Featured
Delivering tangible business impact from generative AI
GAIL from QuantumBlack, AI by McKinsey
QuantumBlack Labs is the R&D and software development hub within QuantumBlack, AI by McKinsey. QuantumBlack Labs has over 250 technologists dedicated to driving AI innovation and supporting and accelerating the work of McKinsey’s 1400+ data scientists. We use our collective experience to develop suites of tools and assets that ensure AI/ML models reach production and achieve sustained impact.
Generative AI: balancing optimism with realism
Projects using artificial intelligence and machine learning (AI/ML) are a key driver of business value. Companies that successfully capture the economic potential of generative AI (gen AI) are those that recognize and invest in areas where gen AI delivers tangible value. As gen AI matures, we can analyze real-world lessons from early adopters to understand the specific domains where investment pays dividends, such as:
- Intelligent chat systems that transform customer and employee interactions.
- Gen AI–enabled “agents” that execute complex, multistep workflows.
- Data exploration tools that use techniques like agents and retrieval-augmented generation (RAG).
- Tools for document and insight generation that connect disparate data sources.
- Co-pilot systems that assist with complex tasks across multiple domains.
For businesses navigating AI transformation there is a potential disconnect between high expectations and practical deployment, which is exacerbated at scale. The challenges of scaling gen AI are both technical, such as overcoming siloed and unstructured data, and organizational, with change management, talent acquisition, and workforce retraining required.
Some gen AI applications, like intelligent chat and data exploration, are gaining traction. Scaling these solutions across diverse industries, however, comes with challenges such as data privacy concerns, the need for domain-specific adaptation, and integration with existing IT infrastructures.
Introducing QuantumBlack Gen AI Labs (GAIL)
To help enterprises codify their expertise, automate processes, and invest in the areas that will yield the greatest impact, QuantumBlack Labs has established Gen AI Labs (GAIL). GAIL makes gen AI accessible, impactful, and easy to integrate, helping organizations navigate the complexities of adoption to drive meaningful results.
The GAIL toolkit for agentic workflows can rapidly accelerate time to impact. Agents represent an opportunity for subject-matter experts to codify complex business expertise into the tech stack in the form of robust software applications.
Agentic systems built using foundation models have the potential to reason and adapt to different scenarios and accomplish a complex workflow. GAIL offers a range of AI-powered agents that can transform complex workflows and automate routine processes, configurable and adaptable for different businesses. The GAIL architecture caters both to developers of complex workflow automation and to the business user wishing to configure agents for reuse.
The GAIL product suite includes production-grade assets built from open-source components using the latest gen AI methods to make them reliable, adaptable, and ready for enterprise deployment. Blueprints are one of the key features of the GAIL product suite, and offer a set of prebuilt solutions for adding out-of-the-box gen AI capabilities to different workflows:
- Chatbot blueprint enhances customer and employee interactions through conversational AI.
- Search blueprint for gen AI-powered search capabilities to gather relevant information quickly from large datasets.
- Label blueprint automates data labeling processes, helping to streamline machine learning model training.
- Co-pilot blueprint offers virtual assistance for complex tasks, enabling users to achieve more in less time.
Blueprints enable teams to maximize the time spent training gen AI systems on what “good” looks like for their specific contexts, then integrate gen AI without incurring the costs, time, and complexity of building a custom solution.
GAIL’s suite of applications continues to expand rapidly through QuantumBlack’s collaboration with industry players across the entire AI stack to learn from and adopt the latest innovations for our clients.
The impact of QuantumBlack GAIL
Overhauling the architectural approach with GAIL
GAIL supported a large financial services group in overhauling its approach to delivering market insights through an investment chatbot. Advisors previously spent excessive time on analyst research, sometimes requiring customers to read lengthy reports on their own.
Over a few months, GAIL deployed an advanced cloud-based architecture, integrated best-practice code, and guided rapid development from a monolithic proof of concept to a deployment-ready product. By focusing on efficient workflows and scalable foundations, the GAIL team helped the organization to expedite its go-to-market plan.
Significant achievements include:
- Accelerated development time, moving from proof of concept to product in just two months (less than half of a typical development period).
- Enterprise-scale readiness.
- Enhanced information retrieval, giving advisors immediate access to relevant research insights.
Encoding domain expertise and accelerating development
GAIL helped a global health organization transform its complex, multi-week reporting processes. By deploying tailored gen AI workflows, GAIL dramatically accelerated report generation, enabling the organization’s experts to focus on strategic decisions rather than laborious data gathering. Within a few weeks, the team incorporated a prompt library, processed more than 2.5 GB of documents, and introduced near-real-time insights.
Key achievements include:
- Reducing drafting time from one month to one day, unlocking over $1Bn+ in accessible funding through faster proposal turnaround.
- At least 85 percent improvement in quality compared to manually created reports.
- <1 week to pivot to adjacent use cases
The impact of GAIL was considered by one Director to have “accelerated the organization’s “broader gen AI adoption by approximately two years”.
Agentic workflows: the next frontier
GAIL collaborated with a leading US bank to automate credit memo drafting, which is a traditionally lengthy process requiring the careful synthesis of data from numerous sources. The bank’s credit analysts and relationship managers often spent days preparing each memo, leading to bottlenecks and a higher risk of error. Using agentic workflows and focusing on automation of both financial analysis and credit-risk narratives, GAIL reduced memo preparation time while improving consistency.
Early results include:
- 20–60 percent productivity gains among credit analysts.
- Approximately 30 percent faster decision-making, thanks to automated data synthesis.
- A scalable blueprint for broader AI initiatives, supporting more lines of business with minimal rework.
Summary
As gen AI technologies have matured, there has been a shift from initial hype to a period of genuine growth. The possibilities of gen AI have switched from theoretical to both tangible and transformative, and organizations are looking for ways to use the technology to extract a competitive advantage.
The high cost and complexity of building custom AI solutions can be prohibitive, and organizations need accessible, ready-made tools to integrate AI capabilities effectively. By focusing on their core capabilities and using GAIL to transform domain expertise and best practice into AI solutions, companies can speed up adoption and integration, shortening time to value.
Successful organizations will be those adept at blending the potential of gen AI innovation with practical implementation strategies; finding a balance between optimism and realism. By capitalizing on their strengths, while tapping into QuantumBlack’s industry knowledge and strategic collaborations, they can lead the way in gen AI advancements, and fundamentally shift their businesses.
Thanks to all who contributed to this article: Ottis Kelleghan, Stephen Xu, Carlo Giovine, Tomás Lajous, Yva Montalvo, Jo Stichbury, Joanna Sych.
