WELCOME, EXECUTIVES AND PROFESSIONALS.

SaaS incumbents are well-positioned to lead, if they move fast. But this requires courage and high-stakes strategic bets.

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

In today’s edition:

• Bain: Will AI agents disrupt SaaS?

• Capgemini’s $3.3B WNS acquisition.

• LangChain: How to build an agent.

• xAI releases Grok 4.

• Transformation and technology in the news.

• Career opportunities & events.

Read time: 4 minutes.

PLATFORM STRATEGY

Bain & Company: Will AI agents disrupt SaaS?

Brief: Bain shared insights on whether AI and agents will disrupt SaaS. When SaaS emerged 25 years ago, it moved software to the cloud. Today, agentic AI threatens to shift knowledge work from “human + app” to “AI agent + API.”

Breakdown:

• SaaS execs can map workflows into four scenarios to set priorities, where AI enhances via automation (A) and replaces via penetration (P).

• AI enhances SaaS (Low A, Low P): AI adds to core SaaS. The importance of domain expertise supports incumbent defensibility.

• Spending compresses (Low A, High P): Third-party agents tap APIs and siphon value. Incumbents should deploy their own and secure endpoints.

• AI outshines SaaS (High A, Low P): Valuable AI agents sit atop SaaS. Incumbents should double down on agents and outcome-based pricing.

• AI cannibalizes SaaS (High A, High P): Most value shifts to AI, diluting SaaS. Here, incumbents must disrupt themselves to stay relevant.

Why it’s important: Will AI and agents disrupt SaaS? Yes. In some cases, that disruption will grow the market; in others, it will commoditize the market. In some cases, the disruption will favor incumbents; in other cases, it will favor new entrants. Disruption is mandatory, but obsolescence is optional.

BUSINESS MODEL TRANSFORMATION

Capgemini’s $3.3B acquisition of WNS

Brief: HFS Research asserted that Capgemini’s $3.3B acquisition of WNS signals the end of legacy BPO, as enterprise demand shifts from FTE-based service models to the adoption of generative and agentic AI.

Breakdown:

• Legacy BPO is on its way out: clients no longer want to just outsource work, they want to reimagine how work gets done altogether.

• Clients increasingly demand AI-driven, orchestrated outcomes enabled by consulting, technology, and domain expertise.

• Capgemini brings consulting and AI scale. WNS adds vertical capabilities, reinvention-ready BPO clients, and outcome-linked commercial models.

• If Capgemini and WNS integrate well post-acquisition, they form a strong AI + Ops combo, and a win-win for the combined client base.

• Capgemini + WNS isn’t alone. Accenture, for example, merged Strategy, Song, Tech, and Ops to drive AI-first, outcome-led reinvention.

Why it’s important: Capgemini + WNS is well-positioned to help shape the next evolution of AI transformation. Over the next 12–18 months, expect consolidation, restructuring, and rebranding as legacy BPO providers pivot. Because in this new era, clients are paying for outcomes, not FTEs.

BEST PRACTICE INSIGHT

LangChain: How to build an agent

Brief: LangChain, a top platform for building AI agents, shared a framework for implementing them in production. Whether automating email triage or orchestrating complex workflows, these six steps provide a practical approach.

Breakdown:

• Identify a realistic agent use case, teachable to an intern, and outline 5-10 concrete scenarios the agent should handle.

• Create a step-by-step procedure, breaking down the task(s) into clear steps and establishing the groundwork for development.

• Design agent architecture, build prompt(s) that handle core reasoning, connect to live data sources and build orchestration logic.

• Evaluate agent against examples, identify failures and iterate to improve reliability and performance.

• Deploy the agent and monitor how users interact, continuously refine based on real usage patterns.

Why it’s important: While many enterprises are talking about building agents this year, far fewer have done it. It’s easy to let your imagination run wild with how agents can transform your business, but many teams are unsure where to begin. So start small, stay user-focused, and keep refining.

INNOVATION INSIGHT

xAI releases Grok 4

Brief: xAI released Grok 4 and Grok 4 Heavy reasoning models. In just 28 months since its founding, it now holds the top-ranked AI model, validated by independent benchmarks, and is showing early signs of enterprise traction.

Breakdown:

• Grok 4 is a single-agent model with voice and vision. Grok 4 Heavy adds multi-agent capabilities for complex task handling.

• Ranks #1 on the AA Intelligence Index (see above) and achieved 15.9% on ARC-AGI V2, nearly 2x the previous lead (Claude Opus's 8.6%).

• Grok 4 is $30/month via SuperGrok; Grok 4 Heavy is $300/month. Available via API with a 256K-token context window and built-in search.

• Delivers fast token output but is costly: ~2x OpenAI o3 per token, and usually needs 2x the tokens to complete tasks of similar complexity.

• Grok 4 is set to launch on Microsoft Azure AI Foundry and Oracle Cloud, where Grok 3 is already available for enterprise use.

Why it’s important: Despite being a newer player, xAI, backed by its Colossus supercomputer and low-human intervention training, continues on a bold trajectory. With increased enterprise access on its roadmap and GPT-5 and Gemini 3.0 nearing release, enterprise AI competition isn’t slowing down.

Capgemini released an 18-page report on Model Risk Management as generative and agentic AI introduce unprecedented opacity, and risk.

PwC shared nine AI-fueled business models leaders shouldn’t ignore as AI transforms how firms design, deliver, and scale products and services.

Salesforce published its 41-page Global AI Readiness Index, and insights on why generic LLM agents fall short in enterprise environments.

Microsoft’s CCO said AI saved over $500M in call center efficiencies last year, following the layoff of 9,000 employees last week.

Databricks shared insights from Accenture and Adobe on how they use agentic AI, from enterprise reinvention to powering data supply chains.

ABBYY published a 23-page handbook including insights on the shift from intelligent to agentic automation and services as software.

Goldman Sachs is testing Devin in a bid to advance its AI coding capabilities with Cognition’s autonomous software developer agent.

Association of Corporate Counsel released a 64-page AI guide for in-house lawyers with prompts, career tips, a maturity roadmap, and more.

Google DeepMind is paying $2.4B to license Windsurf’s AI coding tech and hire its CEO and top talent, after OpenAI’s $3B acquisition deal fell through.

Microsoft introduced an API and SDK-based version of OpenAI Deep Research within Azure AI Foundry Agent Service to support enterprise use.

OpenAI is expected to release its open-source model next week, rumored to be “similar to o3 mini” with reasoning capabilities.

AWS is launching an AI agent marketplace at its NYC Summit next week, with Anthropic as a key partner. Anthropic also released six new certifications.

OpenAI hired four senior engineers from Tesla, xAI & Meta to work on the Stargate data center and infrastructure initiatives, amid a tense talent war.

Perplexity launched Comet, an AI browser embedding search and an assistant that handles agentic tasks like booking meetings and web navigation.

OpenAI plans to launch an AI browser in the coming weeks with ChatGPT-style chat and agentic features, competing with Perplexity and Google.

AI2 introduced FlexOlmo, a new model training paradigm that enables data owners to contribute to AI development without sharing raw data.

CAREER OPPORTUNITIES

Anthropic - Google Cloud Partnership

Morgan Stanley - AI Executive Director

KPMG - AI Director

EVENTS

Gartner - AI Agents for CIOs - July 16, 2025

Snowflake - AI in Action - July 16, 2025

OpenAI - Prompting with Purpose - July 25, 2025

Complete this survey to get more value.

Here's how we can help you:

Promote your business to thousands of executives and professionals.

All the best,

Found this valuable? Share with a colleague.
Received this email from someone else? Sign up here.
Let's connect on LinkedIn.