Why 80% of Companies use AI but see no return

⏱️ Your Friday Brief (TL;DR)

Welcome back!

This week, we're diving deep into four essential reports that map out exactly how to escape pilot purgatory and build AI systems that actually move the needle.

McKinsey calls this the "gen AI paradox", where 80% of companies use AI, but just as many report zero impact on their bottom line.

We take a look at:

• How agents will transform sales

• Three system architectures for re-inventing your product

• A simple mental model to view Agents in your workforce to get the most of your transformation

It requires companies to fundamentally rethink how work gets done.

Let’s dive in 🤖

🚀 BCG's on How AI Agents Will Transform B2B Sales

The GPTLDR

Software development represents a $3 trillion market, and AI is doubling developer productivity today. Andreessen Horowitz (a16z) shared the latest on how AI is transforming software development, redefining coding workflows, and creating new agentic categories with trillions in potential economic impact.

The Details

Companies implementing agentic sales systems are seeing 1.8x margin impact through customer lifetime value growth and go-to-market efficiency. The optimal mix of human and AI depends on deal size and nature, with large strategic accounts requiring human oversight while smaller accounts can be managed autonomously by agents.

BCG breaks down the transformation into three levels:

• Augmented: AI assists with account planning and territory design for strategic accounts

• Assisted: Agents handle complex stakeholder and deal management alongside humans

• Autonomous: AI independently manages inbound demand and smaller accounts at scale

A global logistics firm achieved 30-50% efficiency gains in RFP responses. A Southeast Asian bank increased assets under management by 5-10% and boosted customer conversions four- to sixfold using agents for personalized offerings

Why It Matters

• The long tail opportunity: Finally serve accounts that were previously unprofitable.

• Speed to market: Compress sales cycles from weeks to hours.

• Human capital optimization: Redeploy your best sellers to high-value activities while AI handles the routine.

🎯 McKinsey's AI-Centric Software Frontier

The GPTLDR

According to McKinsey, 40% of software leaders expect AI to unlock more than 20% revenue growth beyond their current trajectory. But becoming AI-centric isn't about bolting AI onto existing processes.

The Details

McKinsey identifies that AI-native disruptors like Anysphere, Gamma, and Lovable are achieving product-market fit in record time and scaling to hundreds of millions in ARR with teams of fewer than 100. Meanwhile, incumbents like Salesforce and Atlassian are racing to rebuild their entire organizations around AI-centric principles.

The transformation brings compelling returns:

• 20-40% reductions in operating costs

• 12-14 percentage point increases in EBITDA margins

• Faster development cycles through AI-powered engineering

Three architectural models are emerging:

1. System-of-record agents: Add-ons to existing databases

2. Horizontal productivity agents: Domain-agnostic tools for any workflow

3. Hybrid specialists: Industry-specific agents with deep domain knowledg

Why It's Important

• Existential imperative: Software companies that don't become AI-centric risk obsolescence

• Business model disruption: The shift from seat-based to outcome-based pricing changes everything

• Competitive dynamics: AI-native startups are rewriting the rules of scale and speed

Read newsletters, not spam

Tired of newsletters vanishing into Gmail’s promotion tab — or worse, being buried under ad spam?

Proton Mail keeps your subscriptions organized without tracking or filtering tricks. No hidden tabs. No data profiling. Just the content you signed up for, delivered where you can actually read it.

Built for privacy and clarity, Proton Mail is a better inbox for newsletter lovers and information seekers alike.

Newsletters without the noise

🤖 Forrestor’s Mental Model for Agentic AI

The GPTLDR

Forrester proposes to stop thinking of AI agents as digital employees and to think of them with a dual identity: agents are both skills (capabilities you can invoke) and products (managed assets requiring governance, roadmaps, and lifecycle management). This mental shift changes everything about how you deploy and scale agents.

The Details

The "agent as employee" metaphor is seductive but dangerous. It tempts firms to view agents as drop-in replacements for people rather than catalysts for process redesign and innovation. This leads to:

• Reinforcing organizational status quo instead of transformation

• Ignoring the need for distinct capability design and system operation

• Missing the opportunity for fundamental process reinvention

Forrester's dual-identity framework requires two synchronized roadmaps:

1. Skills roadmap: Which cognitive capabilities to make available to the business

2. Product roadmap: Platform foundations for scale, governance, and evolution

Why It’s Important

• Clarity drives execution: The right mental model prevents costly missteps

• Scale requires structure: Product thinking enables enterprise-wide deployment

• Transformation, not automation: Agents should reimagine work, not just speed it up

📚 Interesting Reads

• HBR explores how companies like IBM are using generative AI to give leaders more time to focus on human-centered leadership.

• MIT Sloan explores how AI is reshaping the value of human expertise, shifting it from knowing answers to asking better questions and exercising judgment.

• Entrepreneur - True competitive advantage in the AI era comes from nurturing human traits like creativity, empathy, curiosity, and imagination.

• How Block deployed AI agents across a 12,000 person workforce in 8- weeks

• HBR uncovers how leading companies are using experimentation to get real results from Gen AI

 ➜ Until Next Week

If your AI strategy is still focused on "raising individual competencies through copilots and chatbots," you're solving yesterday's problem. The winners are embedding AI into high-value domains and letting it run wild (safely, of course).

The GPTLDR Team

AI, simplified for Decision Makers.