The Frontier Firm : Introducing 'Formula'
Monday 20th April 2026
There’s a moment in most ambitious organisations where the conversation about AI shifts.
It moves away from experimentation and toward something more serious. Less about tools, more about operating models. Less about pilots, more about permanence.
Recently, we began working with a business we’ll call Formula.
Formula is not a specific customer. It’s a composite of the kinds of mid-market and enterprise organisations we work with every day. It runs on Microsoft. Azure underpins its infrastructure. Microsoft 365 powers collaboration. Dynamics 365 manages customer and operational workflows. Power BI dashboards are widely used. Copilot is being explored. The building blocks are there.
And yet, something doesn’t quite connect.
When “Modern” Isn’t the Same as “Structured”
From the outside, Formula looks digitally mature. But under the surface, familiar friction appears.
Data lives across environments and teams. Security policies are active but inconsistently applied. Copilot has been piloted without a unified governance model. Power Automate flows have solved local problems but haven’t always aligned to broader architecture.
Microsoft Fabric has been discussed, but analytics still sit in pockets rather than forming a true data backbone.
None of this signals failure. It signals growth.
Most organisations don’t struggle with ambition. They struggle with sequencing.
AI isn’t the issue. The order in which it’s adopted is.
That’s why we think about this journey in three deliberate stages: Hop. Skip. Jump.
Not as a slogan, but as a structure.
Hop: Control Before Acceleration
When Formula first approached us, the instinct was to accelerate innovation. There was interest in Azure AI Foundry. Curiosity about agent-based orchestration. Early Azure Machine Learning models were being tested. The appetite was there.
But before expanding, we stepped back.
Hop is about control.
We examined their Azure estate, not just consumption, but architecture, resilience and cost visibility. Identity posture was reviewed through Microsoft Entra: roles, permissions, conditional access policies and access sprawl. Microsoft Purview classification and data lifecycle controls were revisited to ensure that future AI workloads would sit on governed data. Defender for Cloud posture was assessed alongside Sentinel visibility to understand where expanding AI could introduce new exposure.
Power BI reporting was rationalised. Fragmented datasets were consolidated. Conversations about analytics shifted toward Microsoft Fabric as a centralised data layer rather than an optional enhancement.
This wasn’t glamorous work. It rarely is.
But AI layered onto poorly governed data amplifies inconsistency. AI deployed without identity discipline increases exposure. Copilot introduced without structured governance creates uncertainty.
Hop is about clarity over the platform before acceleration on top of it.
Skip: Embedding Capability into the Operating Model
Once foundations were strengthened, Formula’s use of Microsoft technology began to change in character.
Skip is where capability becomes embedded.
Copilot was no longer experimental; it was deployed deliberately across Microsoft 365, aligned to specific roles and measurable productivity goals. Copilot for Sales integrated directly with Dynamics 365 data to create context-aware engagement rather than generic assistance.
Power Automate workflows were redesigned to remove repetitive manual processes across finance and operations at scale. Fabric became the backbone for unified analytics, connecting operational data from Dynamics 365 Supply Chain Management with customer insight data and reporting layers in Power BI.
Azure Machine Learning models evolved from proofs of concept to structured forecasting tools that informed demand planning and inventory decisions. Security evolved alongside capability, Defender and Sentinel policies adapted to expanding workloads rather than reacting after the fact.
At this stage, AI wasn’t disruptive. It was embedded.
The difference was subtle but significant. Innovation and governance were no longer competing priorities; they were integrated into a single operating model.
Jump: Optimisation as Direction of Travel
Formula hasn’t “arrived” at some final AI destination. No organisation has.
Jump isn’t about arrival. It’s about trajectory.
The conversations at executive level are now different.
Instead of asking where AI can be trialled, leadership is asking where it should influence decisions.
- How might predictive modelling in Azure Machine Learning inform margin management?
- Can demand forecasting models dynamically adjust supply chain planning within Dynamics 365?
- Where could Azure AI Foundry Agents orchestrate workflows that currently rely on manual coordination?
- How can insights generated in Fabric flow directly into operational action through automated triggers?
These are operating model questions, not technology questions.
They only emerge when the platform underneath is strong enough to support them.
Jump is where Azure, Fabric, Machine Learning, Dynamics and Microsoft 365 stop being separate conversations and start functioning as a connected system.
Why We’re Telling This Story
There is no real Formula headquarters. No product line. No board of directors. We created the story deliberately purely to demonstrate our skills in AI-first transformation.
Because the structural challenges Formula faces are real. We see them across industries and across organisations already invested in Microsoft.
- Azure is in place, but not fully optimised.
- Copilot is being explored, but governance is uneven.
- Fabric is discussed, but not yet central.
- Security is active but evolving.
Hop. Skip. Jump gives structure to that progression.
Control precedes capability. Capability precedes optimisation.
Organisations that approach AI with this discipline are known as Frontier Firms. Not because they are futuristic, but because they evolve deliberately.
In the coming months, we’ll share further chapters of Formula’s journey, focusing on specific layers of the Microsoft stack and what structured AI progression looks like in practice.
But if there’s one idea to take from this first chapter:
AI does not create competitive advantage on its own.
Architecture does. Governance does. Sequencing does.
Once those are in place, AI stops being an experiment.
It becomes inevitable.
Want to know more? Visit our page for more information about becoming a Frontier Firm Now
Or contact the team at [email protected] to accelerate your journey to AI readiness
Want to keep informed? Sign up to our Newsletter