Generative AI Enterprise
Posts
Google reveals proven quick wins

Google reveals proven quick wins

Plus, multi-agent design, red teaming 100+ products, and more.

Lewis Walker
January 15, 2025

WELCOME, EXECUTIVES AND PROFESSIONALS.

Google’s customers are driving impressive results with gen AI, especially in three key areas. The question is, what are they, and how can you capture the value?

Since the last edition, we've reviewed hundreds of the latest insights on best practices, case studies, and innovation. Here’s the top 1%...

In today’s edition:

Google reveals gen AI quick wins driving value today.
Microsoft launches CoreAI division to advance AI mission.
Deloitte showcases multi-agent AI systems in action.
Microsoft’s key lessons from red teaming 100+ gen AI products.
Transformation and technology in the news.
Career opportunities & events.

Read time: 4 minutes.

BEST PRACTICE INSIGHT & CASE STUDIES

Google reveals gen AI quick wins driving value today

Image source: Google Cloud

Brief: Google's 53-slide report offers insights from its AI experts and latest customers, providing a framework for assessing gen AI use cases, a list of proven quick wins delivering value and customer case studies.

Breakdown:

The framework assesses use cases by risk and reward, categorizing quick wins (6-12 months), enterprise intelligence (1-2 years), and transformative (3+ years).
Google's customers achieve quick wins with gen AI in three areas: productivity, customer experiences, and back-office automation.
Increased productivity for knowledge workers and coders by accelerating code development, simplifying DevOps, content generation, and beyond.
Enhancing customer experience by boosting agent productivity, improving self-service, enhancing customer insights, and more.
Automated back-office processes including contract management, HR help desk, compliance processes, accounts payable, and other key processes.
It also includes insights from C-suite leaders seeing gen AI results: Turing, Pfizer, Estee Lauder, Victoria's Secret, GE Appliances, Forbes.

Why it’s important: Leaders face the challenge of navigating the hype cycle to deliver value. Early adopters are seeing ROI (explore more in Google’s gen AI ROI report). The key is balancing immediate gains with long-term potential, focusing on quick wins and demonstrated value before moonshots.

ENTERPRISE INNOVATION

Microsoft launches CoreAI division to advance mission

Image source: Microsoft

Brief: Microsoft established a new CoreAI division, unifying its AI platform and developer tools to accelerate the development of its end-to-end Copilot stack and agentic applications, marking a key progression to AI-first application development.

Breakdown:

Satya Nadella announced a new phase in the AI platform shift, with 2025 focusing on model-forward applications to transform all app categories.
This shift will impact every application stack layer, compressing “thirty years of change into just three years.”
It's akin to the simultaneous introduction of GUI, internet servers, and cloud-native databases into the app stack.
The AI-first app stack will feature new UI/UX patterns, agent runtimes, multi-agent orchestration, and reimagined management and observability layers.
Nadella emphasized that Azure must become the infrastructure of AI, with Azure AI Foundry, GitHub, and VS Code built on top of it.
Microsoft aims to create agents that will drive change across all SaaS categories, enabling custom apps built by software (“service as software”).

Why it’s important: To accelerate and seize the AI opportunity, it's a reminder for enterprises that internal organizational boundaries mean little to customers and competitors. What's important is reinventing for innovation and accountability to fulfill the mission.

BEST PRACTICE INSIGHT

Deloitte showcases multi-agent AI systems in action

Image source: Deloitte

Brief: Deloitte's report, The Cognitive Leap, explores reimagining work with AI agents and multi-agent system design, breaking down a practical example and building on last year’s, Prompting for Action report.

Breakdown:

Multi-agent AI systems can help transform traditional, rules-based business and IT processes into adaptive, cognitive processes.
The report recommends design principles, borrowing from composable design, microservices architecture, and human resources strategies.
It outlines a reference architecture for agent-powered transformation in financial services, detailing interaction, workflow, agents, and operations layers.
Subsequently visualizing this architecture transforming an inefficient IT service desk support process into a multi-agent AI system solution (see pages 11-13)
The report outlines seven steps for success, such as pinpointing the right data in the right context, tapping talent, and more.

Why it’s important: By leveraging the foundational design principles, reference architecture for reuse and rapid adaptation of core components, and process transformation example, enterprises can start to explore the potential and scalability of multi-agent AI systems.

BEST PRACTICE INSIGHT & CASE STUDIES

Microsoft’s lessons red teaming 100+ gen AI products

Image source: Microsoft

Brief: Microsoft’s 21-page white paper details its AI red team ontology, eight key lessons from "red teaming 100 generative AI products", and five case studies from its experience since 2021.

Breakdown:

AI red teaming probes systems for safety and security by “breaking” them to identify weaknesses and rebuild them back with stronger defenses.
Microsoft’s ontology models attack components: actors (adversarial or benign), TTPs (Tactics, Techniques, and Procedures), system weaknesses, and downstream impacts.
Gen AI integration introduces novel attack vectors, but AI red teams must consider both new and existing cyberattack vulnerabilities.
Mitigations don’t remove risk entirely. AI red teaming adapts to evolving threats, raising the cost of successful system attacks.
While automation helps orchestrate attacks, AI red teaming also relies on human expertise, cultural awareness, and emotional intelligence.
Microsoft’s five case studies highlight vulnerabilities across traditional security, responsible AI, and psychosocial harms using their ontology.

Why it’s important: These lessons highlight the importance of both technology and human expertise in securing AI systems. The paper also references Microsoft’s PyRIT (Python Risk Identification Tool), an open-source framework to identify vulnerabilities in AI systems, valuable for enterprises organizing red teaming exercises.

BCG released an article on how the convergence of AI-powered agents and innovative hardware will enable brands to deliver superior customer experiences at a lower cost-to-serve.

McKinsey published an article on scaling gen AI in the life sciences industry, addressing organizational maturity, common scaling challenges, a five-point plan to unlock value, and what a successful initiative typically looks like.

Capgemini's 92-slide report explores the surge in semiconductor demand and opportunities to enable the future of AI, based on insights from 250 semiconductor executives and 800 downstream leaders.

Dentons, the law firm, released a 28-slide report outlining the top legal issues in AI you need to know about for the year ahead, covering AI regulation, data privacy, procurement, talent, intellectual property, antitrust, and more.

Bloomberg predicts that over 200,000 Wall Street jobs may be eliminated within 3-5 years due to AI, with automation and productivity gains potentially boosting banking profits by up to 17%.

Microsoft released AutoGen v0.4, an update to its open-source agentic framework, alongside Magnetic-One, a new orchestration system that coordinates multiple AI agents to tackle complex tasks.

OpenAI released a comprehensive policy framework outlining how the U.S. can sustain AI leadership while promoting equitable access and economic growth, drawing parallels to America's historical handling of transformative technologies.

The United States introduced unprecedented export controls on AI chips, creating a tiered global system to maintain U.S. technological leadership while restricting access for China and other strategic competitors.

Minimax, a chinese AI lab, launched two new open-source models using a novel ‘Lightning Attention’ approach, enabling massive 4M token context windows while maintaining speed and performance.

Google Cloud launched its Automotive AI Agent powered by Gemini, with Mercedes-Benz integrating the technology into its MBUX Virtual Assistant to enable multimodal reasoning for in-vehicle interactions.