#11642 min
詐欺対策の皆さん、こんにちは！Fraud Forward へようこそ！
AI は詐欺の加速要因となっており、その影響は間違いなく 100％です。しかし、より大きなリスク、つまりすべての不正対策リーダーが夜も眠れなくなるほど懸念すべきなのは、キューが決して終わらないがゆえに、不正対策チームが戦略的に考えることをやめてしまうことです。
『Fraud Forward』パート2では、Synchrony Financialで銀行詐欺およびクレジット戦略担当バイスプレジデントを務めるテッド・ジョセフソン氏を再び迎え、戦略、AI、自動化、そして企業文化が交錯するエンタープライズレベルでの不正対策リーダーシップについて掘り下げていきます。
はっきりお伝えしておきますが、大規模な不正防止にはツールだけでは不十分です。必要なのは、不正対策のリーダーシップ戦略や、長期的な計画のための時間を確保できる不正対策チームのリーダーシップ、そして通常の不正のノイズと、本当の不正トレンドの変化を示すシグナルとを見分けられる戦略的な不正対策ソリューションなのです。
次に、AIについて率直にお話しします。AIが現在どこで実際の価値を生み出しているのか、業界がいまだに「実際には学習していないAI」を過大評価し続けているのはどこなのか、そして特にファーストパーティ不正のような一部の判断は、なぜ人間の責任と共感なしに完全自動化すべきではないのかを説明します。
そして不正対策に携わる皆さん、最後にお伝えするのは、いま何が過小評価されているのかというテッドの見解です――それは非常に大きなテーマです。エージェント型コマースによって、新たな説明責任と紛争対応の悪夢が生まれつつあり、さらに「フレンドリーフラウド」のような軽い気持ちで行われる不正が文化的な常態として広がっています。なぜなら、不正防止の進化は単なるテクノロジー導入プロジェクトではなく、銀行の不正対策リーダーシップと信用リスクのリーダーシップが実際に発揮される場だからです。これこそが不正防止のイノベーションであり、戦略的な不正対策の意思決定が求められているのです。
このエピソードでお届けする内容：
- 不正行為が絶えない中で、エンタープライズの不正対策チームがどのように長期的な戦略のための時間を確保しているか
- なぜ計画は「余分なもの」ではなく、他の重要な業務機能と同じように予定に組み込まれなければならないのか
- 通常の不正に伴うノイズと、システム上の弱点を示すシグナルを見分ける方法
- 短期的な急増と本格的な不正の転換点を分けるものは何か
- 一貫したケースストーリーや業務効率化を含め、AI が今日、実際に価値を生み出している領域
- 実際には学習していない「AI」を業界がいまだに過大評価している領域
- 特にファーストパーティ不正に関するような一部の判断は、人間の責任と共感なしに完全に自動化すべきではない理由
- エージェンティック・コマースが新たな責任問題と紛争の悪夢を生み出しかねない理由
- いかにして気軽な「フレンドリーフラウド」が当たり前になりつつあり、その文化的な変化がなぜ重要なのか
- 銀行の不正対策戦略と与信戦略のリーダーシップが、エンタープライズ不正対策チームの中でどのように交錯しているのか
このエピソードは次のような方におすすめです：
- 不正対策のリーダーとして取り組みを主導しているものの、緊急の対応に追われて本当に重要なことに手が回らないと感じている方
- 大規模な不正防止対策を担当しており、ノイズと実際の不正の変化を切り分けるための、より優れたフレームワークが必要だと感じている
- チームのために不正対策リーダーシップの育成に取り組んでおり、実践的で現場に即した戦略を求めている
- 企業の不正対策チームを統括しており、すぐに実務に活かせる戦略的な不正対策の知見を必要としている
- 銀行の不正対策部門の責任者としての役割を担い、自社の戦略的な不正対策ソリューションを検証したいとお考えの方
- 不正リスクや紛争と並行して、信用リスクに関する経営判断を主導している
- エージェント型コマースや紛争における責任の所在、そしてこの先何が起こるのかについて不安を感じている
- 「フレンドリー・フロード」が文化的な常識になりつつある状況を注視しており、それに対応するための不正対策リーダーシップ戦略を必要としている
このエピソードを気に入っていただけたら、ぜひ購読して、iTunes、Spotify、YouTube、またはお使いのポッドキャストアプリでレビューをお願いします。番組を多くの方に知ってもらううえで本当に助けになります。
エピソードノートと主なポイント
メモの内容に踏み込む前に一つだけ言わせてください。マーケティングチームから、人が私のポッドキャストを見つけやすくするために、これらのメモは特定の形式で構成しなければならないと言われました。
不正が絶えない中で戦略的な時間を守る
少し場の空気をリセットしましょう。
不正行為は容赦なく続きます。アラートは、あなたの計画カレンダーにも、四半期ごとの優先事項にも、経営合宿の日程にもお構いなしに鳴り続けるのです。
そのため、計画立案が「やってもやらなくてもいい仕事」とみなされると、キューに押しつぶされてしまいます。
テッドが示している、詐欺対策のリーダーたちが戦略的思考を守るためにできることは次のとおりです。
- 計画は副業ではなく、中核的な業務機能としてスケジュールしてください
- 長期的なリスク評価とシナリオプランニングのための保護された時間枠を確保する
- 業務のプレッシャーが高いときでも振り返りをせざるを得ないリズムを構築する
- リーダーシップの期待値をそろえ、戦略に充てる時間が交渉で削られるのではなく、守られるようにする
これは、実践における高度な不正防止と不正対策のリーダーシップです。
エンタープライズ向け不正対策ソリューションが永遠に後手に回ることを防ぐ、不正対策リーダーシップの戦略なのです。
不正のノイズと体系的なシグナル
ノイズとシグナルの違いについて、もう少し深く掘り下げたいと思います。
その違いが分からなければ、あなたは次のどちらかになってしまいます。
- 過剰反応してチームを振り回してしまう
- 逆に対応が甘くなり、組織的な弱点が根を張るのを許してしまう
テッドが、違いを見分けるための実践的な方法を解説します。
- ノイズは、季節性やキャンペーン、一度きりの要因に結びついた短期的なスパイクのように見えます
- シグナルは 繰り返し現れるパターン、チャネルをまたいだ一貫性、そして何度も再発する統制不備として現れます
短期的な急増と本格的な不正のシフトを分けるものは次のとおりです。
- 時間の経過にわたる継続性であり、単なる件数ではない
- 製品、チャネル、または顧客セグメント間での移動
- 犯罪組織ネットワークによる適応の証拠
- 既存の管理体制に明らかな負荷がかかっており、手動による審査や紛争処理も含まれています
不正対策に携わる皆さん、ひとつお伝えしておきたいのは、変化というのはいつも大きな音を立てて起こるわけではない、ということです。
そのサインは、企業の不正対策チームが、同じ結果を出すためにこれまでの倍の時間を費やすようになっている――そんなところに現れることもあります。
不正対策におけるAI：真の価値と過度な誇大宣伝
それでは本題に入りましょう。
今、詐欺対策のマーケティングではあらゆるところでAIが使われていますが、そのすべてが本物というわけではありません。
テッドが今日、本当に価値があると考えているもの：
- ばらつきを抑え、業務品質を向上させる一貫したケースストーリー
- 基準を下げることなく、チームに時間のゆとりをもたらす業務効率
- 判断を置き換えるのではなく補完として利用する場合の、意思決定ワークフローの支援
業界が「AI」を過大評価している点：
- 実際には学習していないものを「AI」と呼ぶこと
- 自動化を知能であるかのように扱うこと
- 誤検出や紛争、さらにはその後に生じる被害に対する責任を示さないまま、成果を過大に宣伝すること
ここで率直な不正対策リーダーシップの要点をお伝えします。
一部の判断は、完全に自動化すべきではありません。
特に第一者不正では、意図や状況、顧客への影響を踏まえ、人としての説明責任と共感が求められます。
それこそが不正対策チームのリーダーシップです。
それこそが戦略的な不正対策の意思決定です。
それこそが、主導権を握る銀行の不正対策リーダーシップです。
過小評価されているリスク：エージェント型コマースとフレンドリーフラウドの常態化
この部分は重要です。
テッドは、現在過小評価されていると考える2つのリスクを指摘しており、詐欺防止のイノベーションと進化には、誇大宣伝ではなくリーダーシップが求められるのはまさにこの点だと述べています。
エージェント型コマースと紛争の悪夢
エージェンティック・コマースは、責任の所在という問いそのものを変えてしまう。
次の場合、誰が責任を負うのか：
- 代理人が取引を開始する
- 消費者が自分はその取引を承認していないと主張している
- 加盟店または金融機関は、意図や同意を明確に立証できません
- 紛争処理のワークフローがグレーゾーンの申し立てであふれている
これは戦略的な不正対策ソリューションの課題であると同時に、銀行の不正対策戦略の課題でもあります。
つまり、エンタープライズ向け不正対策ソリューションが現実と向き合う局面なのです。
文化として定着したフレンドリーフロード
カジュアルな「フレンドリーフラウド」の増加は、単なる不正行為のトレンドではありません。
それは文化的なトレンドでもあります。
チャージバックや「承認していない」という申し立てが当たり前の行動として定着すると、最終的には次のような事態を招きます。
- 紛争件数の増加
- より逼迫したカスタマーサービスと不正対策業務
- 真の被害と意図的な不正利用の見分けがより難しくなること
- 実際の被害者を罰してしまうような過剰な是正を行うリスクが高まる
そして不正対策に携わる皆さん、これが今回のエピソードがこの結論に至る理由です。
不正防止は、単なるテクノロジーの導入プロジェクトではありません。
それはリーダーシップの鍛錬であり、規律です。
それは、不正対策業界での歩みを通じて育まれる、不正対策リーダーシップの発達です。
それは、不正ファイターとしてのキャリア成長が、不正ソリューションのリーダーシップへと発展していくことであり、さらに、不正対策を迂回するのではなく、不正対策と連携して進められるクレジット戦略のリーダーシップでもあるのです。
常に警戒を怠らず、情報をアップデートし続け、詐欺対策を前進させていきましょう。
Episode transcript
Ted Josephson
45:20
Have the best time. This is great.
Hailey Windham
45:23
Technically, technically you didn't leave, but we're splitting this up into two parts. So we're just we're just giving them more than they asked for[a].
Ted Josephson
45:31
I'm happy and anytime you get a fraud person, you get a chance to talk about your things with other people who enjoy fraud. It's a great day.
Hailey Windham
45:39
Right, because then you don't have to like explain the nitty gritty because they're like, what do you mean by implement a rule? And you're like, I just want to talk to somebody that knows what I mean. Okay, so we're moving on to talking about strategy versus firefighting. We talked a lot about, you know, crisis, things happen. You know, there are those times where it's fraud never sleeps. Right. And strategy requires oxygen. So how do you protect for long-term thinking.
Ted Josephson
46:11
we plan it out. Cause it's the nature of the corporate world, especially the larger your organization, the more overhead you kind of get in your lives. and all, and all, all good, you know, I don't want to disparage those who create overhead in our lives, but a fraud fighters role in your heart is you want to fight fraud. And so if I'm filling out another complaint for, for this or another toll gate for that or another, those do not bring joy. They bring, they bring controls, which is nice. So that being said, if you don't build the time in for your team and for yourself that's blocked out just like a meeting is, or just like a series of meetings are, to do that planning, you will always be fighting the fire. I just had my team in last week, a good chunk of my team, and I had the pleasure of having them all together. And we spent the time and we... generally turned off our phones and you know, there was some fraud fighting going on in the background, but in general, our focus was on what's our next steps related to this over this time frames and how are we building in these new tools and how do we incorporate new data points and things like behavioral monitoring into our existing strategy and how do you plan for that as opposed to, I have it. And I get a lot it. I'll live it all of its alerting do it and I don't know what it's doing But I hope it works what so yeah, it's it's about time and you know, that's what they pay us for it said But it's very easy to just become reaction all day long and look good because you're stopping fraud So you're able to report and do cool stuff but at some point that's gonna bite you because you're gonna take a massive fraud loss because you haven't been planning ahead and There's gonna be something you you didn't have ready and you couldn't fix in two days with something on the fly.
Hailey Windham
48:06
That resonates so much. can truly remember sitting there fighting fraud day in and day out going, okay, I'm stopping this one. I'm stopping that one. But then thinking, you know, a news article came out and it was like, Hey, we told you that this change was going to happen or this, you know, is the new mandatory requirement. I'm like, I'm not prepared because I wasn't expecting this or here's this new fraud trend that's coming out. That's attacking our, you know, online account takeovers. And I'm like, what is happening? Okay. Let me just. pivot, but then everything I was fighting before is going to get left to the side and we weren't prepared for anything new. I truly mean so much hearing that, but it resonates because I have been that fraud fighter that has had to be reactionary because I was in the weeds too much fighting the fraud and not being able to dedicate that time to prioritize different strategy.
Ted Josephson
48:57
And you have to make it so that your team knows it's a priority for each one of them. And if you are not assigning them tasks that help them to set aside time to grow, you know, I've seen some leaders who are so in the weeds and so wicked good at it that they won't let anybody else get good at it. There's other good people on their team, so they could step off and maybe do a little bit more of the other, but they're right in because they love it. It's their happy place. So it's hard. Remember I said it was hard to step away from the weeds.
Hailey Windham
49:14
Mm.
Ted Josephson
49:27
But do your team a favor, let them grow too and have that joy. I've done less fun things because I wanted my team to have the opportunity.
Hailey Windham
49:37
That's hard. It's hard.
Ted Josephson
49:39
It really, I will tell you, it really sucks sometimes. That's the only way to describe it. But because you love your team and you want what's best for them, it's like being a dad is hard. There's sometimes you do bad stuff that it's, you do it because you love your kids. I do it because I love my team and I love my company. I want it to do well and having it all about Ted, that's not the best of any.
Hailey Windham
49:51
Right.
Hailey Windham
50:03
No, no. But we do like having Ted on the podcast. So we are getting the best of Ted right now. But I will say, I think probably one of my absolute favorite moments in my career was when I had a frontline team member. She got it like it clicked and she called me so excited. I mean, she was jovial, you know, and she was like, Haley, I stopped a member from sending out $60,000 in a wire fraud. And I was like, what? Tell me all about it. And so I literally spent 30 minutes with her telling me the whole story. I wrote up a memo for the executive team to say what she did. And then like that got her additional like, you know, points or whatever they called because, you know, it was like they get credits for all the things that they sell to what we were talking about with like account points, but being able to prevent fraud too, started being a part of the point system as well because of that particular scenario. So that was just like such a good win. for Frontline and that it was one moment that proved that showing them how to do and teach a man to fish versus giving him a fish, you
Ted Josephson
51:11
and you incented that person, you gave them an incentive, they were excited and you doubled down on their incentive. So they will never forget that. You built a fraud fighter that day.
Hailey Windham
51:20
Yeah. For sure. For sure. I just got chills. I miss it so much. I miss being in the weeds like that, being able to work with a team and, you know, get my hands dirty in the fraud. But this is also it's rewarding on this side as well, obviously, because I'm talking to people like you, Ted.
Ted Josephson
51:26
you
Ted Josephson
51:37
Life is a box of chocolates. There's always a bunch of good choices, but you can't eat them all at once. So we make a variety.
Hailey Windham
51:43
No. Yeah. Okay. So speaking of variety, how do you prioritize under pressure?
Ted Josephson
51:52
Uh, well, if pressure is defined as a fraud event, what's the, what's the biggest risk? What's my biggest dollar amounts? Um, you know, where, where is it most likely to go off the rails? Uh, and then focus on that, you know, generally if, if we're having an event, uh, I try to figure out, we, we, we've got to figure out where it's coming from. You know, that's for sometimes you don't know what is going on. don't even know how this is happening. Uh, and so focus on what's the point of origin.
Hailey Windham
51:56
Yeah.
Ted Josephson
52:22
of what's going on. Okay, from there I can start to focus in my forensics on trying to stop it. But if I don't even know where it's coming from across my branch network or get my native app or my digital channel or the call center or mail, those are also sneak up on you, you can't control it as well. So at least you can put a bandaid on it and then move.
Hailey Windham
52:44
Yep, can definitely appreciate that. And then it leads me also to ask, how do know when rising fraud losses signal a systemic weakness versus just noise?
Ted Josephson
52:57
you know, we do forensics, know, we do post mortems and we try to look for commonalities in every fraud event that we have. And so we, and we track a lot of you know, we're tracking our alerts, we're tracking our loss avoidance. So we're tracking approval rates and decline rates and things go into manual review. When you start to notice an aberration, in one at the same time you're starting to get some noise or some interesting fraud losses here. We try to pull together to see is there something going on. But you got to stay on top of it. If you're not recording and tracking, if it's only just about losses, you'll wait until you're bleeding out the ice. If every time you're having a loss or at least a decent size loss, you try to get a good feeling for why did that happen? And if you start to hear that two or three or 10 times, you okay, now we got a problem. But it means having people who's willing to be in the weeds, good partnership with your operations teams because sometimes they're going to be seeing some of it and you won't. So, you know, our extended, you know, I have my fraud strategy teams and our analytics team and, we're deep dive on that, but we work seamlessly with our fraud operations team, which is the frontline guys working the alerts, you know, doing the individual, you know, analysis and things using all the other tools we've given them and we provide support. But we are, might as well be on our team. I just don't pay their payroll which is you know cheaper for me but you know somebody they get they're happy but yeah we treat them like they're our team so that they know they're loved and when they people are appreciated they they they give their best.
Hailey Windham
54:31
Right, right,
Hailey Windham
54:42
And I've noticed that they often collaborate more as well and more willingly. know, it's like, Hey, we, see something that's kind of odd. Whereas before they might feel closed off to just keep it siloed in their own group. But whenever you have that openness or that oneness of true, true team spirit, I don't know. I want to call us like all cheerleaders, all the, you know, I feel like that's kind of what we are as broad fighters, at least internally. But once, you know, once you remove that silo from your operations or you're from your front lines.
Ted Josephson
54:43
No, I'm not eating. Yeah.
Hailey Windham
55:11
It really does become this collaborative group effort of fighting fraud.
Ted Josephson
55:15
There's nothing more fun. you know, not every culture, you know, in terms of corporate culture encourages somebody from the frontline to ping me. I'm like, you know, I always would, we would do the like group town halls or different things like this. If I had a chance, I'd say, guys, my team's box is never full. You are welcome to ping me at any point. If I don't answer it's not because I don't love you or maybe because I'm on a call, but I will answer. I would love to hear from you. And if I can't help, I'll get somebody else to. And so a good day for me was when I had somebody I've never heard of before reach out to me and say, Hey, I heard you'd be helpful. And I try to help them on it. And that's, that was our teams. You know, so it was working with fraud teams, working with call center teams, wherever it might be. We wanted to be the accessible, helpful people. And I think we've achieved a pretty good balance on that.
Hailey Windham
56:04
That's our new slogan, Fraud Fighters, the accessible, helpful people.
Ted Josephson
56:06
That's We won't yell at you for calling.
Hailey Windham
56:12
Yeah, I love that so much. I'm probably going to make a sticker. We're put this on our laptops, on our cups everywhere. So anyways, that resonates. It really does. I feel like that's where most fraud teams probably miss the boat is not being accessible, not being reachable where it's like, put it in the queue and we'll get to it when we get to it. I know that at scale it's hard.
Ted Josephson
56:19
you
Hailey Windham
56:41
But at the same time, they've got to be able to reach someone when there is a five alarm fire in their minds. They need to at least to be able to have someone they can call.
Ted Josephson
56:52
Yeah, I mean, I just, I think just very recently, you know, I got a pass forward on one of those and I reached out to one of our senior VPs in another group for outside of fraud, or it was a fraud person, but working with our credit platforms and they could have easily just brushed me off. Let me send you to Susie Q customer service that, you know, was it going to do anything? No, no, no. They sent me to the senior person over here who did that nail thing. And we, we had it resolved in like 20 minutes. Because, again, fraud fighters care. So it's nice. I don't want to brag about this, but I think it's a great culture, the fraud teams.
Hailey Windham
57:29
When you have the mic, Ted, use it. mean...
Ted Josephson
57:32
Well, mean, okay, Grady is a great place. I think part of my favorite.
Hailey Windham
57:37
Yeah, yeah. Okay, so let's transition and talk about AI automation and the human edge, right? Where do you think has AI delivered real value?
Ted Josephson
57:50
to the fraudsters.
Hailey Windham
57:53
Yeah, right.
Ted Josephson
57:55
Sadly, think, boy, it's really made their lives a lot better. My text that I get soliciting me for stuff is much better and has deeper, richer data in it, much more targeted. So sadly, I think the processors are really benefiting from it. From what I have seen, I've seen some very interesting, I would still say it's still early because there's concerns about using PII within AI engines and appropriate. Concern for that you need to have that well done and you need to have a I think a relatively it's probably hard to say mature AI Infrastructure and governance to make sure that you do that. Well, and you're not accidentally going rogue on it that being said I've seen some really interesting things especially about rule strategy or model tuning where it's looking at the performance looking at all the data that's available looking making automated recommendations on how to make it better as opposed to Let me turn my analyst run 15 iterations of your best guess on what other things, features we can might add to this to make things better. So I've seen that. I think the, I'm very excited for what I'm seeing. There's solutions in market and some are still a bit nascent, but it's pretty, for writing up case histories. I mean, we all know that most case write-ups are colorful, you know, and usually on informative and SAR narratives. know, SAR narratives, you read one and it's got three lines in it, one that's got 50, and know, by two different people because everybody does their thing. So consistency of write-ups, inclusion of data, AI is perfect for that. Again, it's large language model. So I'm taking a bunch of data, turning it into a narrative that makes sense, it's consistently applied. I think that is transformative, super operational efficiency. oriented and you have a real person sign off on each one when it's done. But that's grunt work. Writing up a narrative is nobody's best and brightest day. And that is perfect AI process. I do think you will see more and better application of analysis for frontline people. This makes me a little nervous for frontline people. Because when you...
Ted Josephson
60:17
A lot of what they do is they look at disparate systems that aren't tied together well and they do ancillary things to make a best guess at fraud. Now we already know that there are solutions in process that will put an AI overlay on that, looking at all the data and providing a summary of it and maybe a best guess at fraud typology, if it is. It doesn't take a lot to extrapolate that into the future and say that that could have a material impact on it or it will require upscaling of the skills for the people who are currently doing that so that they can provide more value than just collating the data and making a best guess. They will have to provide a more expert level and it will require more of an expert level of skill for it. So. I'm trying to be careful on that one, but I think that's the reality. If my son or daughter was in that space, I'd be saying, you guys, you need to keep raising the bar. And it's not getting your certified fraud examiner title, it's not gonna help you in that. You need to jump a little harder.
Hailey Windham
61:28
Yeah, I've said this before, the only thing that scares me in the sense of AI coming on to the process for fighting fraud is that I fear for what it will do for the new people coming on. Like we do have some really good insights. An example, real world example that I'm trying to... to say here is speak in the words of fraud people, right? Or whatever. But if we think about the cash registers, back in the day, you would say the amount and they'd say $20.22 is your total or $19.22. They'd give a $20 bill, you'd hit $20 now and it tells you 78 cent, but then they go, wait, I've got 22 cent here or I've got two cent, I don't want the pennies. And they go, what? And it's like mind explode. Whereas back in the day, they would not even use the cash register at all and they'd calculate it all up in their head and then, you know, give you the correct change, you know, whatever. You could even add a Coke at the end and they'd know exactly what to do. And so I think the same thing. I'm fearful that we may be too dependent on those alerts that we would get on the frontline versus really empowering them as a fraud fighter to trust their instincts and to trust their gut. AI will never have a gut.
Ted Josephson
62:49
No, no, it won't. And we just, we just had that discussion because we were talking about applications of AI. This was with the team last week. And, uh, that's the downside is that you, you, you will potentially incent people to stop thinking the machine set. is it. Then this must be yet as opposed to no, no, no, you need to check to make sure the machine machine is right. And that's why I'm saying you need that upskilling and you got to pay more for that generally in the market. You know, expertise you need to pour into it.
Hailey Windham
63:03
Yeah.
Ted Josephson
63:19
So, yeah, I think you're spot on that most likely the simplest thing will be, and that may be what some companies want. You know, they just do what the machine tells them to do because I'm going to be consistent process, I'm going to compliant, I'm going to be able go to a regulator and show that there's no bias in anything I'm doing, it's just following this. So that may be actually a business goal. for some companies. And I'm not saying it's an irrational one. It's just not maybe where I'd go, but it's certainly not irrational.
Hailey Windham
63:59
Yeah, yeah. So again, we're talking about AI and automation. What decisions should never be fully automated?
Ted Josephson
64:01
you
Ted Josephson
64:14
Wow. Never. I'm inclined to say first party fraud. think you'll, until AI is a lot better than what I think, there is a certain amount of nuance and heavy moral responsibility when you tag someone as a fraudster, as opposed to identity theft. That if you're gonna do that, you should probably make sure that somebody's really looked at it and thought of the human aspect of it, did it meet this calculation? And that's ultimately what AI is gonna be doing. It's gonna heuristically check that this seems statistically likely that this is first run. Now, maybe super obvious, but I'd rather have a person double check that.
Hailey Windham
65:05
Yeah. Yeah. 100%. I know that I knew it was going to be a tough question by putting the word never in there, but definitely just wanted to kind of gauge your response there. So definitely appreciate that and agree. I think that there are so many wonderful things that the human aspect of fraud fighting really does. It brings to light the, obviously the empathy that we can have for a victim, the what happens afterwards.
Ted Josephson
65:09
you
Hailey Windham
65:34
how we can really work to prevent that from ever happening again versus just say, you're a fraudster, get out of here, you know, whatever. And then they go and they do it again at another FI or then they become an unbankable person. And then what good has that done for them and eventually the economy and, you know, banking as a whole.
Ted Josephson
65:58
All the romance scams, these lonely people who have fallen for it, clearly colluded, done really, in many cases, incredibly stupid things, and that they are personally responsible for it. But I don't think they're morally important people. They're sad folks. And I do think you need to have some heart in that, that would, I think, go beyond what AI's capable of.
Hailey Windham
66:24
I love that. I have been quoted saying AI doesn't have a gut. I think we're going to quote Ted saying AI doesn't have heart. Okay, if AI disappeared tomorrow, what would break first in fraud operations?
Ted Josephson
66:40
Currently, Nothing. Nothing. We, you we, uh, and I would say from what I know, I mean, I'd be interested in your perspective on this from what I know of most, most fraud shops. They don't have AI. They've got, they might have supervised machine learning. They've got, they've got a platform run by so-and-so who says it's AI, but it's not really, it's, it's, you know, it's, it's, it's ML. might be, you know, something like that. Uh, but, yeah, I don't think most shops would shut down. know some cutting edge ones might.
Hailey Windham
66:42
Nothing.
Ted Josephson
67:11
I know some that are let me see, anybody who's so reliant on AI that it would blow them up, who's so cool. I can't think of anybody who's like, it would really be a problem. I think you'd have some potential, they gotta hire some more people because they've been collating, calling some of the volume out with this, maybe a back office SAR group.
Hailey Windham
67:20
Yeah.
Ted Josephson
67:39
I think anybody who's probably using, I think Actimized has a star module that does some pretty efficient stuff and may have AI. I think they say it has AI in it. don't, you know, I'm not a definition guy, but if it does, that would probably mean they need more people.
Hailey Windham
67:55
Yeah. So my opinion is I feel the same way you do is that there's this misleading mindset in the industry is that we're, we are so AI right now in all of our fraud shops. And really it's like, I know what I see now on this side of things and I'm like, no, you don't have AI and you're not using that. And I can remember just being so frustrated with case management software and going, you said you were machine learning. I have answered the same freaking alert. How many times? And I've told you every time, not fraud, not fraud, not fraud. Why is this still alerting?
Ted Josephson
68:45
Nothing is learning. Nothing is learning until an analyst goes in and changes it. And I'm like, that's not learning. That's the model that runs until you fix it.
Hailey Windham
68:49
Exactly. Nope.
Hailey Windham
68:54
Yeah. So stop with your false advertisement, please. Because you are not AI and you are giving me a headache. Which then also, so I'm going to rephrase how I asked this question. Is today's AI making fraud teams stronger or just busier?
Ted Josephson
69:18
Again, because I don't think it's so embedded within most AI shops, I don't think it's having a material difference in most AI shops, or mean, op shops today or fraud shops today, other than possibly where they've applied it for large language models. So if they're doing case write-ups or SARS and things like that, which is the only stuff I've really heard somebody practically do, scraping call notes and doing that, I don't... I don't think there's anything that can have any material impact on them. I talked to Naftali from Sentilink and he was like, everybody's saying, AI is just revolutionizing fraud. goes, no, of stuff is not, especially on acquisition side, because the bad guys are still just taking all the clean data that they've stolen. They're just opening up new accounts and variations on it. And nobody has tools that will stop them from doing it. It's not AI. It's just. Large data analytics applied at scale.
Hailey Windham
70:17
Yeah, yeah. what a good conversation. think that I also think that this is one of those I love. I love uncomfortable conversations. I think that that's one thing that is the beauty of this podcast and the step I'm I'm trying to make with it, which is to say we're having these conversations anyways. And it's we need to make sure that the industry understands like don't be. Don't be lied to. Don't accept things at face value. Really, really look into our whole role as fraud fighters is to trust but verify. Why aren't we doing the verify aspect? We're just, we're just trusting a vendor at face value going, yeah, they have this, this and this. Why? Because their white paper said they had it or, did you really see it in action? Did you really, you know, have that?
Ted Josephson
71:04
or I was in a crisis and I needed something two weeks. So I'm just gonna slam this in and pray to God that's what it does. Most of the time that doesn't happen. Does it work that way? But it made it better. okay, we're happy for any incremental improvement we can get. And normally most of these things are about MoData. If I got MoData, can do better things. So we'll take what we can get.
Hailey Windham
71:07
Yeah.
Hailey Windham
71:27
BWAH!
Hailey Windham
71:33
that mo data. That's great. I'm literally just gonna go back through this entire transcript and just come up with a bunch of tedisms, okay?
Ted Josephson
71:44
I'm going to be colorful. That's my only goal.
Hailey Windham
71:47
I love it. love it. Okay. So let's shift and talk about the future of enterprise fraud. What risk is most underestimated right now?
Ted Josephson
71:58
and at enterprise level, just for fraud or just to, yeah. I think agentic AI is going to blow up, blow up things. I think there, there, I think there's a rush within the industry in the world. And we go through the hype factor. And if I hear the word agentic AI one more time, you know, this week I'm going to vomit, but, but, geez, you know, it's just, it's insane. But.
Hailey Windham
72:20
It's the new buzzword, Ted.
Ted Josephson
72:27
People are building it and but officers are using it. People are building it. And I think this turning over of the decision process, the customer turning that over to a machine. And then that's exposed to merchants. So credit, the banks and they're moving money around and doing all this stuff. It hits me as a huge legal nightmare because you you just need one case. After that goes on that says, well, it's an authorized transaction that I told my agentic AI to do. I told, I said, this type of thing I want you to do the first time it does one thing I didn't agree with. Well, that's not what I meant. It's not authorized. You know, and then you have that applied to fraudsters using that just like we do in malicious intent. You know, it's a, it, that to me feels like we could have a a really scale problem because suddenly now I've got, let's say 20 % of the country is using agentic AI to do their stuff. That means 20 % of the country could potentially, unless the laws get really good and very granular, and we're not great at that, suddenly 20 % of the country could say, I didn't mean to do that. I know that money went to Bitcoin. It wasn't me. It may have been me, but now I can dispute it at scale. for what, a month's worth of activity run by my agent at AI? And then I sued the property who sold me the Seneca. So I'm being very snarky here, but to me, it's always about the, I think a lot of the fraud that we have in the country. It's fueled by our laws and our regulations. We've inherently incented people to abuse it. Unfortunately, most people are honest. But the rules allow me to be a real bad guy at scale and nothing's gonna happen to me. In fact, I'll be encouraged. I'll be potentially in front of Congress said, this poor man was taken advantage of by that evil bank. Give him this $250,000 back. Okay. So yeah, I'm pontificating here, but I do think that...
Hailey Windham
74:45
Yep.
Ted Josephson
74:52
A lot of this could be fixed at a regulatory level. The fact that they let all of the social media solicitation of scams and fraud and won't do a thing. The only person who will get taken down if he complains about it is me. If I could, you know, because I'm harassing this innocent fraud site that's got thousands of members. I'm like, what? There's so much low hanging fruit out there and we're, nobody will do anything about it. And that, that makes me angry. I don't get angry a lot, but I'm like, that, that seems like if I was a politician, boy, talk about a way to really protect people. We'll get them to take down. Don't go after random bits of information. But if it's soliciting fraud and it's pretty obvious and it says, why don't you tell them to take that down?
Hailey Windham
75:41
Yeah. Here's your sign. This is fraud. So I have two thoughts after that statement. So one is, Zimar and I just did an episode where we talked about the agentic commerce that is the change that's coming in. That is super scary. And I just wanted to double click it. That's what you're referring to as underestimated. Whereas when I think agentic AI, I'm thinking also like the, where they will have these bots that will do low hanging fruit cases and things like that? Are you talking about both aspects of it or,
Ted Josephson
76:16
It's a venture commerce. Certainly, I think that there is tons and tons of opportunities once we can get out of our own model governance ways to revolutionize operations, call centers, the day to day, I need to do this transaction, I'm trying to do this. mean, heaven's best, everybody calls into a call center right now. I want you to do something for me.
Hailey Windham
76:18
Got it.
Hailey Windham
76:28
Yeah.
Ted Josephson
76:44
AI should be able to do that because most of what they do 90 % of the time what they're doing is a replicable thing that needs to follow policies and procedures and cut loose. You know, again, there's there's an automation aspect there that it might impact people. But yeah, I think that's that's going to be really revolutionary to the industry. And I don't know how that's going to play out. Because there's personal sides to it, really, super efficient and over time. Bailey, you think?
Hailey Windham
77:16
Yeah. I feel you. My second comment to everything that you said earlier too was, you know, in talking about the regulations and the things that we are governed by, I did a speaking presentation once at a conference and it was on Reg E. And I said, are we protecting consumers or enabling fraudsters? Like, what are we doing with Reg E? Because if I tell you it's fraud, you have to accept that. And then if it comes back as an authorized transaction, no, I'm telling you it's fraud. okay. Well, then I guess it's fraud. Well, here you go. And it's like, no, I proved without a shadow of a doubt, even going on as far as OSINT investigations to pull their Facebook and matching it up with their email address and what's associated in their phone number. And we're pinged and showing you were in fact in Florida for your birthday. And here are the hotel charges. And funny enough, Here's your Facebook photo and the photo that's on the website for this hotel that confirms you were where you said you weren't. What are we talking about? Like, come on.
Ted Josephson
78:26
Well, but again, that's what I'm saying. I'm not a big fan of over regulation, but I do think a lot of the regulation does incent behavior that as a society, we shouldn't incent. Should we protect the people who have been taken advantage of? Absolutely. But should there be some additional guardrails in there? know, there's the ability to non-repudiate. I think there's, you I'm not a big fan of national IDs and things like this, but what's our biggest problem in the country from a fraud stamp? I don't know who's on the other end of transaction. So, you know, and here we're having a moral dilemma within the country because we're saying for a, to vote, we want people to have a good ID. Now, whether it's whatever side you're on the aisle for, but again, I can't get on a plane unless I show an ID. And so there's lots of things where we go, yeah, this is cool. to hold a standard for doing banking transactions or financial transactions that there is a reliable form of identification that is unique to you and across the whole country at the very least. It's not like, this is great in Connecticut, but I'm California, nobody knows. You pray to God that it's the right idea. That seems funky for a country as sophisticated as we are. And it would solve a lot of problems if we knew who each other was from a transactional standpoint.
Hailey Windham
79:50
Absolutely. So, you know, we talked about the risk that's most underestimated. I'm curious, is that also what keeps you up at night or is that something different?
Ted Josephson
80:01
no. you know, again, I'd say that the stuff that keeps me up at night is just the, the casual fraud, the friendly fraud. It's the first, I think it's, it's a statement about us as a country. we're more and more people when I, when I talk to a young person who I like and who makes a casual comment about stealing, like, know, but you know, yeah, I told you, I just told you, you know, the Amazon, didn't get a Walmart. did this. There is it, there is a casualness to low level fraud, that 10 years ago, I didn't hear people talking about it. And I hear over dinner now. You know, like I'm out with people and they say, oh yeah, I did that. just, you know, I'm like, wow. You know, or they didn't charge me for this. So I didn't say anything. You know, it's like a big thing. And you're like, really? Like, seems okay. And they say, yeah, so I got it for free. And I'm like, okay. So there seems to be, and it's probably, you my old guy perspective, you know, like these kids. But it feels like a fundamental change. The difference between like the 80s to the 2000s, you didn't hear me say anything about that. But in the last five, 10 years, it seems like it's okay to screw somebody else, especially if it's a big company. If it's a big company or a bank, they deserve it.
Hailey Windham
81:20
Yeah, I've seen that.
Ted Josephson
81:22
So that keeps me up and that's why I pray.
Hailey Windham
81:25
Yeah, well, for sure. I'll tell you when those conversations happen at my dinner table, you, I get my husband and my 14 year old look at me simultaneously, like, don't say anything. And I'm like, I want to tell him. Now, if it's, if it's other kids, my daughter, she'll hop on them and she goes, you're committing fraud. And I'm like, that's my girl. That's right. You tell them, you tell them.
Ted Josephson
81:48
here one more person, it was a hack. I'm like no it's check fraud.
Hailey Windham
81:52
Yeah, no, just so you know, that's going to come back and bite you on the butt. You know, they're going to know what the account is tied to you. So great job with that hack, I guess. So that's what keeps you up at night. What excites you about the future?
Ted Josephson
82:09
you know, I, I'm, I'm excited about the, the opportunities to get, and again, it probably goes back to the AI side. So much of what we do requires really good smart people to go through reams and reams of data and, and, and reanalyze it constantly and relook at different things to take the grunt work out of that. Like, like to have some, have things running that, make it easier to highlight, abnormalities, you know, I love the idea of some of these unsupervised machine learning models, which I think I'm surprised haven't really taken off or done as well as they should, but I think it's an AI application where I'm looking at all the data that's going on from the bank, your everything like dump everything into this thing. And it should be grouping into these, these are good behaviors, groups of normal behavior. And anytime something kind of whacks out of that, you would say, you know, Ted, you don't understand analytics, but I'm like, no, you know, we should be able to have good sets of normal data and outliers should pop. anytime you and I have both caught fraud or seen fraud caught, it was an outlier. And you go, why didn't somebody notice this? Boy, that seems like a great application. Finally, we'll go for, UMLs didn't do it, but AI sounds like it could.
Hailey Windham
83:29
Love that. Love that. Okay. If you could redesign how the industry measures fraud success, what would you change?
Ted Josephson
83:31
Thank
Ted Josephson
83:37
I would, if we could get better at valuing, I mean, loss avoidance is, is, know, finger in the air stuff, but it's, it's really important. And there's a, there's the valuation of the fraud I didn't let happen and its impact on people or in the fraud I did happen. It's, it's the impact to the, to the, you know, to show you think of the noble. You know all the people that weren't trafficked all the people that upster drugs that weren't done all the only All that horrible stuff that goes on That is enabled by this every dime we removed from them is a huge Moral victory for us in protecting the innocent And I wish there was a less of a just to focus on the dollars and cents as opposed to I saved a million dollars today. my heavens
Hailey Windham
84:21
you
Ted Josephson
84:33
I helped keep a million dollars out of evil people's hands who were going to do evil things with it. like celebrate not just the money, but the bad that didn't happen.
Hailey Windham
84:44
Yes, 1000 % yes. I had a fraud leader the other day call me to say she spent 40 minutes on the phone with a potential victim trying to tell them that, hey, this is following a scam pattern and they just wouldn't believe it, wouldn't believe it and tried to give all the back data or whatever to say how it wasn't fraud. she said, okay, we're just not going to process it today. Let me get back to you in the morning. And the guy was like, okay, well, like, I'm okay if I lose the money, like, it's not going to make her break me. And she said, I appreciate that. But what I'm more concerned of is where the funds would go if it were going to a fraudster. And that mind shift or that mindset shift that happened with that potential victim did so much in that conversation with her. So I completely appreciate that perspective and know that it is definitely something that we need to consider.
Ted Josephson
85:38
Well, I'm sure you've had these conversations with executives saying we can take some more losses on this, which we can't. These are business decisions, it's true. But the, we've got plenty of room in the budget. So let's loosen things up so that we can take more losses and achieve this good goal. But nobody talks about the fact, okay, so I'm funding organized crime this much more with my budget here? that, there's. There's a societal impact to that, that it's not just the bottom line. So again, I'm probably a bit of a Pollyanna here, but I have never heard that perspective. Like, you know, I'm morally troubled by allowing more fraud because horrible things will happen with it, but I'm okay with it because I'm making more checking accounts. Okay, well at least say it out loud. I'm getting more at this so that I've never heard that outside of a fraud-fighter's room.
Hailey Windham
86:32
for sure. just so you know, Pollyanna is where I got my name from. So her name was Haley Mills, and she was my Nana's favorite actor. That's who I was named after. Just an FYI. Yeah. Okay, last question. know, and it's...
Ted Josephson
86:38
really?
Ted Josephson
86:45
Wonderful. Okay, here we go.
Hailey Windham
86:54
One that's like, what advice would you give fraud professionals who want to reach executive roles or as they are preparing for this next era?
Ted Josephson
87:03
you know, the, the, the biggest challenge of a true fraud fighter is we're true speakers and we work within a culture typically that does not value true speaking. It values politically correct speaking. It values nuanced speaking. And so you can speak to truth and love as a fraud fighter and be vilified for it. You you're doing all the right things because they say, this is what we want you to do. So, you know, for any fraud fighter.
Hailey Windham
87:20
Thank
Ted Josephson
87:32
And I've talked to so many of my peers on this. The biggest challenge is not ticking people off by speaking the truth. And so finding a way to do the right thing, but communicate it in a way that doesn't create an inordinate pushback. You need to show that your intent is so good that the truth that you're speaking, that you're reluctantly speaking this thing, because even though you want this other thing, you need to try to protect. So it's a hard one. I have had success. I have had train wrecks and a lot of it depends on who you're talking to. So you have to constantly adapt. it's really manage your communications well. As fraud fighters, we talk to each other very easily and truthfully and because we get in that habit. It's hard for us when we're suddenly in a meeting with a bunch of people that aren't like that to quickly adjust and remember. So just stay on your game. Remember who you're talking to.
Hailey Windham
88:36
Yes, 100%. Ted, thank you again so much for coming on the podcast with me today. It meant more than you could ever know.
Ted Josephson
88:39
Thank
Ted Josephson
88:45
This is, Hallie, this is a treat. I love and appreciate you. I love what you do. I love this field. I love the people who work in it. And a chance to be a part of that just makes my day. So thank you. This is a wonderful way to end my day.
Hailey Windham
89:01
Absolutely. So, fraud fighters, if you listen to both parts of this conversation, you probably felt it. This wasn't just a discussion about fraud losses. It wasn't just about AI. It wasn't even just about enterprise scale. This was about leadership, real leadership, the kind that understands fraud doesn't live in a silo. It intersects with credit, with growth, with customer trust, with culture. And at that level, every decision ripples. Ted, when I think about this conversation, what stands out isn't just the strategy, it's the steadiness, know, the clarity, the willingness to evolve, the intentional promotion of your team. Because at the end of the day, fraud prevention is not a technology program, it's a leadership discipline. And whether you're leading fraud at a national institution like Synchrony or at a, you know, $500 million community bank, the principles are the same. Think long-term, even when the alerts are loud. Protect your people while protecting the portfolio and never forget that fraud prevention is about preserving trust. So to everyone listening, especially those who aspire to lead at that level someday, this is your reminder. The path from analyst to executive isn't about knowing more rules. It's about thinking bigger, collaborating wider and carrying the weight of risk with both discipline and humility. Ted, again, thank you for giving us a transparent look inside Enterprise Fraud Leadership. And to the fraud fighters out there, fraud never sleeps. The threats will keep evolving, the pressure will stay high, but so will we. We lead, we adapt, we think strategically, and we continue to move fraud forward. See you next time.
Ted Josephson
90:34
Amen. Take care.
View full transcript
