Welcome executives and professionals. A new study by McKinsey and MIT finds that AI investments in operations are paying back faster than ever, but leading organizations are pulling away from the rest.

Since the previous edition, we have reviewed hundreds of the latest insights in agentic and generative AI, spanning best practices, case studies, market dynamics, and innovations.

This briefing outlines what is driving material value — and why it’s important.

Note: Previously published as Generative AI Enterprise, this briefing is now titled Enterprise AI Executive.

In today’s briefing:

• McKinsey-MIT: Leading firms pull ahead with AI.

• The new era of enterprise app development.

• MIT’s State of AI in Business 2025.

• CMO shares agentic AI marketing opportunities.

• Transformation and technology in the news.

• Career opportunities & events.

Read time: 4 minutes.

MARKET INSIGHT

McKinsey-MIT: Leading firms pull ahead with AI

Brief: McKinsey–MIT study finds AI in operations is now delivering faster paybacks, with leading firms widening the performance gap. Based on 100+ surveys and 15 interviews, it updates a 4–5 year-old study.

Breakdown:

• Leaders’ AI adoption improves KPI performance 3.8x compared with the bottom 50% of companies, up from 2.7x in the previous survey.

• 89% of leaders use internal AI/ML development capabilities, compared with only 50% of firms in the bottom half of performance.

• Executives citing lack of expertise as a barrier to implementation fell to under 50%, down from nearly 70% in the prior survey.

• Still, over two-thirds of leading firms continue to leverage external partnerships for AI/ML development, versus the 50% of lower performers.

• Payback periods have shrunk to 6–12 months for both leaders and bottom-half firms, down from 12–18 months for leaders previously.

Why it’s important: As the performance gap between leaders and laggards widens, some firms risk being left behind in the race to adopt AI. Yet faster payback periods across the board signal opportunity: with the right strategy and execution, organizations can still leapfrog the competition.

AI-NATIVE INSIGHT

The new era of enterprise app development

Brief: Andreessen Horowitz explored how generative AI is reshaping internal app development by collapsing the gap between idea and execution, outlining how we got here and why the next era may arrive sooner than expected.

Breakdown:

• Internal app building long faced a tradeoff: scarce engineers or the rigid limits of incumbent low-code platforms that rose in the 2010s.

• New AI-native platforms like Replit, Lovable, Figma Make, and Bolt let natural language build UI, spin up databases, and handle deployment.

• Most output is still prototypical, built for validation or low-traffic use, but momentum points toward fully deployable, production apps.

• The biggest step-change over prior platforms comes from natural language interfaces, rapid iteration cycles, and greater customization.

• At Sears Home Services, non-technical staff built 50+ apps in Replit, covering ticketing, SMS alerts, and dashboards for routing.

Why it’s important: AI-native platforms aren’t yet replacing software teams, but they are reshaping how internal apps are scoped, tested, and socialized. With improved integrations, governance, and collaboration, they could move beyond prototyping to become the foundation for production apps.

MARKET INSIGHT

MIT’s State of AI in Business 2025

Brief: MIT NANDA’s 26-page State of AI in Business 2025, draws from 300+ public AI initiatives, 52 interviews, and a survey of 153 senior leaders across four industry conferences to track adoption and impact.

Breakdown:

• 80% of firms "investigated" "general-purpose LLMs" (e.g. ChatGPT, Copilot), yet only 40% were "successfully implemented" (in production).

• 60% "investigated" custom "task-specific gen AI", 20% piloted, and only 5% reached production, in part due to workflow integration challenges.

• 40% purchased official LLM subscriptions, yet 90% of workers report using personal AI tools at work, fueling “shadow AI.”

• 50% of gen AI spend goes to sales and marketing, though back-office initiatives often yield higher ROI (e.g. from BPO elimination).

• External partnerships where firms "procure external tools, co-develop with vendors" deliver 2x higher performance than "internally built tools."

Why it’s important: The starkest “Gen AI divide” is in deployment: just 5% of custom projects reach production. As with initial cloud and RPA efforts in the 2010s, firms often overestimate tech while undervaluing process design, governance, and training. Firms that adapt will likely join the 5% front-runners.

BEST PRACTICE INSIGHT

CMO shares agentic AI marketing opportunities

Brief: The CMO of enterprise AI company WRITER shared how AI has changed marketing: what once took weeks now takes hours. Customers demand hyper-personalized experiences, requiring both speed and custom fit from leaders.

Breakdown:

• Gen AI is the freelancer producing content from a prompt; Agentic AI is the manager and team, turning goals “launch a campaign” into execution.

• The matrix above maps agentic AI use cases across marketing functions and funnel stages, the foundations of an implementation roadmap.

• Teams can start with high-impact, low-complexity cases like competitive research, then progress toward multi-agent campaign orchestration.

• The insights cover six agentic AI use cases in marketing, outlining challenges, agentic workflows, enterprise examples, and outcomes.

• Qualcomm saved 2,400 hours monthly with personalization at scale, while 3,000 Salesforce employees reported a 20% productivity gain.

Why it’s important: Speed without custom fit leads to generic, off-brand content, while fit without speed means losing opportunities to the competition. The winners will master orchestration, building AI agents that embody their brand intelligence and execute their strategic vision at unprecedented speed.

BCG released a 31-page Oil & Gas executive playbook on building competitive advantage with AI, adding to its collection of AI transformation guides.

BlackRock explored role-based multi-agent systems for equity research and portfolio management, weighing their benefits and limitations.

Deloitte’s 2025 CPO survey of 250 leaders across 40 nations shows how procurement execs are embracing agentic AI amid market turbulence.

McKinsey and Workday CEO Carl Eschenbach discussed driving growth with agentic AI and building durable companies through serving others.

Microsoft released its 2nd guide in a 6-part series on building AI agents, covering best practices to help enterprises build and adopt agentic AI.

AWS introduced its Gen AI IDP accelerator, highlighted Amazon Q Business government use cases, and showed Infosys applying gen AI to data processing.

KPMG’s Board Leadership Center outlined AI’s pace, direction, and where boards should focus to harness value creation opportunities.

Infosys shared how agentic AI/MCP can transform engineering, what agentic AI means for financial services, and outlined its AI assurance framework.

Anthropic rolled out Claude Code to Enterprise, adding admin spend controls and policy settings. Claude Opus 4/4.1 can now end harmful chats.

OpenAI’s Altman said GPT-6 will arrive faster than GPT-5. He also warned of an AI bubble and compute limits holding back stronger models.

Microsoft is testing a new COPILOT function in Excel providing broader assistance in cells to generate summaries, classify data, and create tables.

Google reported Gemini chatbot consumes ~five drops of water per query. It will also offer it’s AI technology to U.S. federal agencies at $0.50 each.

DeepSeek launched V3.1 with a larger context window; Chinese media say delays of the R2 release are due to CEO Liang Wenfeng’s “perfectionism.”

Nvidia released Nemotron Nano 2, 9B–12B parameter reasoning models achieving strong performance at 6x the speed of similarly sized models.

The U.S. government rolled out USAi, enabling federal agencies to access secure AI chatbots, coding models, and other applications.

Cohere released Command A Reasoning, an enterprise model that outperforms rivals like GPT-OSS and DeepSeek R1 on agentic benchmarks.

CAREER OPPORTUNITIES

NVIDIA - Enterprise AI Senior Director

BlackRock - Head of AI Transformation

Meta - AI Policy Director

EVENTS

AWS - FSI Executive AI Roundtable - September 3, 2025

C3 AI - Federal Forum - September 9, 2025

The Washington Post - Global AI Summit - September 25, 2025

Originally conceived as a practical communication for executives the editor, Lewis Walker, has worked with, this briefing now serves as a trusted resource for thousands of senior decision-makers shaping the future of enterprise AI.

If your AI product or service adds value to this audience, contact us for information on a limited number of sponsorship opportunities.

We also welcome feedback as we continue to refine the briefing.