Most industrial companies would say they have already completed digital transformation. They run modern ERP systems, operate automated shopfloors, use IoT platforms and customer portals, and manage performance through dashboards rather than paper reports.
Yet when something critical happens – a supply disruption, a sudden demand spike, or a quality issue at a key customer – the response still looks very analogue: emergency calls, war rooms, spreadsheets, and a small group of experts who “know how things really work.” The enterprise is digitally enabled, but not truly steerable as one coherent system.
To move beyond this, a clearer progression is needed:
Digitally Enabled Enterprise → Software Defined Enterprise → Autonomous Business
1. The limits of the Digitally Enabled Enterprise
In a digitally enabled enterprise, most building blocks are in place – but they do not yet form a programmable whole. Typical patterns include:
Digital islands, manual glue
CRM, ERP, MES, WMS and service tools all exist, yet end to end flows like lead to cash or plan to produce are still held together by emails, Excel sheets and individual heroics.
Automation without autonomy
Robots, PLCs and automated warehouses handle repetitive work, but follow fixed programs. When conditions change, humans still need to intervene, re plan and re route.
Reporting instead of real time control
Dashboards show what happened yesterday or in the last hour. Turning insights into action still depends on meetings and manual coordination.
Digital transformation has made operations faster and more transparent, but it has not fundamentally changed how the enterprise is steered. This is where the Software Defined Enterprise becomes essential.
2. Software Defined Enterprise: making value creation programmable
A Software Defined Enterprise (SDE) is not just “more digital.” It is an organisation whose value streams are explicitly modelled so they can be changed and controlled directly in software. Three characteristics define it.
Programmable operations instead of hard wired flows
Today, many processes are physically and organisationally locked in: production lines built for specific products, workflows hard coded in systems, exceptions handled through workarounds.
In an SDE:
- Control logic shifts from isolated controllers to configurable edge and automation platforms.
- Production and logistics flows are modelled as parameterised workflows that can be adapted quickly.
- Changes that once took months become configuration or software updates.
Business capabilities as reusable services
Rather than duplicating logic across systems and regions, an SDE defines core capabilities as services:
- Capabilities such as availability checking, promise to deliver, scheduling, pricing or service dispatch are encapsulated as clearly defined services.
- These services are exposed via APIs and orchestrated into different end to end journeys.
The value chain becomes a set of modular building blocks, not monolithic structures.
Event driven value flows
An SDE is supported by an event backbone:
- Events like “order changed,” “supplier delay,” “machine down” or “shipment at risk” are published across the enterprise.
- Other systems subscribe and react immediately – adjusting plans, informing customers or triggering service actions without waiting for batch jobs.
As a result, value creation is no longer just digitally documented; it becomes software defined, explicit and steerable.
3. Autonomous Business: when agents run the value chain
Once the enterprise is software defined, a new question becomes practical: what if parts of the value chain could run themselves?
That is the promise of the Autonomous Business – systems that do not just automate tasks, but also plan, decide and act within clear guardrails.
- From task automation to goal driven optimisation
Agents optimise for business goals such as service level, margin, CO₂ or risk, rather than just executing predefined steps. - From silo optimisation to end to end orchestration
A supplier disruption no longer triggers isolated fixes in purchasing. An agent can coordinate demand, production, logistics and customer communication across the chain. - From static rules to learning systems
Decisions improve over time as models and policies are updated with new data.
In practice, this includes:
- Autonomous maintenance that predicts issues, schedules interventions and orders parts, reducing unplanned downtime.
- Self optimising production and intralogistics that re plan within seconds when conditions change.
- Proactive service that selects the best mix of remote resolution, technician dispatch or planned retrofit.
Autonomy does not mean removing humans. It means delegating well defined operational decisions to governed systems, while people focus on purpose, priorities and exceptions.
4. A deliberate journey, not a leap of faith
Seen this way, the path is clear:
- Digitally Enabled Enterprise – digital tools on largely traditional structures.
- Software Defined Enterprise – value streams modelled as services and events.
- Autonomous Business – intelligent agents operating on top of that software defined core.
The challenge for leaders is to design this progression consciously.
First, it requires honest assessment. A simple test is to examine one value stream, such as order to cash:
- How many handoffs are still managed by email or Excel?
- How often do we rely on individuals who “know how it really works”?
- How many parallel versions of the same logic exist?
If manual glue and duplication dominate, the organisation is still in the digitally enabled stage.
Second, SDE must become an architectural north star. Key initiatives should be measured against a few questions:
- Does this make the process more modular and programmable?
- Does it reuse shared capabilities and events instead of creating new variants?
- Can future AI agents realistically operate on this layer?
Third, autonomy should be introduced step by step, with strong governance:
- Start with narrow, well understood decisions where impact and risk are manageable.
- Define clear guardrails for autonomous action versus human approval.
- Instrument decisions, outcomes and overrides to enable continuous improvement.
This journey also requires cultural change. Leaders need to actively involve and empower employees so AI driven processes are trusted rather than resisted.
Over time, this journey builds confidence in data, algorithms and governance. Autonomous business then becomes not about replacing people, but about removing low value firefighting so teams can focus on design, innovation and exception handling.
Ultimately, the companies that succeed will not just have more digital tools or more AI. They will own a value chain that can be shaped, steered and improved almost as easily as software – and that is the real foundation of an autonomous business.
As a result, value creation is no longer just digitally documented; it becomes software defined, explicit and steerable.
