Hey 👋 This week it’s all about AI doing the work, not just helping with it. GPT-5.4 launched, AI agents hit the mainstream for small business, and a major new industry report has the numbers to prove AI’s ROI is no longer theoretical. Let’s get into it.
By the numbers
Big Story
GPT-5.4 Is Here — And It’s Built for Business, Not Just Chat
OpenAI launched GPT-5.4 on March 5, calling it “our most capable and efficient frontier model for professional work.” It ships with a 1-million-token context window, 33% fewer factual errors than its predecessor, and record benchmark scores in law, finance, and knowledge work. For businesses, the headline feature is Tool Search — a reworked API system that lets AI reliably chain together multi-step tasks across apps. It’s the step-change from AI as a writing assistant to AI as a workflow engine.
If you’re on ChatGPT Plus or Pro, you have access now. Try it on a real business task this week — a financial summary, a client proposal, or a research brief. The accuracy improvement alone is worth the switch.
The Revenue Race
Healthy competition between OpenAI and Anthropic is good for business users — expect pricing pressure and faster capability improvements as both fight for enterprise contracts. OpenAI is reportedly taking early steps toward a public listing as soon as late 2026.
What’s happening this week
Morgan Stanley: A Massive AI Breakthrough Is Coming — Most Businesses Aren’t Ready
In a sweeping new report, Morgan Stanley warns that a transformative leap in AI capability is imminent in H1 2026, driven by unprecedented compute accumulation at the top AI labs. The bank cited scaling laws holding firm and pointed to the belief that 10x compute translates to roughly double a model’s effective intelligence. The report’s core message: the gap between prepared and unprepared businesses is about to widen significantly.
NVIDIA’s 2026 State of AI: ROI Has Arrived — 88% of Enterprises See Revenue Gains
NVIDIA surveyed 3,200+ companies across finance, retail, healthcare, telecoms, and manufacturing. The findings are stark: 88% report AI-driven revenue increases, 87% achieved cost reductions, and 86% are increasing their AI budgets this year. Agentic AI — systems that autonomously plan and execute multi-step tasks — has moved from pilot programmes to live deployment, led by telecoms (48%) and retail (47%).
AI Agents Go Mainstream for Small Business in 2026
A wave of no-code AI agent platforms — Lindy, Make, Zapier Agents, Relevance AI — has made autonomous workflow automation accessible to teams of five or more. Independent analysis shows these tools can reduce operational overhead by 20–35% within six months when deployed on well-documented, repetitive processes. The key distinction from older AI tools: agents don’t just generate content — they execute tasks across connected apps.
Tool Spotlight
Lindy
AI Agents
Who it’s for: Small teams who want AI agents to handle email, CRM updates, scheduling, and support tasks — without writing a single line of code.
Lindy is a no-code AI agent builder that topped independent testing for small business use in 2026. You describe the workflow in plain English, connect your tools (Gmail, Slack, HubSpot, Notion, Calendly), and Lindy runs multi-step tasks end-to-end. One agent can take a new lead enquiry, log it to your CRM, draft a personalised reply, and book a discovery call — all without a human touch. Starter plans begin at $49/month with a generous free trial.
Gmail, Slack, HubSpot + 50 more
Multi-step task execution
SOC 2 & HIPAA compliant
Free trial available
Quick hits
GPT-5.1 models were retired from ChatGPT on March 11. All conversations migrated to the 5.3/5.4 family. If your team uses older model shortcuts in the API, update your integration now.
OpenAI acquired Promptfoo (March 9) — a developer tool for testing and evaluating AI prompts at scale. A quiet but telling move: OpenAI wants to own the enterprise evaluation layer, not just the model.
Samsung is targeting 800 million Gemini-powered devices by end of 2026. For businesses building mobile-first customer experiences, Google’s AI ecosystem is about to reach an enormous new audience.
Deloitte, EY, KPMG and PwC have all deployed in-house AI agent platforms for audit, tax, and consulting workflows. If the Big Four are betting on agents for knowledge work, the rest of professional services should be paying attention.
This week’s tactic
Run the “Agent Audit” — find the one task to hand off first
McKinsey research shows that businesses deploying AI automation in client-facing and admin workflows reduce operational overhead by 20–35% within six months — but only when they start with one well-documented process. This week, pick the single most repetitive task in your business (lead follow-up, meeting notes, invoice chasing — whatever stings most). Write down every step it involves. Then ask: “Could I describe this well enough for a new hire to follow?” If yes, it’s ready for an AI agent. That document is your brief. Platforms like Lindy or Make let you build it out in an afternoon.
That’s a wrap for this week. The AI doing the work theme is only going to accelerate — so if this is useful, share it with someone on your team who’s been asking “how do I actually start?” 👊
See you next Monday,
Alun · LBDLibrary
