Generative AI Enterprise
Posts
McKinsey: 1265% surge in phishing attacks

McKinsey: 1265% surge in phishing attacks

Plus, Samsung case study, Deloitte on software quality, and more.

Lewis Walker
November 18, 2024

Welcome, executives and professionals.

We identify and breakdown the top 1% of Generative AI for enterprises.

This week:

McKinsey on GenAI opportunities and threats for cybersecurity.
Deloitte on maintaining software quality while leveraging GenAI.
Samsung tackles technical support with RAG.
Fast Fives: Transformation and technology in the news this week.
Career opportunities & events.

Read time: 4 minutes.

MARKET INSIGHT

McKinsey on GenAI opportunities and threats for cybersecurity

Brief: Cyberattacks surged since the launch of ChatGPT in late 2022, driving increased demand for enhanced cybersecurity solutions, with GenAI playing a key role in the problem and solution.

Breakdown:

Since ChatGPT's launch in late 2022, cyberattacks increased rapidly, with phishing (e.g. fake emails, websites) attacks rising 1265%, according to McKinsey.
As enterprises move to fortify their resources and workforce, skills in cloud security, AI, and zero-trust (no user trust by default) are key.
The current market opportunity for cybersecurity providers is $2 trillion with many enhancing offerings with AI-driven upgrades to meet threats.
Significant GenAI opportunities exist in security operations (SecOps), cloud, and endpoint security.
GenAI aids SecOps by suggesting and writing detection rules, analyzing large datasets for hidden threats, and suggesting actions. Providers report time savings up to 20–25%.
Other examples include AI assistants that autofill security questionnaires and reports, saving up to 80% of the time spent on these activities.

Why it’s important: The sharp rise in cyberattacks presents a major challenge for enterprises and professionals. GenAI will be key in determining who adapts, survives, and gains a competitive edge in view of rapidly evolving threats.

BEST PRACTICE

Deloitte on maintaining software quality while leveraging GenAI in enterprises

Brief: Deloitte’s article explores how engineering leaders can maintain software quality while leveraging generative AI for efficiency, addressing challenges such as synchronization dependancies of multi-agent solutions, the evolution of DevSecOps, and more.

Breakdown:

Synchronizing dependencies of multi-agent solutions: Opportunities explored include balancing trade-offs like single- vs. multi-agent systems, contextualizing architecture, adopting flexible middleware for model adaptability, and more.
Balancing code efficiency and accuracy: Key areas of focus include evolving code review metrics (e.g. human-in-loop review repository commit frequency), monitoring API leakage across application connections, and other considerations.
Non-standardized inputs and outputs challenging testing: Considerations include establishing prompt engineering libraries, adopting a microservices-like approach to break down complex tasks for consistent and testable outputs, and further considerations.
Adjusting DevSecOps for automated agents: The scope of analysis includes how GenAI's role in accelerating software development impacts operations typically requiring compressed continuous integration/delivery, and additional strategies.
The article also covers design transparency, data and architecture boundaries, integrating human-centered design early, ownership of AI-generated code, and more.

Why it’s important: As with any innovative technology, implementation challenges are to be expected. These actions together can help leaders and engineers ensure a high standard of software quality and performance, minimizing the accrual of technical debt and enabling long-term sustainability in software development.

CASE STUDY

Samsung tackles technical support with RAG

Brief: Samsung leverages RAG to enhance Kubernetes troubleshooting on its Samsung Cloud Platform (SCP), combining LLMs with external data for better technical support.

Breakdown:

Analysis of SCP support records showed 68% of Kubernetes container issues were resolved by users themselves using guides.
Kubernetes containers, integrated with Samsung Cloud Platform (SCP) as the Samsung Kubernetes Engine (SKE), offers added convenience but can complicate troubleshooting when issues arise.
Samsung created SKE-GPT, featuring a diagnostic area that checks cluster statuses against rule sets and SCP products, and an analysis area that generates solutions.
SKE-GPT overcomes the LLM limitation of lacking real-time, domain-specific data by using Retrieval-Augmented Generation (RAG) to incorporate external data, improving response accuracy and relevance.
Samsung's full case study explores indexing (document loaders, text splitters, embeddings, vector stores) and retrieval generation techniques.

Why it’s important: Samsung’s use of RAG demonstrates how GenAI can deliver quick, precise solutions tailored to user needs, based on years of SKE technical support knowledge.

Transformation

In the news this week:

McKinsey released a 7-page article on the EU AI Act, detailing key challenges for enterprises and preparation strategies. More than 50 percent of survey respondents said they are not clear on AI Act requirements.
Bain's research explores five design principles for improving customer experience with GenAI. BCG explored GenAI in supply chain management, discussing four levels of GenAI adoption, and implementation steps.
IBM published a 20-page report on AI governance for enterprises, addressing scaling challenges, defining roles, and outlining a high-level framework for responsible AI.
Gartner hosted a webinar titled ‘Responsible and Generative AI: The What, Why, and How’, sharing slides on revising GenAI policies, updating guidelines, and establishing evaluation processes.
IDC and Microsoft published its ‘2024 AI Opportunity Study’, surveying 4,000+ global business leaders and AI decision-makers. The 47-page report found that “For every $1 invested in generative AI, companies see an average ROI of 3.7x”

Technology

In the news this week:

Microsoft launched specialized small language models for sector-specific challenges in manufacturing, automotive, and agriculture. It also unveiled a digital AI-powered twin for virtual visits of Vatican City’s Basilica.
OpenAI plans to launch ‘Operator’, a new AI tool for completing complex tasks, in January. It also presented a blueprint for U.S. AI infrastructure and an update for ChatGPT to interact directly with third-party applications.
Google released a standalone Gemini iPhone app featuring Gemini Live voice conversations, image generation, and broader integration with Google services. It also open-sourced AlphaFold 3, its protein prediction model for research.
Anthropic added new developer tools to its Console to automatically improve prompts, manage examples, and evaluate outputs. The company also hired Kyle Fish as its first dedicated "AI welfare" researcher to explore AI consciousness.
Writer Enterprise AI platform secured a $200M Series C investment, increasing its valuation to $1.9B. The startup is set to expand in healthcare, retail, and financial services.