, hosting podcasts, and live demonstrations, and are even considering using it to interview partners to create company content. Teams can change the bot’s “temperature” to control the tone of the response.
BCG also has an internal platform for building AI agents in beta testing.
In a pessimistic story of layoffs and robots coming for work, Wilder said the company’s paper on AI is optimistic. “I think our goal is to work hard and increase joy,” he said. He added that the company estimates that it will reinvest about 70% of the time it saves employees in “higher value activities.”
However, these time savings mean that consultant expectations are fluid. BCG relies heavily on Genai tools and has not changed the way it evaluates performance. However, a company spokesman told BI: “We are thinking about the role they play as these technologies are central to our working style.”
Deloitte, KPMG and PWC see AI Agent Promise
Generator AI appears to be more strictly regulated at Deloitte. For one, ChatGpt is blocked from the company’s internal systems, three consultants told BI.
Related Stories
Business Insider tells the innovative stories you want to know
Business Insider tells the innovative stories you want to know
“I think maybe the only people they’re really trying to avoid here are analysts or just people who put things like client data in generic AI tools,” Andrew Sutton, the company’s senior advisory consultant, told BI. Sutton, which builds internal AI tools for other consultants in the company, said they need to develop them in a safe environment to prevent data leaks.
“If you’re using tools that come from something like Openai, there’s special communication and contracts,” he said. “The amount of bureaucracy we have to go through, everything like that.”
The company has its own ChatGpt alternative called Sidekick. This comes with a disclaimer that employees can only use for non-client work. Deloitte consultants use BI to summarize documents, brainstorm, edit emails, and code.
But Deloitte has invested billions in artificial intelligence. In March, we announced Zora AI, a new fleet of AI agents. The company says they are trained in specific subjects like finance and marketing and designed to think like a human. Last year, we expanded our digital distribution platform Ascend, which has a generation AI function.
Publicly, the company’s leadership is also centered around technology. At NVIDIA’s GTC conference in March, Gillian Wanner, Principal of Deloitte, who leads the company’s AI staff development, acknowledged that the consulting industry is “disturbing” amid the transformation of AI. Jim Rowan, head of AI at Deloitte, previously told BI that senior managers should use AI to demonstrate its effectiveness and give employees time to explore technology.
In a recent statement to BI, Rowan said, “We believe that AI is transforming all industries that will help us transform our own industry, guide new business models and ways of working, and uncover new sources of business growth and innovation.”
KPMG is taking two approaches to AI adoption, according to Ecosystem Head Todd Lohr. “I’m a huge fan of Top Down and Bottoms,” Loa told BI. “It’s really hard to know what hundreds of thousands of people in an organization do every day. But by giving them technology and getting them to use it, they’ve come up with creative ways that are even better than any top-down methodology.”
He said that when the company began rolling out technology two years ago, people were a bit confused about how to use Genai.
“I call it swivel chair processing, which is really difficult for people who have been working for decades to stop what they’re doing for a while,” Lohr said. Since then, the company has harvested data on how its employees are encouraging AI. It uses that information to build new tools for itself and for clients, through searched generation, through methods and open data sources that increase the idiosyncraticity and accuracy of large-scale language models, Lohr said.
As consulting companies develop more sophisticated tools, like platforms for agents, they realised they needed a hub to centralize them. KPMG signed a deal with Google Cloud this month to buy a license for the US workforce for a new platform that integrates AI agents with corporate data.
Deloitte recently announced Agent2Agent, a new platform that improves interoperability between agents. This is the company’s biggest collaboration with Google Cloud and ServiceNow.
PWC announced a similar platform last month called Agent OS. It helps to focus the client’s agents. It also uses over 250 internal agents that we’ve built over the last 18 months. Matt Wood, PWC’s global and US commercial technology and innovation officer, told Bi, Matt Wood is to convert isolated agents from “night-passing ships” to “fleets working together.”
After post-pandemic dry spells where many consulting companies struggled with layoffs, contract losses and cost-cutting initiatives, the generated AI is like a light at the end of the tunnel.
“My bet is that as more agents become available, organizations will see growth as well as efficiency,” Wood said. “And with that growth, they can double what’s working, making them a bigger organization rather than a smaller organization.”
