Consulting powerhouse McKinsey & Company has developed an artificial intelligence tool named Lilli that could radically transform how consultants work. This innovative AI assistant provides data, insights, and strategic recommendations in seconds, freeing up consultants’ time for more high-value work.
What is Lilli and Why Did McKinsey Create It?
Lilli is a conversational AI chatbot that McKinsey built using large language models (LLMs) from partners like Cohere and OpenAI. It is hosted on Microsoft’s Azure cloud platform.
The goal with Lilli is simple – help McKinsey consultants work faster and smarter. The tool digs through McKinsey’s trove of over 100,000 internal documents and transcripts to provide tailored advice based on consultants’ prompts. It can synthesize insights more rapidly than any human.
Lilli’s creation was led by McKinsey’s Chief Technology Officer, Jacky Wright, and the Client Technology (ClienTech) team. They rightly saw an opportunity to leverage AI to enhance productivity and quality of client work.
Top Benefits and Applications of This AI Assistant
In a short beta testing period with 7,000 employees, Lilli has already demonstrated tremendous utility:
Massive time savings – Lilli can reduce research and planning from weeks to merely hours or minutes. The efficiency boost is unprecedented.
Better insights – By analyzing such a vast amount of internal data, Lilli can uncover patterns and insights consultants may have missed.
Improved client deliverables – Higher quality recommendations and strategic plans can be created for clients.
Enhanced teamwork – By automating tedious tasks, Lilli lets consultants spend more time collaborating and innovating.
Standardization – Solutions have the consistency and quality assurance that comes from McKinsey’s institutional knowledge.
On-demand access – Lilli is available 24/7 and can concurrently support many client teams.
How McKinsey Designed This Tool for a Familiar Experience
The creators of Lilli designed the interface to be immediately intuitive. In some ways, it is reminiscent of ChatGPT and other conversational AIs. However, Lilli has differentiating features tailored for consulting use cases:
Expandable sidebar – Users can save prompts in a sidebar to easily modify and reuse.
Cited sources – Lilli provides citations for all data to ensure transparency.
Contextual responses – Remembers user’s company, industry, past questions, etc. for relevant insights.
Follow-up prompting – Users can ask questions to clarify or expand on Lilli’s answers.
Secure collaboration – Lilli can be used by multiple team members working on an engagement.
The Cutting-Edge Technology Behind Lilli
As an industry leader, McKinsey leveraged best-in-class AI capabilities to build Lilli:
Large language models – Lilli uses LLMs like GPT-3 and others to generate human-like text.
Microsoft Azure – The cloud platform provides data storage and high scalability.
Proprietary algorithms – McKinsey added custom models to improve Lilli’s consulting knowledge.
Built-in guardrails – Lilli has constraints to prevent sharing confidential data.
Ongoing training – Lilli’s capabilities will expand as McKinsey continues feeding it industry documents.
What’s Next for This Game-Changing Consulting Tool?
McKinsey plans to expand Lilli usage internally and explore external applications as well:
Wider internal rollout – Lilli will eventually be provided to all McKinsey consultants globally.
Client access – Some clients may be granted limited access to Lilli for improved collaboration.
Licensing to other companies – McKinsey sees monetization potential in licensing Lilli out.
Solving more complex problems – With more training data, Lilli will handle increasingly sophisticated queries.
Integration with McKinsey software tools – Lilli can power up other McKinsey solutions like QuantumBlack’s Ada.
Lilli represents a massive leap forward for augmenting human consultants with artificial intelligence. It has the potential to radically transform consulting by generating insights at speeds not humanly possible. Other firms will surely follow McKinsey’s lead, but the venerable consultancy appears to have gained a significant competitive advantage.FAQsWhat types of questions can Lilli answer?
Lilli can provide strategic advice, data-driven insights, and recommendations on a wide range of consulting issues. However, it is optimized to leverage McKinsey’s proprietary knowledge base.Is Lilli’s advice unbiased?
Lilli forms responses based solely on McKinsey internal data. However, McKinsey implements guardrails to prevent sharing of confidential information.How was Lilli trained?
Lilli was trained on over 100,000 McKinsey documents and transcripts. McKinsey continues to feed it more data, expanding its knowledge.Will Lilli replace human consultants?
No. Lilli is a productivity-enhancing AI assistant for McKinsey’s consultants. It augments human expertise rather than replaces it.Can anyone access and use Lilli today?
Currently, Lilli is available only to McKinsey employees. However, McKinsey plans to provide access to clients and license it to other companies in the future.How does Lilli ensure data privacy?
Lilli has built-in constraints to prevent confidential data leakage. It also attributes any insights provided to their original McKinsey source materials.What makes Lilli better than other AI assistants?
Lilli is optimized specifically for consulting use cases, leveraging McKinsey’s unmatched industry knowledge base. No other tool has this level of tailored business acumen.