Ambient ERP – breaking free of the enterprise software triad

This is a guest blogpost by Claus Jepsen, Chief Technology Officer, Unit4.

Imagine a future where your back-office applications are context-aware, operating in the background, only surfacing what matters when it matters. Thanks to progress being made in artificial intelligence (AI) this is no longer a fanciful dream. It requires a complete shift in the mental model for delivering business applications, from viewing it as software that humans are heavily involved in operating to tools that either act for us autonomously (within parameters designated by us) or support us by providing relevant analysis and context in real-time. Essentially, this means automating what can be automated and augmenting human judgement in other situations. For me, this is the best marker of when software becomes pervasive. It does not disappear entirely from the user’s orbit, but it reduces how often they must interact with it. A shift from invasive to ambient business applications.

With the arrival of more robust AI tools like large language models (LLMs) and machine learning, we will move on from the rigid constraints of existing enterprise software to more adaptive, event-driven interactions based on the user’s context, rather than a need to follow a sequence of instructions. There will likely still be a need for robust rules and structures operating in the background to safeguard data fidelity and transaction integrity, but the visible rigidity in current enterprise software will dissolve and users will be guided not constrained. With AI embedded, systems will reduce the cognitive load on users by providing the right insights at the right time to improve decision-making and automate other tasks to relieve users of digital bureaucracy.

Escaping the triad: flowcharts, databases and user interfaces

Without intending to down-talk existing enterprise applications, the reality is that innovation is redefining human-to-machine interaction. Back when enterprise resource planning (ERP) systems first emerged in the 1970s it was a significant revolution in how core back-office functions like HR and finance were run. Paper-based processes were turned into flowcharts automating how business processes worked. The challenge is that this created rules which had to be adhered to, leaving users imprisoned within the rigidity of these same flowcharts.

Consequently, interactions with enterprise software became dominated by a tightly knit triad which impacted user productivity, positively and negatively: flowcharts dictated processes, databases shaped what data we saw and user interfaces (UIs) enforced structure and workflow. This meant the user’s experience was predefined by systems designed for data accuracy and compliance, not real-world flexibility. While it did give structure to complex processes, such as supplier invoice management, it became an inflexible, stepwise interaction that reflected the requirements of the application’s architecture rather than the human need. Fundamentally, users were being stripped of their agency and reduced to transactional processing entities.

The introduction of AI sees enterprise software move from enforcing specific structures in business processes to enabling flow, from limiting users to predefined workflows to allowing them to respond dynamically to events. Ultimately, this will empower us with software that is more relevant and purpose-driven. There are several technical changes at the heart of this transformation:

• The “databasification” of enterprise software: databases are traditionally systems of record hardcoded with business logic that takes users through predefined screens and database tables. Traditional enterprise software is very quickly heading in the same direction. In one sense standardisation is good, but it also commoditising the value of business applications. Innovation is disrupting this situation. Instead of the typical record-based view of information, AI-enabled systems will replace this rigid view with semantic context, made possible through ontologies, state models and defined relationships between information, what is called the meta-information layer. As a result, business logic will not be buried in screens and interfaces, but made accessible as computational tools that algorithms can use to learn and optimise. AI Agents will follow instructions based on execution logic derived from detected patterns in data and statistical correlations.
• Data is the foundation of context-driven computation and algorithmic execution: in this AI-imbued world, the importance of data must be elevated. It must be organised not just by schemas, but by purpose, relationships and state. By introducing these categories it will be possible to construct algorithms that interpret data in context, understanding not just what it is by why it matters and how it connects to the broader flow of work. The goal is to create a foundation that decouples data from interfaces and workflows allowing algorithms to draw connections, track transitions and operate with context-driven logic.
• The rise of predictive algorithms: these will be important to enable to software to “learn” when to anticipate, recommend and act. For the user this will mean relevant information is surfaced when it is needed, or the AI Agent completing routine tasks in the background with requiring the user to navigate complicated menus. It will mean the burden of knowing how to use the system will shift from the user to the software itself.

Enterprise software should support its users, not ask them to support it. And it should extend human capability, not consume it. As AI tools mature, we will see a significant shift in how we view such business applications. It will create a new type of infrastructure supporting organisations that will affect how decisions are framed, priorities are surfaced and operational knowledge is applied. Infused with AI, these applications will become the scaffolding for business operations. They will no longer be place where you must go to process invoices and HR requests. They will be the pervasive engine recording changes, interpreting data, providing insights and prompting actions based on real-time analysis of all the information in the organisation. Are you ready for the era of Ambient ERP?