With several years of experience as a project manager in the financial sector, I have worked extensively on automating operational tasks through IT systems. In this setting, process optimisation and Straight Through Processing (STP) have been well-established principles for creating efficient, reliable, and scalable workflows. While process optimisation defines and improves the structure of work, STP realises its value through end-to-end automation.

I find it particularly encouraging that widely accessible AI technologies are now available. They offer a practical opportunity to accelerate process optimisation and extend STP principles into areas such as the public sector, where automation has traditionally been more challenging.

The concepts

In line with the digital transformation and increasing complexity of information systems, organisations’ ability to streamline and automate workflows has become a critical success factor. Two key concepts in this context are STP and process optimisation. While process optimisation targets the systematic analysis and improvement of existing workflows, STP focuses on realising a completely automated processing of transactions without human intervention. Taken together, they are fundamental mechanisms to support organisational efficiency, quality, and scalability in IT operations.

Straight Through Processing as an automation paradigm

STP represents an advanced level of digital process automation maturity. Eliminating manual checks and entries not only minimises the risk of human error but also significantly shortens lead times.  The literature highlights that STP is particularly relevant in domains where transaction volumes are high, such as in financial services, supply chain administration, and IT service management. In information technology, STP refers to the ability to complete routine tasks, such as creating users, integrating systems, or transferring data, automatically from start to finish, without any need for manual intervention.

Process optimisation as an organisational discipline.

Process optimisation represents a methodical approach to reducing complexity, identifying bottlenecks, and ensuring better resource utilisation. Using techniques such as Lean, Six Sigma or Business Process Management (BPM), organisations can create more rational workflows where redundancy and waste are minimised. Within IT organisations, process optimisation often means a move from silo-based workflows to more holistic and value-chain-focused approaches. A good example is DevOps, which combines development and operations teams to produce deliverables more quickly and reliably.

The interaction between STP and process optimisation

Although optimising processes is often necessary for successful STP, it is STP itself that acts as a catalyst, unlocking the complete advantages of these improved processes. To automate a workflow from start to finish, it must be clearly outlined, standardised, and stripped of any unnecessary differences.  The interaction between the two disciplines thus creates a synergy effect, where the organisation not only achieves efficiency, but also robustness and compliance.

Benefits and strategic importance

Implementing STP and process optimisation offers several key benefits:

Efficiency and speed: Tasks are finished more quickly, which boosts overall productivity.

• Quality and consistency: Reduced errors and deviations create higher data integrity.
• Cost reduction: Automation frees up resources for more value-adding activities.
• Compliance and traceability: Standardised processes ensure better documentation and regulatory compliance.
• Scalability: The organisation can handle growth without proportionally increasing human resources.

At a strategic level, STP and process optimisation support the organisation’s ability to adapt to changing market conditions and technological developments. They are thus not only operational tools, but essential elements of the digital transformation agenda.

The news: “AI as an accelerator for process optimisation and STP”

Artificial intelligence represents an important next step in the development of process optimisation and STP. Where traditional automations are based on fixed rules and standardised workflows, AI adds an adaptive dimension that enables the systems to learn and improve continuously. Using machine learning, AI can identify patterns in operational data, predict bottlenecks, and suggest optimisations in real time.

In addition, in the context of STP, AI enables exceptions and unstructured inputs to be handled automatically, expanding the potential for end-to-end automation. At the same time, technologies such as Natural Language Processing and predictive analytics can support automated case management, customer dialogue, and capacity management. The result is a more agile and data-driven operating environment, where efficiency, quality and scalability are lifted to a new level.

Why STP and process optimisation are more straightforward in banking than in the public sector.

The banking industry has long been a natural environment for the introduction of process optimisation and STP. Core banking processes are typically transaction-oriented, high-volume, and governed by clearly defined business rules. Financial transactions such as payments, settlements, and account updates are based on structured data, standardised formats, and deterministic decision logic. Moreover, strong competitive pressure and direct cost–income incentives have historically driven banks to invest heavily in automation and end-to-end process integration. As a result, the organisational, technological, and economic conditions required for STP are often already in place.

In contrast, the public sector operates under a fundamentally different set of constraints. Public-sector processes are frequently shaped by complex legislation, political objectives, and legal requirements for due process, transparency, and equal treatment. Case handling often involves discretionary judgment, individual assessments, and unstructured information submitted by citizens, which limits the degree to which workflows can be fully standardised. Additionally, public-sector organisations are typically risk-averse, as errors in automated decision-making may have legal, ethical, and societal consequences. These factors make the introduction of STP and aggressive process optimisation more challenging than in banking, even where digital maturity is high.

Nevertheless, advances in digitalisation and artificial intelligence are gradually narrowing this gap. By enabling risk-based automation, intelligent handling of unstructured inputs, and selective straight-through processing, AI allows public-sector organisations to adopt STP principles in a controlled and compliant manner. This suggests that while the path to STP is more complex in the public sector, it is increasingly feasible when supported by appropriate governance and technology.

Four Illustrative examples of AI as an accelerator for process optimisation and STP in the Danish public sector

1) AI-enabled straight-through processing in public benefits administration

In Danish public administration, the automation of benefits processing provides a clear illustration of how artificial intelligence can accelerate the interaction between process optimisation and STP. Organisations like Udbetaling Danmark process many benefit applications through standardised digital workflows. AI-supported routing and validation mechanisms are used to distinguish between routine and non-routine cases based on predefined eligibility criteria and learned risk patterns.

AI helps streamline processes by automatically handling straightforward, low-risk cases, which reduces unnecessary manual differences and improves efficiency.  Human intervention is reserved for cases that deviate from expected patterns. This approach increases overall process throughput while maintaining compliance and legal accountability, thereby demonstrating how AI can function as an enabler of STP without undermining governance requirements.

2) Risk-based automation in tax administration

Another prominent example can be found in tax administration, where AI is applied to optimise control-intensive processes. Within the Danish Tax Agency (Skattestyrelsen), machine-learning models analyse tax declarations and VAT submissions to identify anomalies indicative of potential non-compliance. Instead of subjecting all cases to uniform manual review, AI enables a differentiated, risk-based approach.

This represents a significant shift in process design. By optimising control activities and targeting them selectively, a substantial proportion of compliant cases can be approved automatically, effectively following a straight-through processing logic. The example illustrates how AI contributes to both efficiency and effectiveness: operational resources are concentrated on high-risk cases, while low-risk cases flow through the system with minimal friction.

3) Intelligent document processing as a foundation for STP

Document-intensive administrative processes have traditionally posed a barrier to STP due to unstructured inputs and data quality issues. Several Danish municipalities now use AI solutions with optical character recognition and natural language processing for document processing.  Incoming documents such as invoices, applications, and citizen correspondence are automatically classified, and relevant data elements are extracted and validated.

In terms of process optimisation, this reduces manual preprocessing activities and standardises the quality of input data. As a result, downstream workflows become more predictable and better suited for automation. AI thus functions as a foundational enabler of STP by transforming unstructured information into structured inputs that can be processed end-to-end by IT systems.

4) Selective straight-through processing in healthcare administration

The application of AI in healthcare administration further illustrates the concept of selective STP. In Danish regions such as Region Hovedstaden, AI-based decision-support systems are used to assist in the triage of referrals and diagnostic workflows, for example, in radiology. Algorithms pre-screen cases and identify those that are likely to be routine versus those requiring specialist attention.

While clinical decision-making remains the responsibility of healthcare professionals, administrative and preparatory steps can increasingly be handled automatically for routine cases. From a process optimisation standpoint, this reduces administrative bottlenecks and improves flow efficiency. Although full STP is neither feasible nor appropriate in all clinical contexts, AI enables a partial and controlled form of straight-through processing that improves overall system performance.

Synthesis: AI as a mediator between optimisation and automation

Across these examples, AI emerges as a mediating technology that connects process optimisation with operational automation. Traditional STP initiatives depend on strict standardisation and deterministic rules, which limit their applicability in environments characterised by complexity and variability. AI extends the reach of STP by enabling adaptive decision-making, probabilistic assessment, and automated handling of exceptions.

Consequently, AI should not be understood as a replacement for established process optimisation disciplines, but rather as an accelerator that enhances their effectiveness. By supporting risk-based routing, intelligent validation, and dynamic control mechanisms, AI enables organisations to achieve higher levels of automation while preserving transparency, compliance, and organisational control.