Highlights
- Three catalysts drive AI evaluation: regulatory change, transformation initiatives, and efficiency demands.
- Each catalyst introduces blind spots that can distort what gets measured during evaluation.
- Defining operating constraints before vendor engagement creates stability across stakeholder perspectives.
Tax teams are evaluating AI for indirect tax under more pressure than ever, but the force driving that evaluation often distorts what gets measured. Identifying yours is the first step to correcting it.
The pressure of needing to keep pace with rapid technological change is measurable for indirect tax professionals. According to the Thomson Reuters 2025 Corporate Tax Department Technology Report, 94% of corporate tax professionals feel positive about the future of tax technology with 73% feeling hopeful and 21% feeling excited. The vendor market is responding to this trend. AI solutions are multiplying, capabilities are expanding, and organizations are moving forward with decisions faster than their evaluation processes can keep pace.
The urgency of keeping up isn’t the problem, but rather how different catalysts shape what gets prioritized and what gets missed.
In most organizations, indirect tax AI evaluation is set in motion by one of three forces. Each one is legitimate while also introducing a blind spot.
Jump to ↓
The 3 catalysts and what each one misses about indirect tax AI evaluation
How to navigate these catalysts to ensure a realistic evaluation for indirect tax AI solutions
The 3 catalysts and what each one misses about indirect tax AI evaluation
Catalyst 1: Regulatory change
Compliance obligations in indirect tax are expanding. New e-invoicing mandates, continuous transaction controls, cross-border reporting requirements, and evolving state and local rules are increasing the volume and speed of what tax teams must manage. When regulatory change is the primary driver, AI evaluation focuses on coverage. This applies to which jurisdictions are supported, how quickly regulatory updates are reflected, and whether the system can keep pace with expanding compliance requirements.
That focus is appropriate, but it can lead to overvaluing vendor claims about jurisdictional breadth at the expense of testing those claims against real operating conditions.
A system that covers many jurisdictions on paper behaves very differently from one that handles those jurisdictions accurately when data is incomplete, rules have recently changed, or local variations apply. Coverage claims are easy to make while operational accuracy under realistic conditions is harder to evaluate and harder to fake.
Blind spot 1: Regulatory-driven evaluations often optimize for coverage breadth while undervaluing accuracy depth. The question that gets missed is, “How does the system perform in the specific jurisdictions where our transactions are most complex?”
Catalyst 2: Technological transformation
Many organizations evaluating AI for indirect tax are doing so inside a broader transformation initiative, like an ERP migration, a finance modernization project, or a push toward integrated tax technology. In these contexts, AI evaluation is shaped by architectural considerations and invites questions including, “How does this solution fit the target state?” “What are the integration requirements?” or “Does it work with the systems we’re building toward?”
That framing surfaces important questions, but it can lead to over-emphasizing architectural alignment while under-emphasizing operational performance. A system that integrates cleanly into the target architecture still needs to produce accurate, auditable results in an environment where data quality is variable and workflows aren’t yet stable.
Transformation-driven evaluations also face a timing challenge. ERP migrations and system consolidations often mean that the environment in which AI will eventually operate doesn’t fully exist yet. Evaluations that involve future-state architecture may not surface the friction that will emerge during the transition.
“Almost 70% of tax departments remain in chaotic or reactive stages of digital maturity, with only 6% operating optimally. Nearly 60% lack confidence in their ability to upgrade systems within two years.” — 2025 State of the Corporate Tax Department
Blind spot 2: Transformation-driven evaluations often optimize for architectural fit while undervaluing production readiness in an environment that’s still changing. The question that gets missed is, “How does this system handle data conditions during transition, not just in the target state?”
Catalyst 3: Efficiency demand
Tax departments are being asked (and are expected) to do more with less. Headcount constraints, increased transaction volumes, and growing stakeholder expectations for speed are pushing teams toward automation. When efficiency is the primary driver, AI evaluation focuses on how much the system can do without human intervention. This involves automation rates, processing speed, and the ability to scale output without scaling cost.
This catalyst can produce overly optimistic vendor claims and even misleading evaluation dynamics. Automation rates are easy to demonstrate under controlled conditions and difficult to sustain in production, where data variability, exception rates, and workflow dependencies all reduce the effective rate.
More fundamentally, efficiency-driven evaluations may not factor in governance. A system that automates a high percentage of determinations but provides limited visibility into how those determinations were made, or limited ability to review and correct them, doesn’t eliminate compliance risk. Instead, it concentrates the risk in everything that wasn’t reviewed.
“Without human interaction, how do we know that tax return or other processes are good the way they should go? AI is great to catch abnormalities, but if you’re doing something a certain way for a reason, will it know that?” — US Tax Firm President, 2026 AI in Professional Services Report
Blind spot 3: Efficiency-driven evaluations often optimize for automation rate while undervaluing auditability and oversight. The question that gets missed is, “What happens to accountability when automation is high and human review is limited?”
Why this pattern matters
These three catalysts shape who gets involved in the evaluation, how vendor conversations are framed, and which criteria for AI indirect tax solutions remain in focus as the process progresses.
When regulatory compliance drives evaluation, the tax function leads, which is appropriate but often means that downstream IT and finance considerations arrive late.
When transformation drives evaluation, IT often leads, which surfaces integration requirements but may not adequately value the operational needs of the tax function.
When efficiency drives evaluation, finance needs are prioritized, but tax accuracy and governance may not receive equal scrutiny.
In practice, most organizations face some combination of all three. When evaluation criteria shift along with different catalysts (as they often do when broader stakeholders engage), the process loses stability. What’s being measured changes depending on who’s in the room.
That’s when the gap between evaluation performance and production performance appears.
How to navigate these catalysts to ensure a realistic evaluation for indirect tax AI solutions
The starting point for any of these catalysts is the same. You must define the operating constraints that exist regardless of what’s driving evaluation, including:
- Where human oversight must remain, no matter the automation rate
- What data conditions the system will face, regardless of the target architecture
- What jurisdictional accuracy requirements exist, beyond headline coverage claims
- What integration dependencies cannot change
Defining those constraints before engaging vendors creates stability. Evaluation criteria that start from operating realities are harder to shift mid-process because they’re anchored in conditions that won’t change.
It also creates a common reference point for stakeholders. Finance, IT, and tax leadership are each responding to one or more of the three catalysts. Shared operating constraints give them a basis for evaluation that doesn’t collapse when perspectives diverge.
Where to start
Before moving further into evaluation, it’s worth naming the primary catalyst driving your organization’s decision.
If regulatory pressure dominates: Build in deeper testing of accuracy at the jurisdictional level. Don’t just test coverage or breadth.
If transformation drives the decision: Ensure evaluation accounts for transition conditions, not just the target state.
If efficiency is the primary metric: Define governance requirements explicitly before assessing automation rates. Include what must be reviewable, traceable, and correctable.
Knowing your catalyst helps you compensate for its associated blind spot before it shapes a decision.
The full framework
The six-question evaluation framework in our guide, Before You Commit – a practical guide to evaluating AI in indirect tax, was designed to remain stable regardless of which catalyst is driving evaluation. It covers domain expertise, data handling, human review, integration, and ongoing validation, which are the dimensions that matter in production, no matter the path to selection. Read the guide in full.
