From applications to outcomes: How AI is reshaping indirect tax compliance

Driving the shift away from the applications era

The applications era is not ending because tax teams are underperforming. It is ending because the compliance environment has moved beyond what it was designed to support. Governments are increasingly requiring real-time data submission through mechanisms such as e-invoicing. Across LATAM and Europe, these mandates have given tax authorities real-time visibility into business activity that previously arrived only through periodic filings. Authorities are now pre-filling returns using that data, often incompletely, requiring teams to reconcile their own filings against authority-generated prefills. Real-time submissions and periodic filings must be kept in alignment, a cross-functional task that historically sat in separate silos.

Humans move data, reconcile discrepancies, confirm alignment, and keep the workflow progressing. As the compliance environment becomes more continuous and interconnected, that manual coordination model becomes harder to sustain without adding cost, delay, and risk.

The outcomes era

As regulatory demands and data complexity increase, a different model has taken hold.

In this model, the role of software shifts from tools that professionals operate to systems that deliver results.

Rather than executing steps across multiple applications, tax teams engage at the outcome level. They review system-generated results rather than assembling them manually. AI agents execute workflows across the indirect tax lifecycle, pulling data, applying rules, reconciling outputs, and preparing filings without requiring additional human coordination.

From toolbox to outcome: How work gets done differently
In the applications era, tax professionals worked directly within applications, much like a builder working tool by tool, carrying results from one stage to the next. In the outcomes era, those tools still exist, but professionals no longer use them directly. Instead, an intelligent system draws from the full toolbox, selecting and sequencing the right tools to deliver the result. The professional's role shifts from builder to inspector, focused on ensuring the structure is sound rather than laying every brick.

The human in the loop

Indirect tax compliance is inherently an outcome-driven problem. This is reflected in how modern solutions are being designed.

A unified, end-to-end platform can allow AI agents to orchestrate the full lifecycle of tax determination, e-invoicing, and compliance. These systems deliver results that tax professionals review, with human involvement triggered only at defined control points.

This approach aims to make the process of producing review-ready compliance outputs automated by default. AI agents perform work on behalf of professionals, while human-in-the-loop workflows ensure outputs are reviewed, validated, corrected, and approved as needed.

This doesn’t reduce the importance of human expertise. It elevates it. It shifts it to a different layer of the workflow, where professionals focus on judgment, validation, and accountability rather than execution.

FROM APPLICATIONS TO OUTCOMES
Applications era: Humans complete work in systems. Outcomes era: Systems complete work, humans validate outcomes.

What this transition looks like in practice

The platform foundation

The shift from applications to outcomes is achievable when the system can operate across the entire workflow, not just within its individual parts alone.

In organizations where the shift is underway, these processes operate as part of a continuous end-to-end workflow rather than as separate steps across disconnected systems. Data flows across each stage without requiring manual transfer or reconciliation.

Reconciliation becomes embedded in the workflow rather than layered on top of it.

The AI differentiation

This shift is not driven by AI alone, but by how AI is applied within the platform.

In indirect tax compliance, accuracy, repeatability, and auditability are essential. Outputs must be consistent, traceable, and aligned with regulatory requirements across jurisdictions. “It’s not enough to get the right answer; you have to show the steps you took to get there,” notes Alex Bunnett, senior product manager with Thomson Reuters.

What distinguishes the platforms where this shift is working is how the AI is grounded. Rather than applying generic models to tax problems, these systems operate within expert-designed frameworks built on established tax workflows, applying domain-specific logic encoded over decades of compliance practice. This approach can be described as domain-founded agents — AI systems guided by expert-defined rules and processes, operating within clearly defined boundaries that give tax teams confidence in the outcomes they are validating.

Thomson Reuters ONESOURCE Indirect Tax is one such platform, bringing together tax determination, e-invoicing, and compliance in a single integrated system built on decades of compliance expertise.

From AI processes to repeatable execution

To address the need for repeatability, AI-generated processes are translated into structured, auditable templates that execute programmatically.

ON TRACK: HOW AI DELIVERS REPEATABLE, CONTROLLED EXECUTION
Think of it as a railroad track. The system moves through defined checkpoints in a consistent order — data validation, reconciliation, review, and approval. There is controlled flexibility at each stop, but the overall route is fixed. Once established, the workflow runs the same way every period, and the audit trail shows every station visited along the way.

"It's not just a freehand experience," says Bunnett. "Once you've defined what that looks like, it runs on a repeatable basis."

In effect, the system combines the scalability and configurability of AI with the reliability of rule-based execution.

These workflows are governed by four principles operating together — accuracy in outputs, transparency in process, security in data handling, and integration across the full indirect tax workflow.

The critical advantage comes from how these elements are combined.

How roles transform in the outcomes era

This shift represents a fundamental evolution in how tax professionals create value. It is the next step in the continued evolution of the compliance function. In recent years, tax professionals have employed tools such as Alteryx to automate data workflows and have leveraged robotic process automation (RPA) to reduce manual effort. This, in many ways, represents the natural next step in that evolution.

As systems take on more of the execution workload, professionals focus on higher-impact work, including analysis, planning, and oversight. The center of activity moves away from performing tasks and toward validating outcomes. With it, the nature of expertise in the function changes.

New roles are emerging from this shift. Examples include compliance process orchestrators, domain experts responsible for validating automated AI actions, and AI reasoning model specialists. These roles focus on monitoring workflows, managing exceptions, and maintaining the integrity of automated processes.

Bunnett also points to the growing importance of knowledge-curation roles, professionals who translate practical, hard-won workflow expertise into the reasoning, controls, and orchestration logic these systems depend on. It requires people who have lived the compliance process; people who know not just what the rules say but how they work in practice, where exceptions arise, and how decisions get made under pressure. That knowledge, encoded into the system, is what makes the AI trustworthy.

The skillset is shifting. Manual execution becomes less central, while pattern recognition, risk assessment, process design, and judgment-based validation become more important. This shift expands opportunities for career development and specialization, particularly at the intersection of domain expertise and technology.

Three organizational models

As execution becomes more automated, indirect tax teams are adapting their structures accordingly. Three patterns have emerged, each reflecting how different organizations are responding based on their size, geography, and operational complexity. None of these are a prescription.

Three patterns are emerging in practice.

In a centralized model, indirect tax activities are managed within a single, consolidated team. Standardized processes and shared visibility make compliance easier to manage centrally, while automated workflows reduce the need for manual cross-functional coordination.

In a regional or federated model, indirect tax responsibilities are distributed across geographies, domains, and functions. Regional teams retain ownership of specific workflows, particularly where jurisdictional complexity requires closer oversight. A shared platform provides alignment across those distributed teams without requiring full centralization.

In a hybrid model, a central team, often functioning as a center of excellence, defines standards, manages the platform, and oversees governance. Regional teams handle execution, validation, and exception management within their areas of responsibility, contributing jurisdiction-specific knowledge while operating within a shared framework.

Across all three models, the underlying shift is the same. Organizational structure becomes less about coordinating manual processes and more about how oversight, validation, and accountability are distributed. The platform reshapes how these structures operate. The people within it determine how well it performs.

WHY DISCONNECTED SYSTEMS BREAK THE MODEL:

• AI can’t orchestrate a workflow it can’t see.
• Point solutions create fragmented visibility.

Unified platforms create end-to-end continuity.

Conclusion

The shift from applications to outcomes is not a pending transition. For the organization ahead of it, it is already here. The gap between them and those still operating in the applications era is widening.

This transition is changing not just how compliance work gets done, but what the compliance function is and how it is valued. Teams that move from execution to oversight become different kinds of partners to the business. They have the capacity for the analysis, foresight, and strategic input that the applications era rarely left time for. They earn a different conversation with the CFO.

The organizations capitalizing on this shift share a common foundation — platforms with broad cross-domain capability that can connect the indirect tax workflow end-to-end, enabling outcome-driven orchestration to be built quickly and with confidence. That foundation is not a technology decision alone. It is a strategic one, with implications for how the function scales, how talent is developed, and how compliance is led.

The compliance leaders who understand this shift now and who are building their teams, governance models, and infrastructure around the outcomes model will define what this function looks like for the next decade. That is not a technology story. It is a leadership one.

BUILT FOR WHERE COMPLIANCE IS HEADING
Thomson Reuters has spent decades building the compliance infrastructure that indirect tax teams rely on and is actively building the outcomes-driven model described in this paper. To explore how ONESOURCE Indirect Tax supports the shift from applications to outcomes across your indirect tax compliance workflow, visit us or speak with your Thomson Reuters representative.