When we talk to B2B commercial leaders about AI, one of the most common sources of confusion is the gap between what they're already doing with AI and what's actually possible. A company where the sales team uses ChatGPT informally thinks it's "doing AI." A company that has just deployed an agentic commercial intelligence system thinks it's still "early in the journey." Neither is quite right.
To cut through the noise, we use a simple four-level framework that maps where organisations actually are in their AI adoption journey — and what the realistic next step looks like from each position.
Individuals across the organisation are using AI tools — ChatGPT, Claude, Copilot — in their personal workflows. This usage is largely informal, ungoverned, and disconnected from company systems and data. There may be informal sharing of prompts and techniques, but no systematic approach to what AI is used for, by whom, or with what safeguards.
Signals you're here
- AI tools are expensed individually or used on free plans
- No AI governance policy exists
- AI usage is personal productivity, not commercial process
- Senior leaders talk about "exploring AI" without clear programmes
AI is being integrated into specific, defined workflows — generating first drafts of proposals, summarising CRM notes, automating routine reporting, flagging invoice exceptions. These are typically point solutions: AI applied to a specific task with a defined input and output. There is increasing governance around what AI tools are approved and how they're used with company data.
Signals you're here
- Licensed AI tools deployed for specific use cases
- Some AI governance/data handling policies in place
- Measurable time savings in specific workflows
- IT/digital teams actively involved in AI tooling decisions
Autonomous AI agents are operating continuously across company data and systems — not just responding to prompts, but proactively monitoring, analysing, and surfacing intelligence. These agents weave together multiple data sources, operate with defined goals, and take actions or make recommendations without requiring a human to initiate each interaction. This is where Cai operates.
Signals you're here
- AI operates 24/7, not just when prompted
- Agents have access to multiple connected systems
- Commercial decisions are routinely informed by AI-generated intelligence
- Significant measurable commercial outcomes attributable to AI
The organisation is fundamentally structured around AI capabilities. Hundreds or thousands of specialised agents run core business functions — commercial, operations, finance, supply chain — with human leadership focused on strategy, judgement, and governance of the AI system rather than managing traditional processes. This is not science fiction, but it is genuinely early-stage at the time of writing.
Signals you're approaching here
- Org design explicitly accounts for AI agent capabilities
- Hiring strategies include "AI agent management" skills
- Core business processes rebuilt around AI, not retrofitted to it
- Leadership thinking about governance of AI systems, not just tools
Why the levels matter for B2B commercial
The practical implication of this framework isn't about pride of position — it's about sequencing investment correctly. The trap we see B2B companies fall into is trying to jump levels: a Level 1 organisation that decides to buy an agentic AI platform before it has the data connections, governance, and internal capability to operate it effectively.
The transition from Level 2 to Level 3 — from workflow automation to agentic AI — is the biggest value-creating step available to most established B2B companies right now. It's also the step where most mid-market companies are sitting, uncertain about what "agentic AI" actually means for them practically.
The honest answer is that moving from Level 2 to Level 3 doesn't require a technology transformation. It requires an intelligence layer — built on your existing data, connected to your existing systems — that operates continuously rather than on demand.
That is, precisely, what Heyxio's agents are designed to provide. Not a rip-and-replace of your technology stack. An intelligence layer on top of what you already have, starting with your commercial organisation.
How to use this framework
Use these four levels as a diagnostic, not a trophy case. The question is not "which level are we at?" but "what does moving to the next level actually require, and what's the highest-value outcome we can get from that transition?"
For most B2B companies in the £200m–£2bn revenue range reading this: you are at Level 1 or Level 2. The move to Level 3 is available to you now, with the data and systems you already have. The gap is not technology. It is the intelligence layer that makes your commercial data work for your commercial team — in real time, continuously, proactively.
That's the move worth making.