Edition in partnership with

Welcome executives and professionals. Adversaries are using frontier and open-source models to reason through exploit chains, inspect codebases, and weaponize vulnerabilities.

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.

In today’s briefing:

• Palantir’s agentic security strategy.

• Why organizations need an AI hub.

• Taming the enterprise AI token trap.

• AI-driven IT modernization for CIOs.

• Transformation and technology in the news.

• Insights for Executive+ members.

• Career opportunities & events.

Read time: 4 minutes.

BEST PRACTICE INSIGHT

Palantir’s agentic security strategy

Brief: Palantir shared lessons from more than a year of building an agentic software security strategy, from its internal multi-agent review harness through to the model-agnostic reviewer platform it now sells to enterprises.

Breakdown:

• Palantir's Product Security Team began testing agentic AI across security workflows over a year ago, well before Anthropic's Project Glasswing.

• It now runs multi-agent code review, analyst-directed vulnerability hunting, product triage and runtime validation in production.

• Palantir has productionized its model-agnostic AI reviewer platform, Security Forge, for customers to use for their own cyber defense.

• Five insights shaped both its strategy and the product, among them the principle that an organization's own context is its alpha.

Why it’s important: Drawing on that experience, Palantir outlines a nine-step starting sequence for building an agentic security pipeline, from a software inventory and a controlled harness around models through to a shared workflow for security and product engineers.

IN PARTNERSHIP WITH TELEPORT

Contain, audit, and control every agent

Brief: AI agents run 24/7 and spawn sub-agents at scale, so teams increasingly have to treat nearly every agent as a potential insider threat. Teleport Identity Security for AI extends least-privilege, ephemeral access, and cryptographic identity to agents.

With Teleport Identity Security for AI, you can:

• Capture what agents did and what they were thinking

• Enforce security intent, even when actors try to bypass policy

• Quantify and score the potential impact of an agent's actions

Agents don't fear consequences. See how Teleport contains them.

BEST PRACTICE INSIGHT

Why organizations need an AI hub

Brief: BCG detailed how companies often scatter their AI efforts across a wide range of pilots. An AI hub creates a single entity to oversee the entire AI portfolio, with the right capabilities, data and governance.

Breakdown:

• Companies are making massive investments in AI, but they often end up with fragmented programs that don't add up to meaningful value.

• Similar to a center of excellence, an AI hub serves as a central point of coordination, speeding up implementation.

• Hubs typically range from 10 to 100+ full-time employees. All need cross-functional expertise and authority to oversee AI efforts.

• Hubs evolve through four phases, from directly implementing AI projects to coordinating business units’ own capabilities (image above).

Why it’s important: AI is becoming ubiquitous and every company can access powerful models, infrastructure and tools. The real advantage comes from the operating model: the structure, governance, leadership and talent that allow AI value to compound across the enterprise.

BEST PRACTICE INSIGHT

Taming the enterprise AI token trap

Brief: Oliver Wyman detailed how procurement controls can curb rising enterprise AI spending, giving COOs, CPOs, and finance leaders a way to make AI a high-value investment rather than an uncontrolled cost.

Breakdown:

• Token usage may attract the most attention, but it represents only one component of the exponential rise in the total cost of AI.

• Costs are also increasing fast for licenses, cloud, implementation, security, governance and AI premiums embedded in tech contracts.

• These costs behave like a hybrid model, mixing fixed subscriptions with variable usage, so partial control leaves spending unchecked.

• Rather than tackling categories individually, treat AI as a single budgetary category spanning all suppliers, contracts and business units.

Why it's important: As adoption accelerates, costs risk scaling faster than the value AI delivers. The most pressing challenge for COOs and CPOs is governing AI spending with the visibility, commercial discipline and cross-functional control that enables cost-effective innovation.

AI-NATIVE PROFESSIONAL

How to model strategic scenarios with ChatGPT

Brief: In this guide, you'll learn how to give ChatGPT a financial model, KPI dashboard, planning documents, market context, and decision criteria, then ask it to compare options and produce a reviewable trade-off model.

Step-by-step:

1. State the decision, options, time horizon and criteria, then attach financial models, planning docs, market and operational data.

2. Ask ChatGPT to identify assumptions that are shared, option-specific or missing, then run the starter prompt to build the model.

3. Review the model's cost, timing, risk, ownership and customer impact, changing assumptions explicitly to compare across scenarios.

4. Use the follow-up prompt to test sensitivity, varying the assumptions that matter most to see when the preferred option changes.

Best practice: Keep the source model intact and save each scenario output separately, noting which assumptions would change the recommendation.

For the full guide, including prompts, upgrade to Executive+ or The Boardroom.

BEST PRACTICE INSIGHT & CASE STUDIES

AI-driven IT modernization for CIOs

Brief: BCG shared common misconceptions about AI-driven IT modernization that can undermine value, increase risk and slow progress, along with what leaders should do instead at each stage of the modernization journey.

Breakdown:

• IT modernization is back at the top of CIO agendas, as core systems from home-grown apps to decades-old COBOL become liabilities.

• Generative and agentic AI have renewed optimism, but capturing value at scale means recognizing and resisting misconceptions (image above).

• Modernization spans understanding business intent, system analysis, architecture, coding and testing, each needing different AI support.

• Using the right AI pattern for the right task and context can reduce costs by 25% to 35% and speed delivery by 30% to 40% (image above).

Why it’s important: The real promise of AI in IT modernization comes from disciplined use of the right AI, grounded in the right context, at each stage of the journey. When executed well, it cuts costs, accelerates delivery and produces higher quality systems that are easier to maintain over the long term.

Everest Group shared insights on open-weight AI, from the capability gap and geopolitical risk to the enterprise adoption paradox.

Deloitte published a 30-page report on controlling cloud spend as infrastructure, data transfer, and AI services drive disruption.

Bain outlined what a modern data platform is, mapping five enterprise workloads and three architectures against the semantic layer beneath.

McKinsey explored where AI agents pay off, mapping economic opportunities and what drives the unit economics of agentic workflows.

EY explored how AI can build resilience against risks not yet imagined, and how geospatial data turns earth observation into business insight.

Porsche signed a five-year, $1.5B deal with TCS to run a hub overseeing AI deployment, including transferring its IT consultancy for €320M.

Google Cloud launched industry-tuned Gemini Enterprise editions for financial services and legal in preview, with healthcare and life sciences next.

Anthropic rolled out a single Claude memory shared across chat and Cowork, now able to save topics in real time mid-conversation.

OpenAI brought its GPT-5.6 family to AWS coding agent Kiro, with testing showing Terra completed tasks at roughly 82% lower cost.

Nvidia struck a $6B deal to license Poolside's model-development technology and bring 100+ engineers onto the open-weight Nemotron team

SpaceX announced it will build its Starmind space data centers around Nvidia's Vera Rubin NVL72 rack, with Musk targeting orbit by Q4 next year.

Hugging Face, which hosts over 2M models and 1.5M datasets, is exploring a sale at $13B or more, up from its $4.5B valuation in 2023.

CAREER OPPORTUNITIES

Databricks - AI Transformation Leader

AXA XL - Head of Artificial Intelligence

U.S. Bank - AI Strategy Leader

EVENTS

OpenAI - ChatGPT Work for Marketing - September 3, 2026

EY - Preparing for Agentic Commerce - September 23, 2026

Oracle - AI World - October 25-28, 2026

Reach enterprise AI decision-makers:

• 66% of readers are C-level executives or VP and Director-level leaders.

• 63.2% of the audience is based in the U.S., EU, UK, ANZ, and Singapore.

• Read by leaders at Microsoft, Deloitte, the Fortune 500, and more.

Guaranteed impression and custom sponsorship packages available, with post-send performance reporting.

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

We welcome your feedback.