Across our collective we’ve seen one pattern in the last 18 months. 

• An agency founder decides to lean into AI. 
• They roll out Claude and ChatGPT across the team. 
• Productivity improves. 
• They build a few automations. 
• Productivity improves again. 
• They connect the automations into a workflow. 
• Content flows faster. 
• Delivery gets tighter. 
• The founder declares victory. 
• And AI adoption is complete.

Six months later, the same founder is feeling frustrated. Because the tools work, the systems work but the margins haven't improved as much as they expected. The team is doing the same amount of work in less time, but nobody quite knows what to do with the freed-up capacity. Meetings that used to feel necessary now feel performative and roles that made sense two years ago don't map onto how the business actually runs today, but nobody has done the work of redesigning them.

Here's the Unusual model for 5 stages of AI adoption in an agency.

The team replaces Google searches with Claude prompts. They use ChatGPT to draft first versions of emails, briefs, and copy. 

This is where every agency starts, and where most agencies still are. 

In this version, individual outputs get faster but the workflow, org chart and the pricing model is unchanged. The founder is running the same operation at 1.2x speed, with the productivity gain either absorbed as slightly higher margin or invisibly transferred to clients through faster turnaround times they aren't paying more for.

The cost of stopping here is opportunity cost. The agency stays in the same commercial category it was in before AI existed but every competitor gets better at the same tools within twelve months. Any temporary advantage quickly evaporates.

The founder builds specific automations. 

• An email that reads client feedback and drafts a response
• A workflow that generates weekly content briefs from a shared drive
• A system that pulls project data into a Monday.com board without manual entry

The productivity gains at this stage are immediate and measurable. 

With Unusual Group's AI strategy, the baseline targets result in a reduction of admin time by five hours per week per person. When that target is hit consistently across a team of ten, you're recovering the equivalent of more than a full-time role every week.

An agency from our collective is a standout example. The founders implemented structured AI operations alongside a redesigned sales process and recovered fifteen hours per week each in admin time. That capacity went straight into strategy work. The commercial results followed almost immediately with a 202% year-on-year revenue growth and an increase in average project value from £18k to £42k.

In Stage 2, the point isn’t about productivity gain but more about what the recovered capacity gets redirected to. The founders who move beyond Stage 2 use the recovered capacity to change what the agency does.

This is where most agencies think they've completed AI transformation.

The founder connects the individual automations into full workflows. One decision cascades through the system. Content creation triggers image generation, which triggers social copy, which triggers email drafts, which triggers CRM updates. One move produces ten outputs.

The execution layer is minimised completely. The team is no longer doing the mechanical work of producing deliverables. They're feeding decisions into a system that produces the deliverables.

The margin uplift at this stage is significant. AI-native businesses that integrate AI deeply into their departments are seeing a 10% to 14% uplift in net profit compared to traditional, human-heavy operating models. The traditional agency structure carries too much non-billable overhead relative to billable output but AI absorbs a significant proportion of that non-billable work.

This is what we call "services on rails" operating model where manual, execution-heavy delivery gets replaced with tech-enabled workflows and defensible margins follow.

At Stage 3 the system also exposes the quality of the decisions feeding it. The workflow is producing three blog posts a week, but only one is remotely making sense. The system is producing ten pieces of client-facing collateral, but only three of them actually get approved. The founders who see this and act on it move to Stage 4.

The founder stops optimising for speed and starts optimising for decision quality.

The workflows built in Stage 3 get pulled apart and reassembled around better inputs. Decisions are influenced by data of which content actually converts, which client engagement patterns predict retention and which projects generate the strongest referrals. 

The system stops being a machine for producing output and becomes a machine for producing insight about what to output.

Agencies that differentiate by proprietary methodology produce the highest gross profit of any approach: 49%. That's significantly ahead of undifferentiated firms working from generic frameworks. 

Here’s what Stage 4 looks like in practice: the founder isn't optimising the tools, they're building a system that reflects their business's specific decision logic, then using that system to produce outputs no other agency can match at the same speed or cost.

The approval layer built in Stage 2 becomes redundant here because if the decision logic is transparent and the system shows the outcome, the founder doesn't need to sign off on every output. 

The founders in our collective who've reached Stage 4 describe it as the moment AI stopped feeling like a tool and started feeling like a business partner.

The final stage is where the money comes in.

At Stage 5, the agency stops being a service business with AI tools and becomes a platform business with a services layer. The proprietary system built during Stage 4 becomes an asset in its own right and the methodology becomes IP. The data accumulated through delivery becomes a moat. The system that started as an internal productivity tool becomes something clients might eventually license directly.

The founders who complete Stage 5 build businesses that look nothing like their pre-AI versions. They have different pricing models, different team structures, different meeting rhythms, different relationships with clients and most importantly, different valuations. 

According to our 2025-2026 PE Playbook, traditional project-heavy creative shops trade at 7-8x EBITDA multiples. But tech-enabled agencies with defensible IP and proprietary data command 12x+ multiples. That gap is the difference between an exit that funds the founder's next chapter and an exit that funds a life-changing outcome.

Acquirers are actively looking for these hybrid models right now and they are paying premium multiples for agencies that have completed the organisational redesign because those agencies operate as tech-enabled platforms, not as service shops.

Most agencies come to us stuck between Stage 2 and Stage 3.

Where, they've built individual automations which have led to productivity gains. They know something bigger is possible but they can't quite see the path from where they are to where they need to be. 

What they're missing is a structural framework. The Unusual Method addresses this at each stage. 

• In Stage 1, we help founders see which workflows are candidates for structural change versus which are just tool substitution. 
• In Stage 2, we push founders to build systems with the end-state architecture in mind. 
• In Stage 3, we help founders to start looking at what the system is revealing about the business. 
• In Stage 4, we bring the pattern recognition from working across multiple agencies simultaneously, so the founder doesn't have to rebuild the decision architecture from scratch. 
• In Stage 5, we do the hard organisational and commercial redesign work alongside the founder, because that stage requires substantial senior input.

The window to make this transition is open now but it won't be open indefinitely. 

If you've recognised your own agency somewhere in this map, we can help you through this transformation. Book a call with us and we'll walk through what the specific next steps look like for you.