Introduction
Earlier this year, we updated our annual Grading the IRS report, commending the Service for making operational changes that should result in better services for taxpayers. NTUF’s annual evaluation of the performance of the IRS serves as a more insightful review than its own previously self-reported report cards.
Our report measures improvements in key areas including taxpayer service, modernization, and closing the tax gap. This year, we raised the letter grades for technological modernization from D to C and taxpayer services from C to B-, conditional upon improvements made during the 2026 tax filing season. Efforts to close the tax gap moved slightly from a C- to C+.
Since then, the Treasury Inspector General for Tax Administration (TIGTA) continues to release new information on the tax filing season and other IRS operations. While full results of the filing season are not expected until later this year, the Government Accountability Office (GAO) released its own findings in August regarding the filing season.
This new information provides enough insight to provide an update in one key area: taxpayer service. Metrics used to evaluate service levels have changed recently, but the new figures make it difficult to assess whether quality of service has improved, declined, or remained the same. The IRS should clarify its goals and take measures necessary to provide better service before the next filing season, whether those measures involve hiring more staff or deploying additional technology resources.
Telephone Service and Live Chat
When taxpayers encounter issues, they generally prefer to call and speak with a live IRS agent. Around 80% of all taxpayer assistance was provided over the phone in recent years.
In our 2025–2026 update to our Grading the IRS report, we were encouraged by news that the IRS would ditch its faulty phone metrics in favor of a new metric which would also account for new technologies. Part of the improvement of our grade for IRS taxpayer services was contingent upon successful implementation of the new metric as well as effective phone service provided. Unfortunately, during its implementation of a new metric, the IRS seems to have moved the goalposts and performed worse than before.
The IRS typically evaluates the quality of its telephone assistance using its flawed Level of Service metric. In past years, when the IRS has reported a Level of Service of 87%, just above its typical target rate of 85%, it actually only answered 35% of phone calls. However, the Level of Service metric typically only tracks calls to one phone line, leaving out the majority of calls to the IRS.
In April 2026, Treasury Secretary Scott Bessent testified that the IRS would be using two new metrics, the Assistor Service Rate (ASR) and the Enterprise Service Completion Rate (ESCR), with the latter being a more holistic calculation. Recent reports only account for the former calculation, with TIGTA noting that the IRS began using a metric called Assistor Service Rate for this filing season.
TIGTA claims that the ASR is very similar to the Level of Service, stating “the only difference is that the Assistor Service Rate includes contacts answered through Live Chat,” a service where taxpayers can submit questions to IRS customer service on a webpage. For 2026, this feature had an extremely narrow scope, only being available to international taxpayers seeking assistance with individual income taxes.
Prior to the filing season, the National Taxpayer Advocate described the difference between the two new metrics as follows: “ASR would serve as a resource-driven measure to help determine live assistance funding. It would revise the existing LOS measure and include service provided by assistors over the phone and through live chat. ESCR would provide a holistic view of the live assistance services provided to taxpayers across the IRS, including those provided by new technology.” While the ESCR accounts for more taxpayer interactions, there are still questions as to whether it covers all telephone lines. The official descriptions of the metrics are outlined below.
Due to staffing shortages, the IRS reduced its typical Level of Service target rate from 85% to 70% for the new ASR metric. At the time of GAO’s most recent report, the IRS had reported exceeding its target to reach a 73% ASR, falling well below the 87% Level of Service reported last year. This is even after accounting for an 11% decline in call volume since last year, suggesting that the IRS is struggling to meet service goals even as fewer taxpayers call for assistance.
Areas for Improvement
If the only difference between the Level of Service and the ASR is the inclusion of live chat metrics, the measurement still excludes the majority of its inventory of over 100 phone lines. Worse yet, call wait times doubled from the 2025 filing season to the 2026 filing season. In its interim filing season report, TIGTA reported its in-house Level of Access metric, which compares the total number of calls received with the number that receive assistance, stood at 44%. This is well below the 73% ASR at that time.
Regardless, none of the existing metrics convey the quality of service provided to taxpayers. This is increasingly important as the new metric in use accounts for both phone service and live chat services. TIGTA has recently called attention to the fact that the IRS does not have a process to evaluate the accuracy and completion of live chat interactions. In fact, the system currently used to assess live chats is so inaccurate that it showed one live agent handling more than 600 chats simultaneously. Metrics also do not account for how many taxpayer inquiries are “out of scope”, meaning agents cannot answer those questions on the phone.
GAO and TIGTA agree that staffing shortages contributed to the lackluster taxpayer services during this filing season. Although Erin Collins, the National Taxpayer Advocate, has suggested that hiring more people does not necessarily lead to better service, the IRS must have enough properly trained and qualified employees to handle peak assistance demand times throughout the filing season.
IRS officials told GAO that reassigning customer service representatives from phone lines to answering correspondence helped reduce correspondence backlogs and late correspondence, but this undoubtedly worsened phone and chat metrics. The IRS should have enough employees to do both functions.
Modernization can help fill in some of the gaps. However, it is still unclear what technological upgrades are being made to address the challenges presented by low staffing and higher demand for online services. Efforts we have highlighted previously seem to be ongoing, such as allowing more taxpayer correspondence to be uploaded electronically. We look forward to monitoring further updates on IRS efforts to modernize technology and close the tax gap.
Conclusion
The purpose behind NTUF’s Grading the IRS report is to measure success in a way that the public can easily understand. The IRS is taking the opposite approach by moving the goalposts, dramatically reducing its taxpayer service target with the addition of its new Assistor Service Rate metric. In addition, the IRS seems to have provided service well below prior expectations with the new metric only tracking a service rate of 73% instead of last year’s 85% for the Level of Service metric.
Our next annual update to the Grading the IRS publication will take the 2026 filing season into account as well as any upgrades made in its aftermath. In the meantime, the IRS should hire enough customer service representatives to adequately staff all service functions. It should also implement new technologies to both promote taxpayer access to service and streamline job functions for agents.
Allowing the IRS to measure its own performance, as it did with its previous report cards, carries the risk of exaggerating progress and overlooking the most important goals. Congress can take a more active role in demanding progress through provisions included the Taxpayer Assistance and Service Act, which would enhance many aspects of tax administration.