← 返回

2026.09.23 | TB | Fintech | Bread Financial Fraud & Risk

原标题：Bread Financial – Fraud Prevention, AI-driven Risk & Approval Strategy

#financials

Third Bridge · 2026-09-23

Analyst-led • 11 Sep 2026

Companies: BFH US, SYF US, C US, LOW US, HD US

Adam Johnson (AJ) Senior Director, Fraud Analytics at CNG Holdings Inc

Jonathan King (JK) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

JK: Welcome to Third Bridge's Interview entitled Bread Financial – Fraud Prevention, AI-driven Risk & Approval Strategy. My name is Jonathan King, and I'll be facilitating today's Interview. Very pleased to welcome Adam Johnson on the platform today. Adam is the former Director, Fraud Strategy at Bread Financial. Thank you so much for joining us today.

Just before we begin, could you please state either I agree or I disagree to the following statement: You understand the definition of material non-public information and agree not to disclose any such information, or any other information which is confidential, during this Interview.

AJ [00:00:33]: I agree.

JK [00:00:34]: Could you share a little bit about yourself and your background?

AJ [00:00:44]: Yes, for sure. I graduated with my MBA from Indiana University in 2011. My first couple of roles after business school were in consulting, where I was focused on both model validation and risk within many large US banks. After several years of doing that, Synchrony Financial reached out to me and asked me to be one of the initial members of their acquisition fraud strategy team. That was my first post-consulting role, where I led acquisition fraud strategy for Synchrony, responsible for all of their fraud strategies for all of their private label and co-brand products. I was promoted several times at Synchrony. One of the roles I was in was Vice President of Emerging Fraud Strategy, and then also held some portfolio credit roles there as well. All told, I believe I was at Synchrony for eight or nine years and then made the move to Bread Financial to be the Director of Fraud Strategy, where I led a fraud strategy team of 14 or 15 people and was responsible for really all types of fraud, but the big ones I dealt with were third-party fraud, account takeover, synthetic and transaction.

Questions

[00:02:21] Bread Financial has delivered several consecutive quarters of improving delinquency and net loss rates while simultaneously shifting its portfolio mix towards co-branded cards. How much of that credit improvement do you think reflects better underwriting and fraud discipline vs simply having a better-quality borrower mix?

AJ [00:02:54]: The borrower mix is the bigger part of it, maybe 85-15. While I was there, we were going through a couple of major projects that improved underwriting, improved fraud strategy, but when we see a shift like this, I think the biggest piece of the puzzle is just underwriting more qualified borrowers.

[00:03:33] Could you walk me through some of the projects that Bread was working on to improve the underwriting and some of the fraud strategy? What did that look like in practice?

AJ [00:03:46]: Yes, absolutely. The big one was transitioning to the Alloy fraud orchestration platform. Alloy is a pretty significant fintech company in the marketplace right now. Their product is really strong. It basically allows Bread to leverage all of their fraud tools, be it fraud scores, authentication, device intelligence, whatever it may be in real time and provide just a better experience for the customer. It allows them to do a better job of splitting out high-risk applications vs low-risk applications and applying the appropriate authentication type based on that risk. Certainly, I think that that is a piece of the puzzle as we look at their improved loss rates.

[00:04:51] What does Alloy change in Bread's underwriting process? Could you outline how it integrates with Bread's system?

AJ [00:05:10]: Yes, absolutely. Previously, they were in a very manual environment, where if they wanted to make a change, it required a project that would take months to deploy and put together. The different tools and scores and things didn't work together. You could call iovation and apply a certain type of authentication based on your result with iovation, but you couldn't use, say, an iovation result along with a Socure fraud score to split out a specific high-risk or low-risk segment. In addition to that, when you think about authentication, it streamlines that process for the customer. When you think about going through a credit card application, the landing pages that you would end up on previously at Bread saw a significant drop-off, whereas now the authentication process is much more streamlined. I imagine they see much lower drop-off as people go through that authentication process.

[00:06:49] Among different types of frauds, which do you think poses the biggest risk to Bread today, and why?

AJ [00:07:00]: I'm a big believer that they probably have a little bit of a synthetic fraud problem. The reason being several fold. (1) With AI, it's becoming easier and easier to make synthetic identities. They're certainly seeing more application volume from synthetic identities. (2) Their push towards co-brand. Co-branded cards have higher credit limits, which makes these accounts much more valuable to synthetic fraudsters. With the private-label cards, oftentimes you see lower credit limits. Maybe a synthetic fraudster would use one of those private-label credit cards to build up their credit score, but oftentimes, what you'll see with synthetic fraudsters is that they'll pay off those accounts in order to build up their credit score. They'll use a Bread private-label account to build up their credit score and then go somewhere else where they can get a bigger credit line now that they have a higher credit score and bust out there. Now with Bread shifting towards co-brand products, they're going to be more of that target where if you have a good synthetic identity, you're going to go there and bust out. Then my last comment in terms of Bread and synthetic fraud is they aren't leveraging the Social Security Administration's electronic consent-based Social Security verification product to authenticate applications which are high risk for synthetic fraud. In my opinion, that's a very large gap for them that makes them susceptible and a target for synthetic fraudsters.

[00:08:58] You said Bread has been dealing with synthetic fraud problem. What impact could that have on the company? How would you see that manifest in the financials, and where do you expect the biggest impact to show up going forwards?

AJ [00:09:26]: It manifests in credit losses. This is going back to my time at Synchrony. Synchrony was part of the pilot programme for the eCBSV product with the Social Security Administration. This is public information. If you fill out a credit application, you can see there's a disclosure there, whether or not they're sending you through eCBSV. Synchrony Financial does, Bread does not. If you do not use that product, what eCBSV does is, it sends the name SSN and date of birth on the application to the Social Security Administration and checks, does that combination of name SSN and date of birth exist in the Social Security Administration's database. If it does, it's a match, you know it's a real identity. If it doesn't match, you have a synthetic identity. That is the only way to definitively identify a synthetic identity. With Bread Financial not using that product, they don't know if an account that charges off is synthetic fraud or if it's just somebody who disappeared and they can't get a hold of and won't pay their bill. Long story short, it all shows up in credit losses at this point for Bread.

JK [00:11:13]: Why do you think Bread is not leveraging the Social Security Administration's electronic consent-based verification product for authentication?

AJ [00:11:29]: It's a cost thing. I was a part of it. I led Synchrony's pursuit and pilot testing with that, with the Social Security Administration. Initially, the cost to run eCBSV was pretty low, but then it didn't meet the requirements the government had for… They were split. The government wasn't getting back to break-even on the product nearly as quickly as they needed to. The Social Security Administration dramatically increased the price. Synchrony's response to that was, "Okay, we can't pay this on every application, but if we run a synthetic fraud score and identify applications that are high risk for being synthetic fraud and then run those through, then the maths works out." I tried to sell that when I was at Bread and just wasn't able to get it done while I was there.

JK [00:12:46]: What did the price increase look like?

AJ [00:13:02]: It depends on how many transactions you run through eCBSV. I don't remember the exact tiers, but I think if Bread were to do it right and only run it on high-risk applications, I think it would be around USD 750,000 for them.

JK [00:13:31]: Every year?

AJ [00:13:33]: Yes.

[00:13:34] Bread processes a high volume of new account applications through Comenity Bank. How vulnerable do you think that funnel is to scaled AI-driven bot attacks? Can Bread's current fraud stack respond quick enough when that attack volume spikes?

AJ [00:13:58]: Yes. Now that they're on Alloy, I think they're in a pretty good position to respond to bot attacks. When I was there, we saw bot attacks very regularly. We had all the reporting set up to identify it and respond quickly. Frankly, on many of the bot attacks, you see so much application volume so quickly that it's very obvious when it's happening. It's honestly much more difficult to identify when it's more of a slow drip, but generally speaking, yes, I think some of the bot protection stuff sat outside of my team. My team did lots of monitoring and reporting to identify bot attacks, but in terms of what we're doing with Amazon Web Services and things like that that all sat outside of my team, but generally speaking, yes, I think they're in good shape with their current stack to identify bot attacks.

JK [00:15:07]: It seems as if Bread is more conservative on fraud than peers such as Synchrony. As AI-driven fraud becomes harder to detect, do you think that conservative approach provides better loss protection? How do you think about that dynamic?

AJ [00:15:31]: I left Bread just after they were moving on to Alloy, completing the project to roll over to Alloy. Their fraud strategies, one of the big things I did was work to modernise their fraud strategies. Prior to me being there, they were very rudimentary, using a chain saw instead of a scalpel. I doubt that they've gotten to where they need to go in terms of improving their analytics and their fraud strategy, but I think they were, by using simplified strategies that didn't really connect all the things you need, device intelligence, behavioural biometrics, fraud scores, things like that, they were impacting a lot more good customers than they needed to. I think they have a lot of the right tools. It's just a matter of doing a better job, a more analytical job with strategy development is the key for them. If they're able to do that, they can define risk more clearly and assign the appropriate amount of degree of authentication to the correct customers.

[00:17:18] You said Bread was impacting more good customers than it needed to. Can you break that down for us? What does that look like in practice?

AJ [00:17:28]: Yes, absolutely. This is just an example. I don't remember the specifics of their fraud strategies, but I work on this stuff every day. This is just a high-level example. Let's say, their main fraud score, they get from a company you may or may not have heard called Socure. They've made the decision to send everyone with a Socure fraud score greater than 0.6, which is probably, let's say, 25% of applications through a facial biometric, which is harder for people to complete. It's more friction. There's a lot of drop-off there, whereas a better strategy would say, let's say, they have that exact same cut-off with the Socure fraud score, what did I say, 0.6, 0.65 whatever I said, but let's say their IP distance, the distance of the device that they're applying on from their home address on their application is only 10 miles. That application is much lower risk than an application with the same Socure fraud score where the IP distance is 400 miles.

Let's send the application with 0.65 Socure fraud score and 10 miles of IP distance through one-time passcode where you see very little drop-off of good customers and the application where they have the 0.65 Socure fraud score and a 500 mile IP distance through the facial biometric. You're able to make more granularly define risk and assign the appropriate degree of friction.

[00:19:47] If you were leading Bread's fraud strategy today, what would be the single highest-priority investment heading into 2027, or if you could make investments to improve the company's fraud strategy, where would you make those?

AJ [00:20:09]: Two things come to mind. One, as I already talked about with eCBSV, I think especially with shifting more towards co-brand, higher credit limit products, it is absolutely critical that they're able to split out synthetic identities. With AI, as I said, AI is just making those easier and easier to create. I would absolutely invest in eCBSV. I pushed for it while I was there, we didn't get it across the finish line. The other thing that I would do is I would invest in highly analytical, skilled strategy developers to really get the most out of Alloy and use advanced econometrics to create better strategies that, (1) Increase overall approval rate, and (2) Decrease fraud. With where their strategies were at when I was there, I think there's an incredible opportunity there to do that. Right now, they just don't have the people in place to get that done.

[00:21:38] Bread's management has guided to a full-year 2026 net loss rate of around 7% with a longer-term ambition near 6%. What do you think needs to go right to get down to that level? What are your thoughts on the trajectory going forwards?

AJ [00:22:08]: To get down to that 6%, I think you need to see unemployment stay at least as low as it is right now. Really to get down to 6%, I think that you would need to see a little bit better wage growth vs inflation in that prime to subprime grouping where Bread operates to get their payment rates up. Those are going to be the big drivers over the next 12 months.

[00:22:51] Payment rates and net loss rates have been improving. How much of that do you think is a genuinely healthier consumer today vs a shrinking pool of taking the weakest borrowers and rolling them off the book?

AJ [00:23:06]: I'm less pessimistic than most people about how things are going for Middle America in terms of the economy, but at the same time, I think in this situation, it's 80% the shift in focus that they've had towards higher credit quality borrowers and lending less to people in that middle-to-low-income group.

[00:23:54] You're less pessimistic on the macro, but if something were to happen such as unemployment were to tick up, what impact do you think that would have on Bread? How quickly would you expect that to show up in delinquency data?

AJ [00:24:15]: It would have a significant impact. Unemployment is always the first thing that I look at. You'd start to see that in the 30-60 day delinquencies really quickly. There's more of a lag in terms of charge-offs, but you would start to see in the 30, 60, 80 day delinquency within one month or two. It would show up very quickly there.

[00:24:46] A big part of Bread has been releasing loan loss reserves as credit metrics improve. At what point do you think further reserve release starts to look aggressive rather than warranted?

AJ [00:25:11]: Any indication that unemployment is ticking up they're going to want to slow down that plan dramatically. Again, looking at unemployment as the initial driver, but also looking at wages vs inflation and just the cost of living for people in that category. Oftentimes, for middle to especially low-income people, if it comes time to pay their bills, a private label credit card is going to be the first one thing that they take a pass on. They're particularly sensitive to any shift in unemployment rate, wages, etc

[00:26:10] You talked about improving the analytics around Alloy. How much approval rate improvement do you think Bread can realistically unlock from this investment with Alloy?

AJ [00:26:32]: On credit-approved applications, it could be anywhere from I would say that it could be as high as 8% even. If we're talking about total applications through the door, it's going to be lower, maybe 3-4%…

JK [00:27:00]: What explains the difference between that 8% vs the 3-4%?

AJ [00:27:08]: It's just a difference in the denominator. Of the credit-approved applications, if our denominator for approval rate is credit-approved applications, the denominator is smaller. These are all creditworthy people. If we're talking about total applications, then it's just a bigger denominator in the approval rate.

[00:27:35] Where do you see the bigger opportunity today – reducing fraudulent approvals or curbing unnecessary friction on legitimate applicants?

AJ [00:27:54]: Reducing unnecessary friction. They have a huge opportunity there. As I said, when I got there, their fraud strategy was really, again, using a chainsaw. I think that if they were able to get some people who could do some really strong strategy development, they could really unlock a lot of approval rate there because they have all the tools, especially now that they're on Alloy. They've got a lot of the right scores and tools, best in the industry. There's a great opportunity there if they were to get the right strategy development team in.

JK [00:28:41]: How much of the benefit from Alloy do you think is already reflected in Bread's loss rates vs being an unrealised opportunity?

AJ [00:28:49]: 65% is probably already in there. The rest of the 35%, again, would just be from improved strategies.

[00:29:15] Bread has been pushing Bread Pay and it's now bundled into partner programmes, which is putting it in a more direct competition with Affirm and Klarna at the POS [point of sale]. From a fraud perspective, does the card issuer's existing fraud infrastructure and risk-tiering translate well to BNPL [buy now, pay later] underwriting? What do you think about that? Is it a meaningfully different fraud problem than revolving credit?

AJ [00:29:47]: Yes, there are some key differences. With buy now, pay later, you're more concerned about merchant fraud and then also first-party fraud as well. I would think that they would need to invest in a first-party fraud score to identify intentional misuse, people who are likely to just not pay or never pays, but generally speaking, a lot of the fraud tools and scores are also applicable in that space. There's just some nuance with some different fraud types.

JK [00:30:48]: Is there anything else from a fraud perspective that we haven't talked about for Bread today?

AJ [00:31:05]: No, I think we've covered it pretty nicely. Again, they're using best-in-class products, fraud products, Plaid, Prove, the behavioural biometrics, they've got it all there. It's just a matter of optimising everything that they have.

JK [00:31:32]: When you say optimising everything, to me, that sounds as if getting all the data you can out of it and applying that data. Is that how we should be thinking about it?

AJ [00:31:45]: Yes, exactly.

[00:31:48] Bread recently landed Ford as a co-brand partner. The company continues to add partners such as Ethan Allen and others. What do you think wins a retailer away from an incumbent issuer such as Synchrony or Citi? How repeatable do you think that pitch is across other categories?

AJ [00:32:17]: I do not see any structural advantages that Bread has over a Citi or Synchrony. I think that they are willing to do some customisations for the right client that a Synchrony or Citi may not be willing to do, but they're in a position where a lot of these agreements are impacted by what other customers or what other companies a Synchrony or a Citi is working with. If Ethan Allen, for example, I think they sell furniture. If certain competitors of theirs are working with Synchrony, Synchrony either may not be willing to work with Ethan Allen or Ethan Allen may not be willing to work with Synchrony. There's an element of that that really works in Bread's favour in some of these instances.

JK [00:34:05]: It seems to be an exclusivity thing, or even if it's not explicit, they just might not work with their competitors.

AJ [00:34:13]: Exactly.

JK [00:34:15]: You said that you don't see Bread having any structural advantages over some of its competitors. Besides scale, Do you see the company at disadvantage anywhere outside of fraud against those competitors?

AJ [00:34:38]: Yes. Synchrony's technology is much better than Bread's. From a talent perspective, I think Synchrony has been very focused on analytics and bringing in analytic talent for a long time. Bread has not been, and I think they have a hard time attracting some of those people. I think that's an area where Bread is at a disadvantage.

[00:35:20] Partner concentration is a big risk factor industry-wide. History shows partners sometimes walk away from agreements. If Bread were to lose one of its top 3-4 branded partners, how long do you think it would realistically take to backfill that receivables volume? What impact do you think that would have?

AJ [00:35:48]: The receivables volume is really, really tough. Sales less so, but receivables volume is really tough. Not to go back to my Synchrony days again, but Synchrony lost Walmart while I was there. It took them two or three years to fill that gap. That was a really big deal. It's not easy… With Bread specifically, I think they are very concentrated within their top three or four partners, very, very concentrated. They've got three or four key partners and then a lot of far less significant ones.

JK [00:36:46]: What I gather from that is there's a lot of risk at the top.

AJ [00:36:49]: Yes, absolutely.

JK [00:36:52]: What do you think might drive any of Bread's top partners to change providers?

AJ [00:37:02]: Thinking about them, the relationships there are very strong. Some of them are connected to Columbus, Ohio, where Bread's headquarters are. They've been with them for a long time, but as we've seen, if one of those companies is struggling and is knocking on every door to see where they can improve, I think that they would certainly consider a competitor and a competitor could come in, offer them a deal that Bread probably couldn't match, offer them better technology and certainly steal them away that way.

[00:38:01] We haven't talked about agentic commerce yet. As AI shopping agents become more prevalent, what do you see as the biggest fraud and authentication challenges for issuers when an AI agent is acting on behalf of the consumer rather than the consumer directly?

AJ [00:38:27]: The key here is we've got to be able to authenticate the agent itself. A lot of companies are working on this problem. Right now, the main solution is to assign a token to the agent, but end of the day, you have to be able to authenticate the agent. I think that big picture, it's just a matter of using all of your signals. You've got to, again, authenticate the AI agent, but then you've also got to be able to compare all of your different other signals, device fingerprint, network intelligence, device intelligence, your standard fraud scores, your behavioural biometrics. That's where a product like Alloy becomes so valuable. You've really got to lean on does everything line up for this application. I think that's really the key.

JK [00:39:57]: Do you think assigning a token to the agent is a sufficiently robust long-term solution? It makes a lot of sense now, and it seems as if you can layer on all these different things.

AJ [00:40:18]: To be honest with you, I don't know if I'm a technical enough person to break down for you, the weaknesses of that approach. Right now, I think it's effective and working well, but as we know with fraud, these fraudsters are always trying to find new ways to beat you and things are always changing.

[00:40:49] Where do you see the biggest risks and challenges for Bread over the next 1-3 years? What are the biggest headwinds for the company?

AJ [00:41:09]: A couple of things. I don't think they're well-positioned to compete with buy now, pay later. If we continue to see growth like we have from Affirm and Klarna and all the fintechs doing buy now, pay later, eventually, that's got to eat into Bread's private-label business. I think that's part of what we've seen in terms of their shift towards co-brand, but co-brand is a highly competitive space and creating a top-of-wallet product that competes with a JPMorgan Chase, Citi, those type of companies, financially, Bread just can't afford to do that. They're certainly in a vulnerable place from that perspective. They're in between these major banks on the co-brand side and then the buy now, pay later fintechs type of products on the other side. I know we haven't gotten to this yet, but the regulatory environment could change dramatically in two years. Bread would be in a very, very difficult spot.

[00:42:45] How do you think regulation or policy change might impact Bread going forwards? I understand it's going to depend on the different administrations, but what do you see as the worst-case scenario?

AJ [00:43:14]: Worst-case scenario, the cost of living becomes a major campaign issue and the Democratic candidate decides they're going to make a pledge to eliminate the late fee on credit cards, something that they want to do on day one of their new administration and then that candidate wins, then you're in a really tough spot. I went through this process where the company I was working with believed the late fees were going to be reduced. We went through an exercise where we had to find every way we possibly could to make up for that revenue. It even came down to things like adding a paper mailing fee to these products. I think that just tells you how important the late fees are to Bread Financial.

JK [00:44:32]: If we think through the impact of removing the late fees, that would be a big loss in revenue. That's the main thing there.

AJ [00:44:44]: Yes, huge.

[00:44:47] What would you encourage us to ask Bread's management team?

AJ [00:45:13]: I would ask them where they see themselves as being better than their larger competitors. I would want to have a great understanding of the deal structures that they have in place for their key clients, their biggest four or five partners that are so critical. I would want to really ask them about and test how durable those agreements are because in my mind, they're going to be very hard to replace if they were to ever lose those. Yes, I think from a technology perspective, Bread probably has a way to go. I would certainly press them on the talent they have from a technology perspective and then also from an analytical perspective as well.

[00:46:32] From what we've talked about, it seems as if Bread is sitting between the larger banks and the fintechs such as Affirm and Klarna. With that in mind, what do you think about the longer-term positioning of the company? Is its niche defensible over the next 5-10 years?

AJ [00:46:55]: Over the next 5-10 years, especially if they were to get another favourable administration, I think over the next 5-10 years, they're probably in decent shape. Part of that is I do think that they have really good relationships with their main key partners. I do think that they benefit from some companies not being willing to work with a Citi or a Synchrony because a competitor is already engaged with one of those companies. They naturally get some of that falloff. That can be really, really significant in some cases. They are positioned well there. Their strategy has clearly paid off over the last one year, 1.5 years. The stocks obviously performed very well. Yes, I think they should be fine over the next 5-10 years.

JK [00:48:03]: Is there anything that you would encourage us to spend more time on or dig deeper into as we're trying to understand Bread, the fraud strategy and the outlook going forwards?

AJ [00:48:29]: No, I think we covered it pretty well.

[00:48:34] Do you have any closing remarks to summarise the key points about Bread?

AJ [00:48:51]: Yes. The real opportunities from a fraud perspective with Bread are leveraging eCBSV to identify synthetic fraud before they book the accounts. The other big opportunity is improving their analytics so that they can develop better fraud acquisition fraud strategies to book more accounts and also stop more fraud. From a credit perspective, they've made a good move that's paid off here recently to shift and focus more on the co-brand products, but I think that introduces risk from a fraud perspective, obviously, with higher credit limits. AI is only increasing the fraud risk that they're going to face. Circling back to the late fees. I think that they've really benefited from the current regulatory environment. They put in a plan as though the late fees were going to go away or be reduced dramatically. When that didn't end up happening, they just left that plan in place. It's been a boom time for them right now, but again, they're positioned in a tough spot from a standpoint that the fintechs and the buy now, pay later are really well positioned on the private label side, and then they also face the challenges from the major co-brand issuers like JPMorgan Chase and Citi.

JK [00:50:55]: I just want to say thank you again for taking the time for the discussion today. I found it to be very helpful. Thank you. With that, I will end the recording.

AJ [00:51:05]: All right. Thank you.

Transcription ends at 00:51:08 of the recorded material