This story is available exclusively to Business Insider subscribers. Become an Insider and start reading now.

In recent years, the landscape of consulting has witnessed a remarkable transformation driven by the rise of generative artificial intelligence (AI). Once, consultants at renowned firms like McKinsey & Company dedicated hours to meticulously reviewing reports to ensure they conformed to the company's signature writing style. Today, however, an AI agent known as "Tone of Voice" has taken over this task, streamlining the process and allowing consultants to focus on higher-value activities.

At Boston Consulting Group (BCG), consultants are leveraging a cutting-edge tool named Deckster, designed to significantly reduce the time spent refining PowerPoint presentations. Meanwhile, at Ernst & Young (EY), consultants no longer need to reach out to payroll for clarification; they can simply engage with a chatbot that explains their pay slips in a matter of seconds.

These consulting firms are emerging as pioneers in the generative AI revolution. They are not only utilizing AI within their organizations but are also at the forefront of helping other companies train their employees, enhance their technological capabilities, and navigate the evolving landscape of AI regulation.

Internally, these firms have made substantial investments in generative AI over the past two years, resulting in a suite of chatbots, agents, and applications that have fundamentally altered the way consultants conduct their work.

McKinseys AI Powerhouse: Lilli

At McKinsey, one of the flagship AI innovations is an in-house chatbot named Lilli. This sophisticated tool synthesizes the firm's extensive intellectual property, which spans a century and comprises over 100,000 documents and interviews. Users can input their requests into Lilli, which then aggregates key information and identifies five to seven relevant content pieces from the firm's vast repository while also directing users to appropriate experts within the organization. Users have the option to receive answers derived from the firms internal knowledge base or external sources.

Since its rollout in 2023, Lilli's adoption has surged, with over 70% of McKinseys 45,000 employees utilizing the tool. Regular users turn to Lilli approximately 17 times each week, according to Delphine Zurkiya, a senior partner at McKinsey. Initially, when Lilli was launched, some employees experienced what McKinsey termed as prompt anxiety, characterized by uncertainty over how to effectively engage with the chatbot. However, the firm discovered that providing just one hour of training significantly boosted employee engagement and proficiency with the tool. Lilli has since evolved to provide functionalities that were not part of its original design, such as parsing PowerPoint presentations, where much of the firm's knowledge resides.

Consultants at McKinsey have reported substantial benefits from using Lilli for various tasks, including research, document summarization, data analysis, and brainstorming sessions. In a case study shared on its website, McKinsey revealed that employees could save up to 30% of their time when utilizing Lillis capabilities.

Zurkiya, who considers herself a frequent user of Lilli, emphasized its collaborative potential: teams often consult Lilli to determine the most effective approach to client challenges. We almost have AI in the room with us because we often ask, what does Lilli think? she explained.

Partners at McKinsey have been at the forefront of developing AI products for years. Notably, the firm acquired data analytics and design company QuantumBlack in 2015, which now operates as McKinsey's AI consulting arm and boasts a workforce of 7,000 tech experts across 50 countries.

About 40% of the work we do is analytics-related, AI-related, and much of it is transitioning to generative AI, explained senior partner Ben Ellencweig. The firm has collaborated with a network of 19 AI companies, including industry giants like Microsoft, Google, Anthropic, and Nvidia, resulting in the successful completion of over 400 generative AI projects for various clients.

Interestingly, the meteoric rise in popularity of tools like ChatGPT has highlighted the intrinsic value of conversational interfaces. Zurkiya remarked, There wasnt a significant shift in our strategy since we had already been developing many tools internally. What has changed is that these tools can now deliver value more rapidly, thanks to a natural user interface. Notably, McKinsey consultants do not have access to ChatGPT.

BCGs Vision: Increasing Joy in Work

Meanwhile, Boston Consulting Group has taken proactive steps to train its workforce in AI technologies. In 2023, the firm rolled out ChatGPT Enterprise to all employees, ensuring that all data remained under its control. Since then, the firms 33,000 employees have developed over 18,000 custom GPTstailored versions of ChatGPTdesigned for a variety of internal applications ranging from summarizing documents to generating automated email responses and addressing HR-specific queries.

BCG has also created several internal generative AI tools. Among the most prominent is Deckster, a presentation editor trained on 800 to 900 slide templates, which allows consultants to rapidly create compelling presentations. Scott Wilder, a partner and managing director at BCG, stated that one of Decksters most popular features is a review this button that assesses slides against best practices employed by mid-level managers and leaders. Remarkably, about 40% of associates utilize Deckster on a weekly basis.

The tools growing popularity has sparked some concerns among consultants about job security. One consultant raised questions on the professional networking app Fishbowl, asking, BCG folks who've tried Deckster: how worried should we be about our jobs? Is it already creating groundbreaking productivity that more junior folks won't be needed as much?

BCG has also experimented with a conversational chatbot named GENE, built on the GPT-4o model by ElevenLabs, which features a deliberately robotic voice. This decision was made to remind users that they are interacting with an AI and not a human, and it adds a touch of retro charm to the interaction. GENE has been utilized for brainstorming sessions, podcast hosting, live demonstrations, and even discussions about potentially interviewing partners to generate content for the firm. Teams can adjust the bots temperature to modify the tone of its responses.

In addition to GENE, BCG is beta testing an internal platform for creating AI agents. Despite widespread narratives surrounding layoffs and automation, Wilder maintains an optimistic outlook on AIs role within the firm. We aim to eliminate tedious tasks and increase the joy in our work, he stated, adding that employees can reinvest approximately 70% of their time savings into higher-value activities.

However, as consultants adapt to these new tools, expectations are shifting. Although BCG has not revised its performance evaluation criteria in light of increased reliance on generative AI, a spokesperson for the firm noted that it is thoughtfully considering the role they play as these technologies become more integral to our workflows.

AI in the Consulting Landscape: Deloitte, KPMG, and PwC

Other major players in the consulting arena, including Deloitte, KPMG, and PwC, are also exploring the potential of AI agents. At Deloitte, the use of generative AI is regulated more stringently. For instance, ChatGPT is blocked from the firms internal systems to prevent analysts from inadvertently inputting sensitive client data into an unregulated AI tool. Andrew Sutton, a senior advisory consultant at Deloitte, explained that all internal AI tools must be developed in secure environments to ensure data protection.

Deloitte has introduced its own alternative to ChatGPT, called Sidekick. However, employees are permitted to use this tool only for non-client work. Consultants have reported using Sidekick for tasks such as document summarization, brainstorming, email editing, and coding. Deloitte has made substantial investments in AI technologies, unveiling Zora AIa suite of AI agents trained in specific subject areas like finance and marketingdesigned to emulate human thinking. Additionally, the firm expanded its digital delivery platform, Ascend, to incorporate generative AI capabilities.

Publicly, Deloittes leadership has embraced the notion of AI as a transformative force. Jillian Wanner, a principal at Deloitte responsible for AI staff development, acknowledged the ongoing revolution in the consulting industry at Nvidia's GTC Conference. Jim Rowan, the head of AI at Deloitte, has emphasized the importance of using AI effectively and encouraging employees to explore its capabilities.

KPMG is also taking a balanced approach to AI adoption. Todd Lohr, the head of ecosystems at KPMG, explained to BI that it is essential to balance top-down and bottom-up methodologies. Its incredibly challenging to understand what hundreds of thousands of employees do daily. By providing them with technology and allowing them to use it, they are coming up with even better and more innovative solutions than any top-down strategy could ever produce, he noted.

Initially, some employees faced confusion over how to effectively utilize generative AI. Lohr likened this to swivel chair processing, as it is challenging for individuals who have conducted tasks for decades to shift their mindset. However, KPMG has since gathered data on how employees are prompting AI, which has been instrumental in crafting new tools for both the firm and its clients by employing retrieval augmented generation techniques to enhance the specificity and accuracy of large language models.

As these consulting firms develop increasingly sophisticated tools, they recognize the need for centralized hubs to integrate their offerings. KPMG recently announced a partnership with Google Cloud to acquire licenses for Agentspacea platform that amalgamates AI agents with company datafor its U.S. workforce. Meanwhile, Deloitte unveiled Agent2Agent, a platform designed to improve interoperability between agents, marking the firm's largest collaboration with Google Cloud and ServiceNow. PwC has also introduced a similar platform called agent OS to streamline the operation of clients agents and facilitate collaboration among its internal agents, which have been developed over the past 18 months.

Amid a tumultuous period marked by layoffs, lost contracts, and budget cuts following the pandemic, the rise of generative AI represents a beacon of hope for consulting firms. As these firms continue to explore the capabilities of AI agents, they believe that the potential for increased efficiency and growth could lead to more significant organizational expansions rather than downsizing. I believe that as more agents become available, companies will not only see efficiency but also growth, said Matt Wood, PwCs global and U.S. commercial technology and innovation officer. This growth will enable them to double down on successful initiatives, ultimately resulting in larger organizations rather than smaller ones.