Edition in partnership with

Welcome executives and professionals. Agentic AI doesn’t just change what work gets done, it changes who owns execution. That shift carries real accountability.

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:

• Genpact’s agentic AI buyer’s guide.

• Cognizant: Most workers want more AI.

• Advantage comes from difference.

• Not every workload needs sovereignty.

• Transformation and technology.

• Insights for Executive+ members.

• Career opportunities & events.

Read time: 4 minutes.

Important note: Starting this Wednesday, October 14, Enterprise AI Executive will be published once a week, on Wednesdays. We will no longer publish on Sundays. The number of Executive+ insights in each edition will double.

BEST PRACTICE INSIGHT

Genpact’s agentic AI buyer’s guide

Brief: Genpact published a 30-page buyer’s guide to agentic AI, arguing that because agents take over execution, enterprises need a higher bar for buying them, and that trust and readiness are the main barriers to scaling them.

Breakdown:

• Leaders must first settle four decisions: who owns outcomes, how to measure success, workforce impact and process design.

• A build-vs-buy matrix scores workflows on six factors, three shown above: differentiation, internal capability and speed to value.

• It flags five costs business cases often miss, including integration, oversight design, workflow redesign and accountability rules.

• An eight-point due diligence checklist weights criteria by buyer type and flags missing audit trails or explainability as red flags.

Why it’s important: Unlike assistive tools, agents make decisions and act across business systems, so a poor purchase carries greater risk. Executives who set accountability, governance and metrics before they buy will be better placed to scale agents and show a return.

IN PARTNERSHIP WITH THE HACKETT GROUP®

The emerging 75% AI performance gap

The first phase of enterprise AI focused largely on experimentation and deployment. Increasingly, attention is shifting to whether AI is changing operating performance.

New benchmark research from The Hackett Group® suggests a measurable gap is emerging between organizations that redesign end-to-end processes around AI and those that primarily automate existing workflows.

Across multiple business functions, the modeled performance difference reaches up to 75%, reflecting improvements in cost, productivity, speed and operating leverage.

Order-to-cash illustrates the pattern. Organizations operating at AI World Class performance levels achieve 52%-59% lower process costs, 56%-64% lower staffing requirements, 43% faster dispute resolution, and 85% fewer delinquent days than peer organizations.

Rather than resulting from a single automation initiative, these outcomes stem from changes across credit, order capture, billing, cash application and collections that reinforce one another throughout the revenue cycle.

The research examines where AI-enabled process redesign is producing measurable business outcomes and how benchmark data can be used to prioritize enterprise AI investments.

Read the research.

MARKET INSIGHT

Cognizant: Most workers want more AI

Brief: Cognizant published a 34-page report surveying over 10,000 workers, finding that most want to use AI more than their employer allows, and that attitude predicts how people use AI about six times better than seniority does.

Breakdown:

• It maps seven adopter types into three groups (image above): about 20% are out in front, 75% are held back and 4% are unconvinced.

• Between 72% and 82% of workers in every group say their AI tools fall short, citing missing company context and poor workflow fit.

• Use remains mostly basic, led by data analysis and summarizing, with only 18% directing agents and 25% using AI for strategic work.

• Cognizant urges leaders to drop one-size-fits-all rollouts, remove permission barriers and build role-specific training by type.

Why it’s important: Rolling out the same tools and training to everyone misses the biggest opportunity: the three in four workers who are willing but held back. Leaders who tailor access, training and tools to each group will be best placed to turn AI spending into returns.

BEST PRACTICE INSIGHT

Advantage comes from difference

Brief: Arthur D. Little argued that as AI becomes cheaper, widely available infrastructure, much like electricity, it stops being a differentiator, shifting advantage to human judgment, diverse thinking, and proprietary know-how.

Breakdown:

• ADL warns of a “convenience trap,” where leaders who hand decisions to AI gradually lose the independent judgment they depend on.

• Shared models push firms toward similar strategies, while over-reliance erodes skills, as doctors used to AI got worse without it.

• To counter this, ADL urges leaders to make judgment part of roles and performance reviews, and to audit how teams decide with AI.

• It also recommends rewarding constructive dissent, and running small real-world experiments to build knowledge no model has.

Why it's important: When every company can access the same models, AI alone won’t set firms apart. Executives who protect their people’s judgment, encourage different perspectives and build proprietary know-how will be best placed to turn AI into competitive advantage.

AI-NATIVE PROFESSIONAL

How to review forecast risk with ChatGPT

Brief: In this guide, you'll learn how to give ChatGPT your forecast data, deal records, calls, emails, legal status, usage data and owner notes, then ask for a sourced risk review of which deals belong in commit, upside or pull.

Step-by-step:

1. Define the forecast period, teams or deals in scope, and what counts as commit, then attach your CRM export and forecast snapshots.

2. Add calls, emails, deal threads, legal, procurement, usage and owner notes, and ask ChatGPT to separate facts from inferred risk.

3. Run the starter prompt, then review the rationale, blockers and next owner action ChatGPT gives for each deal before acting.

4. Confirm changes with the sales owner before updating the forecast, then use the follow-up prompt to build an efficient call agenda.

Best practice: Don't infer confidence from stage alone. Each recommendation should show the evidence, missing context, urgency and close-path risk.

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

MARKET & BEST PRACTICE INSIGHT

Not every workload needs sovereignty

Brief: HCLTech and Everest Group published a 26-page paper encouraging firms to match sovereign cloud controls to each workload’s risk, warning delays build “sovereignty debt” that grows costlier to undo as dependencies deepen.

Breakdown:

• A gap score sorts firms into four levels (image above), from sovereignty-aligned to structurally vulnerable, where fixes cost most.

• It scores workloads on five exposure tests, including regulation, data sensitivity, operational criticality and geopolitical risk.

• It estimates about 25% of enterprise workloads are sovereignty-sensitive, with sovereign cloud costing about 15% more than standard.

• The paper urges firms to extend controls to AI training data, model hosting and inference, and to make compliance checks continuous.

Why it’s important: With geopolitics and AI regulation shifting fast, and under 1% of cloud customers switching providers each year, today’s cloud choices are hard to reverse. Leaders who classify workloads now can avoid paying sovereign premiums where they aren’t needed.

Deloitte shared how CIOs can use an investment portfolio approach to tie AI spending to value and decide where capital goes next.

McKinsey explained why AI’s easiest wins are misleading CEOs, and shared five signals from Dreamforce 2026 on why speed is the edge.

EY published an 11-page guide to the total cost of AI agents, prompting firms to manage their run-rate spend as a growth investment.

Bain explored realization rate, the metric reshaping enterprise AI, and spoke with Prudential’s Pankaj Banerjee about scaling AI.

AWS explored building the business case for agentic automation, and shared a playbook for closing the AI knowledge-capability gap.

Infosys detailed how an AI capability maturity model shows organizations where they stand on AI adoption and where they need to get.

Anthropic released Claude Haiku 5.5, matching GPT-6 Luna’s $0.10 per million input tokens while beating it on multiple benchmarks.

OpenAI debuted “intelligent UI,” a GPT-6-powered feature that lets ChatGPT answer with interactive charts, maps, tools and more.

Google Cloud unveiled the Gemini agent at Gemini at Work 2026, working across Workspace apps and orchestrating Gemini and Claude.

Anthropic launched Claude Dashboards and Motion, which turn data into interactive dashboards and prompts into animated explainers.

Nous Research raised $90M at a $1.5B valuation, revealing its open-source Hermes agent now handles about 2.5% of global token usage.

Anthropic launched OSS Scanner, a free tool to find open-source flaws, and added its first usage rule protecting Claude from “sustained” abuse.

CAREER OPPORTUNITIES

Anthropic - AWS GTM Partnership Lead

Ralph Lauren - Head of AI Strategy

Accenture - Head of AI, Microsoft

EVENTS

Google Cloud - Executive Roundtable - October 22, 2026

Deloitte - Tokenomics and AI Observability - October 28, 2026

Snowflake - Pilot to Production - December 3, 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.