Strategic briefing, April 2026
The Messy Middle
AI, your business, and what to do right now. A practical guide to the transition phase every company is living through.
Last updated: April 2026
AI, your business, and what to do right now. A practical guide to the transition phase every company is living through.
Last updated: April 2026
Every week, a frontier AI lab ships a model that makes last month's model look quaint. Anthropic grew from zero to over a billion dollars in annual revenue in under three years. OpenAI reached 400 million weekly users. Google, Meta, and xAI are spending tens of billions on infrastructure. The velocity is unlike anything the technology industry has seen, and it is dragging every other industry along with it.
If you run a business, you feel it. Your LinkedIn feed is AI. Your competitors are talking about AI. Your board is asking about AI. Your team is either excited or terrified, sometimes both. And yet, when you look at your actual operations, your CRM is the same CRM. Your HR system is the same HR system. Your accounting software still requires the same manual data entry it did two years ago.
You are in the messy middle.
The messy middle is the gap between what AI can do and what your business is actually doing with it. Your tools work. They are not broken. But they are not connected, not intelligent, and not ready for what is coming next.
This is not a technology problem. It is a timing problem. The AI-native versions of every tool you use are being built right now. Salesforce is embedding AI across its platform. HubSpot has AI assistants. Xero has automated categorisation. But none of them are finished. None of them are mature enough to bet your business on. And the pure-play AI alternatives, the startups promising to replace everything, most of them will not survive the next 18 months.
So what do you do? You cannot wait, because your competitors are not waiting. You cannot rip and replace, because your data, your workflows, your team's muscle memory are all embedded in the tools you have. The answer is in the middle: build an intelligent layer on top of what you already have.
Understanding where you are is the first step to deciding what to do.
Your CRM has years of customer data. Your accounting system knows your chart of accounts. Your team knows where everything is. These tools are not broken. Replacing them carries real risk: data migration, retraining, downtime, lost institutional knowledge.
AI-native CRMs, AI accounting platforms, AI customer support, AI HR. Every category has multiple challengers. Some will survive. Most will not. Choosing the wrong one means migrating again in 12 months. The market has not settled.
Will Salesforce's AI be good enough, or will a startup eat their lunch? Will your industry's tools rebuild in time, or will a horizontal AI platform absorb them? The honest answer: nobody knows. The winners will not be clear for another 12 to 18 months.
A snapshot of how fast the AI landscape is moving, and why your window for action is narrowing.
Mid 2024
GPT-4o goes multimodal. Claude 3.5 Sonnet becomes the developer favourite. Cursor goes from novelty to essential. Meta open-sources Llama 3, competitive with GPT-4 for the first time.
Late 2024
OpenAI ships o1, the first reasoning model. DeepSeek R1 from China matches it at a fraction of the cost, wiping USD 600 billion from Nvidia's market cap in a single day. Google ships Gemini 2.0. API pricing drops 90% in 18 months. AI stops being expensive.
Early 2025
Agentic AI arrives. Claude can use tools, browse the web, execute code autonomously. Cursor hits USD 100 million ARR from near zero in 12 months. McKinsey reports 72% of organisations using AI, but only 10% at production scale. The gap is massive.
2025 into 2026
Anthropic crosses USD 1 billion ARR, growing from USD 200 million in a single year. Claude Code, OpenAI Codex launch. Shopify's CEO mandates AI as a baseline expectation for all employees. The conversation shifts from "should we use AI" to "how do we integrate it with what we already have."
Now: April 2026
You are here. The tools are ready. The question is whether your business is.
Navigating the messy middle is not about choosing the right AI tool. It is about making the right decisions at the right time.
Connect your existing systems with an AI layer on top. One command centre. Your team keeps using what they know. Your business gets smarter.
Read more
Not every AI opportunity needs action today. A practical framework for deciding what to automate now, what to buy off the shelf, and what to leave alone.
Read more
The people using your systems are not the people who want to change them. How to bridge the gap between ambition at the top and anxiety on the ground.
Read moreThe messy middle is uncomfortable, but it is also where the advantage is built. While most businesses are frozen, either overwhelmed by the pace or paralysed by the choice, the ones that act now are quietly building something their competitors will spend the next three years trying to catch.
Not by betting on one tool or one platform. By building the connective tissue: the integration layer, the data flows, the AI-augmented workflows that make their existing operations faster, cheaper, and smarter. When the dust settles and the winning platforms emerge, these businesses will already have the architecture in place to plug them in.
You need someone who lives in this chaos every day and can help you make sense of it. p0stman builds the integration layer, the AI agents, and the connected workflows that bridge where you are and where you need to be.
Written by Paul Gosnell, founder of p0stman. 20 years building digital products, now building with AI every day. p0stman is an AI-native product studio that helps businesses navigate the messy middle: connecting existing systems, building AI agents, and shipping production-ready platforms. We have built for healthcare, finance, hospitality, manufacturing, and more.
Read more about Paul