The State of Chargeback Management:How AI Is Reshaping Performance

Fundamentally, enterprise scale merchants are articulating their position clearly – chargebacks are challenging, and while AI has delivered incremental improvements, it is not yet enough to meet enterprise expectations. Introducing automation more fully in decision making is vital for achieving these demands.

Exploring Experiences within Chargeback Management

Chargebacks are a persistent problem for merchants of all sizes, but enterprise-scale merchants are struggling in particular with them. The rise of friendly fraud is driving ever greater fraud volume; creating a sizeable challenge for enterprises to manage.

With the value of chargebacks being raised set to reach almost $57 billion annually by 2031 according to Juniper Research, effective chargeback management is more important than ever.¹

So, in this context, what is effective chargeback management? It means many different things – it means recovering revenue, reducing manual costs, and many other elements. To gain a true sense of the challenges in the market, we surveyed 121 enterprise-scale respondents across the US and UK to understand their challenges with chargebacks, and how the market needs to evolve. Within this report, we will examine their attitudes, and chart a course to a more effective chargeback management solution.

Chargeback Volumes Are Increasing - Causing Difficulties for Merchants

Given the scale of chargeback challenges, increases in volume are perhaps not surprising, but the vast scale is. Enterprises are under siege from chargebacks; creating a problem for the market to solve.

How Are Chargeback Volumes Changing over Time?

How Much Are they Increasing by?

When those who responded that it had increased were asked by how much, they gave a relatively consistent answer, with a majority (62%) selecting that they had increased by 10-25%, and 23% saying they had increased by 26-50%. This was consistent across countries.

10% or more growth over the past 18 months in chargebacks creates a major problem for enterprises; requiring increasingly effective solutions.

What Total Volume of Chargebacks Are Enterprises Seeing?

The volume of chargebacks received a mixed set of responses, but 49.6% of respondents had more than 25,000 chargebacks a year, with over one in four (26%) having over 100,000. This demonstrates the significant scale of chargebacks across the digital ecosystem.

What Factors Are Driving Chargeback Growth?

When asked which factors were most impactful in driving these rises, the top ranked answer was “increase in legitimate chargebacks from fraudulent transactions”, however, all the answers scored relatively well.

Interestingly, the importance of each factor varies significantly depending on which department is involved. Friendly fraud was least important to finance department members, but was the most important factor to risk and fraud department members. Risk and fraud respondents felt that friendly fraud was driving the increase, but that increases in legitimate chargebacks from fraudulent transactions were ranked fourth.

Fraude amicale

We then asked about the prevalence of friendly fraud specifically, as this is a major trend impacting merchants of all sizes. The data was clear - friendly fraud was considered a major problem. 43% of all respondents believed that friendly fraud represented above 25% of their chargebacks. 11% even thought that it exceeded 50%.

This is not surprising – data from Juniper Research that was calculated independently shows that the value of transactions lost to friendly fraud will reach nearly $16 billion in 2031; a growth of 96% from its 2026 value of $8.1 billion, representing 22% of chargebacks in 2026.²

How Are Enterprises Tackling Chargebacks, and Is it Working?

We have already seen the drastic scale of chargebacks, and this is clearly an issue that enterprises need to tackle effectively, or they will experience significant revenue loss. So how is this playing out – what tools and approaches are they using, and what is the current level of effectiveness?

In-house, or Outsourced?

For this element, the answers were uniform across countries – in-house has a lead (54%), but a split between both was a close second (42%). Complete outsourcing was rare in both countries and overall.

As enterprises are relatively unlikely to completely outsource chargebacks, chargeback management vendors will need strong integration capabilities.

Level of Automation

In terms of automation, the survey found that the majority of respondents were using a mix of manual work and automation (68%), however:

Which Tools Are Used?

When asking which specific tools were most popular for chargeback management by enterprises, fraud prevention and risk mitigation was top (84%), closely followed by chargeback alerts (77%).

Indeed, automation needs to evolve. It is not just about automating existing tasks; it is also about improving the decisions and processes that determine financial outcomes, such as which disputes should be prevented or intercepted, when is a refund preferable, which chargebacks are economically worth contesting, what evidence and strategy should be used, and where humans should be involved. Moving to this next level will maximise financial effectiveness through better end-to-end decisions.

Pre-dispute Service

This question found that over four in five (82%) of enterprises use a pre-dispute service; showing that this is considered table stakes for chargeback management by enterprises.

However, while 82% used pre-dispute alerts, 65% are still seeing chargebacks rising, showing that this is not a cure for rising chargeback volumes by itself.

Chargeback Outcomes

Now we have seen how enterprises are handling chargebacks, it is important to assess effectiveness. We asked respondents into which categories the outcomes of their chargebacks fell. The answers were clear: outcomes are not sufficient.

Companies only disputed on average around 32% of chargebacks. Of these, around 32% were disputed successfully. Ultimately, only around 33% of chargeback value was recovered through successful disputes.

This shows that despite heavy investment and progress to date in chargeback automation, there is still a large gap the market needs to address.

Time & Cost of Chargebacks

Looking at this through the lens of time and cost, chargebacks still represent a depressing picture for enterprises.

Collectively, 47% of responses selected categories indicating that chargebacks took up over 25% of their team’s time each week; showing the gravity of the challenge.

In terms of cost, the results were equally noteworthy. Almost half of respondents (49.6%) reported that chargebacks cost between 0.5% and 1% of revenue.

Assessing chargebacks by these metrics, it is clear that despite some technology adoption and automation, chargebacks represent a problem area for enterprises; requiring innovation to solve.

To What Extent Are Enterprises Using AI in Chargeback Management, and What Results Are they Seeing?

When it comes to AI use, the answers here were very consistent, and showed broad adoption for AI systems – 80% of enterprise respondents overall said that AI was used for chargeback management.

Even those who do not currently use AI are planning to. Of the 20% of respondents who did not currently use AI (24 respondents), 54% are planning to adopt AI within the next 24 months. Only two respondents were not planning to use AI at all; showing how important it has become.

An interesting aspect here is how AI use interacts with chargeback trends. There is a clear correlation between chargebacks increasing, and using AI, with 72.2% of respondents who use AI saying chargebacks are increasing, versus 37.5% for those that do not use AI. This suggests that businesses being beset by chargebacks have turned to AI systems.

How AI Is Being Used

Here, the survey found that integrating a third-party AI solution into an existing chargeback management system was the most popular way of deploying AI services at 48% of responses, with third-party, end-to-end AI chargeback management solutions being a distant second at 24%. This suggests that typically enterprises are buying solutions in, and it is more likely to be a system integrated to into an existing system, rather than an end-to-end service. This does suggest that there is major room for growth for third-party end-to-end services.

For those who do not use AI yet but are planning to, “Integrating a third-party AI solution into our existing chargeback management system” was the most desired deployment method at 67%, followed by “Third-party, end-to-end AI chargeback management solution” at 24%; broadly in line with the models chosen by those who have already deployed AI solutions.

Impact of AI Adoption

We asked about what impact had been seen after AI was adopted. The results validated the impact of AI solutions. The vast majority (81%) responded that their win rate had increased by 10% or more, with 43% saying 25% or more, and 5% even saying by more than 50%. Critically, only 5% suggested that their win rate had declined or stayed the same.

Of the 42% of AI users who answered that they had an improvement of over 25% in performance, 48% believed that win rate justified the investment, but only 12% said that reduced staff time did. As such, win rate is a key metric that is top of mind for enterprise merchants.

To see if any particular deployment model was more effective, we calculated the average win rate increase, based on which model was chosen in question 23. This indicated that the average increase overall was 16.5%, but the difference between models was not meaningful at this base.

Of those who have not yet deployed AI, but are planning to (21 respondents), 33% expect win rate increases of 25%, and 43% expect increases of 10%.

It is clear that AI is having a major impact, and the value is recognised by enterprises. However, it is not a silver bullet, and enterprises are still struggling with the impact of chargebacks. It is also important to note that these are self-reported AI improvement values, so they do not have the same weight as independently measures or causally attributable improvements.

One finding that reinforces the state of chargebacks today is comparing chargebacks that are disputed and won by AI use versus non-AI users. Across the survey responses, the respondent-level median share of eligible chargebacks disputed and won was the same if AI is used or not: 6.65%. Exactly the same rate. The survey also found that the median dispute rate was the same whether AI was used or not, 17.5%, but the median win rate differed significantly, at 38% for AI users, versus 27.75% for non-AI users. It is important to consider these findings in conjunction with Q24, where 81% of respondents reported at least a 10% increase in win rate. Two separate questions producing complementary findings provide strong evidentiary support for the relationship between AI adoption and improved dispute win rates.

The identical respondent-level median share of eligible chargebacks disputed and won between AI and non-AI users does not mean AI is ineffective. The reported improvements suggest meaningful, albeit incremental benefits. However, stronger performance on individual disputes does not necessarily translate into proportionately greater success across the full portfolio of eligible chargebacks.

The next opportunity lies in applying intelligence more broadly, identifying economically viable recovery opportunities, improving decisions about which chargebacks to contest, and strengthening the effectiveness of each dispute through more effective, case-specific representment.

What Do Enterprises Want from a Solution?

Decision Making around Adoption of AI Chargeback Management Solutions

When analysing the decision factors around choosing a tool, when looking at what elements were selected as very important by respondents, “relevance and quality of data” and “security & regulatory compliance” were clear winners, both in the 70-72% range. By contrast, “commercial model and total cost” scored lowest.

Enterprises want trust - trust in the data, and trust that they are compliant, and this is seen as much more important than cost.

Data is important for a number of reasons. Relevant transaction, customer, behavioural, and historical dispute-outcome information can support better predictions, stronger evidence, more informed decisions, and continuous performance improvement, thus why it is seen as a priority by respondents.

Priorities for Automation

The survey found that evidence gathering was the most important task to automate, with 60% selecting it as very important. The other tasks were scored very similarly; showing that evidence gathering is perceived as more onerous than the other tasks.

Concerns When Choosing a Solution

The respondents also had clear priorities for issues and challenges when choosing a solution. Data privacy, security or compliance risk was top ranked, matching the answer to question 28. This was followed by similar metrics, such as lack of auditability and incorrect evidence, showing that trust is again central.

What Outcome Is Most Important

For this question, the top two answers were “Improvement of win rate” at 39% and “Increase in revenue recovered from chargebacks” at 34%, and “Reduction in staff time” (16%) and “Reduction in overall cost” (12%) scored relatively poorly.

When comparing this to the answers in Figure 16 and Q24, it creates an interesting case for the adoption of AI in chargeback management. The most desired benefit of adoption is an increase in win rate, which the results in Figure 16 and Q24 support. However, the second most desired benefit is increased revenue recovery from chargebacks, but Figure 16 suggests there is no difference in the proportion of chargebacks contested and won.

Fundamentally, this means that trust and data quality are prerequisites for adoption, but financial performance ultimately justifies the investment.

Buyers also want systems that can adapt and improved based on quality data, but they must remain transparent and controllable.

Minimum Improvements Required

Expectations from enterprises are quite varied, with 45% of respondents accepting a less than 10% percentage of cost saved, versus 31% for the other three metrics. “Increase in funds recovered” has the highest expectations, with 28% “Wanting over 25%” percentage growth.

Importantly, the survey indicates that 69% of respondents require an improvement of at least 11 percentage points in dispute win rates, while 68% require recovered funds to increase by at least 11%.

Enterprise buyers aren't merely asking for incremental efficiency improvements. They're setting substantial performance expectations.

The level of improvement being demanded by enterprises is notable when compared to the reported improvements from Q14. The majority of enterprises expect at least an 11-percentage-point improvement, with AI users reporting a median win rate 10.25 percentage points higher than non-users. However, the survey found no difference in the proportion of chargebacks disputed and won.

Are Enterprise Merchants Worried About the Holiday Season?

For many merchants, the holiday season typically causes anxiety, as when eCommerce activity surges, so do chargebacks. Here, we will explore what that means for enterprise-scale merchants.

Do they Increase?

The results here were clear – 77% of respondents expected to see an increase in chargebacks over the holiday season.

How Important Is Planning?

When such disruptions are expected, planning is an important element, but how important?

The answers to this question showed that nearly all respondents considered it important to prepare. “Net important” was 93%, with 60% considering it “Very important”. Only 2% thought it was ”Unimportant” to prepare for chargebacks during the holidays.

This was relatively consistent between the countries, though the UK had a slightly lower net important score compared to the US (88% versus 96%).

Will it Be Worse this Year?

There were much more mixed views about the 2026 holiday season. Overall, a majority (58%) of respondents thought this year would be better for chargebacks than last year, which was relatively consistent across countries (UK – 46%, US – 64%).

However, around one in five did still believe this year would be worse (19%). Expressed as an average score out 5 (5 being significantly better), the score was 3.496/5.

For the holiday season then, the takeaway is clear – prepared optimism is what the respondents are feeling. The holiday season creates challenges, but with careful preparation, they expect this year to be easier than the last

The Next Phase of Chargeback Management

Enterprise merchants face a fundamental challenge. Chargeback volumes are rising, even as investment in prevention, automation, and AI-powered management has become increasingly widespread.

The findings demonstrate that technology adoption alone is not sufficient to address this challenge. AI adopters report meaningful, but often incremental improvements in dispute performance, yet substantial inefficiencies and unrealised recovery opportunities remain.

The next phase of chargeback management will require a broader approach to how true intelligence is applied throughout the dispute lifecycle.

Rather than focusing exclusively on automating individual tasks, enterprises need to improve the decisions that determine financial outcomes. This includes preventing avoidable disputes, intervening earlier where appropriate, identifying economically viable recovery opportunities, and strengthening the evidence and strategies used to pursue them. For example, we foresee the transition from dynamic representment to true, case-specific strategy that prioritises the most effective evidence for each individual case. This transition marks the leap from process automation to data-informed AI decisioning, and promises to move the needle in both reducing manual effort, and most importantly, actually improving dispute outcomes.

Data availability and quality will be central to this evolution. Relevant transaction, customer, behavioural, processor and bank information, and historical outcome information can support more informed predictions, better decisions, and more effective recovery strategies.

Equally important is the ability to connect intelligence across workflows, learn from outcomes, and continuously refine performance. For enterprises, this must be accompanied by the transparency, security, auditability, and human oversight necessary to operate AI confidently at scale.

The research also demonstrates that enterprise expectations are substantial. Buyers overwhelmingly prioritise stronger win rates and recovered revenue, with more than two-thirds requiring double-digit improvements in these measures. However, with only incremental gains, the true promise of AI automated representment has not yet been realised.

Ultimately, the opportunity is not simply to automate chargeback management. It is to make chargeback management materially more effective.

As AI capabilities mature, the next generation of solutions will be judged by their ability to turn intelligence into better decisions, coordinate those decisions across the dispute lifecycle, and deliver measurable financial improvements.

Survey Methodology

In September 2026, Juniper Research conducted a survey of C-level executives and senior managers from enterprise scale businesses across the UK and US. 121 respondents took part in the survey; 80 in the US and 41 in the UK.

Survey respondents had the following characteristics: