1
00:00:00,000 --> 00:00:07,000
So fraud sweep, so the main part we were just talking about there, is primarily for sort

2
00:00:07,000 --> 00:00:10,000
of small community banks or mid-banks as well.

3
00:00:10,000 --> 00:00:13,000
Well, there's no reason it couldn't go up.

4
00:00:13,000 --> 00:00:15,000
Right, I mean, it could scale up.

5
00:00:15,000 --> 00:00:16,000
Marketing and business-wise.

6
00:00:16,000 --> 00:00:21,000
Right, a lot of the customer base has historically been of that nature.

7
00:00:21,000 --> 00:00:28,000
And so, you know, on the business side of things, what's it like selling to banks?

8
00:00:28,000 --> 00:00:33,000
Well, I mean, I've sat in on some of the sales, but obviously Greg does most of it.

9
00:00:33,000 --> 00:00:41,000
It's, banks are an interesting market because they don't, there's no competition in banks

10
00:00:41,000 --> 00:00:44,000
in how we would normally conceive it.

11
00:00:44,000 --> 00:00:51,000
It's such a regulated market, more or less, if you get in there and you don't, as long

12
00:00:51,000 --> 00:00:56,000
as you do what the regulators say to do, you're pretty much guaranteed to make money.

13
00:00:56,000 --> 00:01:02,000
And even if you mess up, they're guaranteeing all the funds and insuring all the funds,

14
00:01:02,000 --> 00:01:04,000
so you can't even really lose that much.

15
00:01:04,000 --> 00:01:05,000
Right.

16
00:01:05,000 --> 00:01:12,000
So there's not a lot of drive to be different.

17
00:01:12,000 --> 00:01:20,000
And in fact, so, well, actually, so what happens is you get, you've got the market pretty much

18
00:01:20,000 --> 00:01:22,000
starts from a point of stagnation.

19
00:01:22,000 --> 00:01:25,000
Then a new technology, a new process will come along.

20
00:01:25,000 --> 00:01:27,000
Some of the fringe early adopters pick it up.

21
00:01:27,000 --> 00:01:33,000
If it makes sense for the industry, it'll start to get some momentum after they prove out the value for it.

22
00:01:33,000 --> 00:01:36,000
Then you'll have kind of a second wave of adopters.

23
00:01:36,000 --> 00:01:38,000
Eventually it moves to mainstream.

24
00:01:38,000 --> 00:01:41,000
Then the market completely saturates and then stagnates again.

25
00:01:41,000 --> 00:01:46,000
So you have these periods of fluidity where kind of product waves sweep through.

26
00:01:46,000 --> 00:01:47,000
Right.

27
00:01:47,000 --> 00:01:51,000
And then you'll have stagnation for sometimes years at a time where all the vendors,

28
00:01:51,000 --> 00:01:56,000
all the people that aren't the banks themselves are simply living off support contracts,

29
00:01:56,000 --> 00:01:58,000
waiting for the next product to come through.

30
00:01:58,000 --> 00:02:04,000
Then they'll either develop such a solution themselves, buy up or make a partnership with someone else that has.

31
00:02:04,000 --> 00:02:05,000
Right.

32
00:02:05,000 --> 00:02:09,000
Then try to go in, lock up as much of the market as they can before the stagnation happens.

33
00:02:09,000 --> 00:02:14,000
And then they do that and then eventually it's stagnation and then you're done.

34
00:02:14,000 --> 00:02:20,000
It's very hard within the market to dislodge an existing supplier for a particular product.

35
00:02:20,000 --> 00:02:21,000
Yeah.

36
00:02:21,000 --> 00:02:24,000
So this same thing has happened a number of times.

37
00:02:24,000 --> 00:02:32,000
It happened in the '80s and early '90s with digitizing checks.

38
00:02:32,000 --> 00:02:35,000
So check scanners and automating that process.

39
00:02:35,000 --> 00:02:42,000
And then it happened in the late '90s, early '00s with online banking.

40
00:02:42,000 --> 00:02:43,000
Right.

41
00:02:43,000 --> 00:02:46,000
And it is happening with fraud detection now.

42
00:02:46,000 --> 00:02:51,000
So I mean, I think Greg, and that's really what Greg was trying to do was time the market to kind of hit this wave.

43
00:02:51,000 --> 00:02:54,000
And that's what you do in banking is you say, what's the next product?

44
00:02:54,000 --> 00:02:59,000
Okay, we need to get in late enough in the process that we can actually make sales somewhere.

45
00:02:59,000 --> 00:03:00,000
Right.

46
00:03:00,000 --> 00:03:04,000
But early enough that we're before the crest of the wave and we have time to build up

47
00:03:04,000 --> 00:03:09,000
so that by the time the wave crests, we're big enough to be noticed by someone that can buy us out

48
00:03:09,000 --> 00:03:16,000
and help us inject into the thousands of banks all around America.

49
00:03:16,000 --> 00:03:17,000
Right.

50
00:03:17,000 --> 00:03:23,000
So it sounds like more than being sort of a kind of a tight loop of boom and bust

51
00:03:23,000 --> 00:03:28,000
or even an innovation sort of driven way of deploying technology in banking

52
00:03:28,000 --> 00:03:33,000
that it's more of, there's sort of a collective, all right, it's time to upgrade our technology.

53
00:03:33,000 --> 00:03:38,000
And you've got to time the window when the industry, the banking industry as a whole thinks that

54
00:03:38,000 --> 00:03:41,000
now's the time to upgrade instead of now's the time to lock down everything.

55
00:03:41,000 --> 00:03:42,000
Yeah.

56
00:03:42,000 --> 00:03:44,000
Yeah, I think that's correct.

57
00:03:44,000 --> 00:03:47,000
The upgrades are always driven though by some external factor.

58
00:03:47,000 --> 00:03:51,000
If the banks didn't have to upgrade, they would never do it just to upgrade the system.

59
00:03:51,000 --> 00:03:52,000
Right.

60
00:03:52,000 --> 00:03:56,000
Because, you know, obviously they're not, banks aren't, as much as they may rely on IT,

61
00:03:56,000 --> 00:04:02,000
they would be happy processing everything manually if they could make money off it.

62
00:04:02,000 --> 00:04:03,000
Right.

63
00:04:03,000 --> 00:04:07,000
So the two reasons they're great is the growth in the customer sizes

64
00:04:07,000 --> 00:04:11,000
and the customer base has become substantive that they simply need a new product

65
00:04:11,000 --> 00:04:16,000
because old processes are simply imploding in trying to handle today's volumes of things.

66
00:04:16,000 --> 00:04:21,000
And that's been the driver of a lot of these factors.

67
00:04:21,000 --> 00:04:24,000
And then the other factor is simply regulatory issues.

68
00:04:24,000 --> 00:04:31,000
Like the next thing is you have to do a lot of BSA stuff, which is Banking and Syncrasy Act.

69
00:04:31,000 --> 00:04:34,000
So you have to look for the name Bin Laden on all your accounts.

70
00:04:34,000 --> 00:04:35,000
Ah, right.

71
00:04:35,000 --> 00:04:39,000
And banks have no way of doing that, you know, so they buy products to do it.

72
00:04:39,000 --> 00:04:44,000
And so now a lot of BSA products are starting to emerge in the market.

73
00:04:44,000 --> 00:04:49,000
We're actually looking at ways to add BSA tests to our engine

74
00:04:49,000 --> 00:04:53,000
because it's a very similar type of question as fraud detection.

75
00:04:53,000 --> 00:04:54,000
Right.

76
00:04:54,000 --> 00:04:56,000
And so that's kind of the next wave.

77
00:04:56,000 --> 00:05:02,000
Fraud actually came out because fraud loss, typically the bank just got insurance and ate it all.

78
00:05:02,000 --> 00:05:03,000
They didn't care.

79
00:05:03,000 --> 00:05:07,000
They would pay everything and take a few losses and turn them over to their insurance company

80
00:05:07,000 --> 00:05:08,000
and we paid out.

81
00:05:08,000 --> 00:05:11,000
And so it was a controlled cost for them.

82
00:05:11,000 --> 00:05:16,000
They were still making money, so they could maybe make a little bit more if they took the effort,

83
00:05:16,000 --> 00:05:17,000
but they'd rather just not take the effort.

84
00:05:17,000 --> 00:05:18,000
Right.

85
00:05:18,000 --> 00:05:23,000
Then they started to lose more on fraud for various reasons.

86
00:05:23,000 --> 00:05:31,000
And so fraud actually became a bigger loss than bad loans, which was a tremendous turning point in the industry.

87
00:05:31,000 --> 00:05:34,000
It's the type of thing no one ever thought would really happen.

88
00:05:34,000 --> 00:05:39,000
And once that did, then they had to start thinking of ways to control fraud a little bit better

89
00:05:39,000 --> 00:05:46,000
simply because it was outrunning them and the insurance companies wouldn't pay as much fraud as they were having,

90
00:05:46,000 --> 00:05:47,000
so they had to find alternatives.

91
00:05:47,000 --> 00:05:48,000
Right.

92
00:05:48,000 --> 00:05:52,000
So they were really driven to it by the fact that they were already losing money on it.

93
00:05:52,000 --> 00:05:53,000
Right, right, right.

94
00:05:53,000 --> 00:05:58,000
And so what are the kinds of fraud that people commit in banks?

95
00:05:58,000 --> 00:06:05,000
Probably one of the most damaging and common is a kiting where you go to one bank, you open an account,

96
00:06:05,000 --> 00:06:06,000
you deposit some money.

97
00:06:06,000 --> 00:06:13,000
You go to another bank, you open an account, you deposit money and you have to use checks for all this from the first bank.

98
00:06:13,000 --> 00:06:19,000
And so if you start with the initial $1,000, let's say, then you deposit $1,000 here for real.

99
00:06:19,000 --> 00:06:21,000
You write a check, deposit $1,000 over here.

100
00:06:21,000 --> 00:06:26,000
Then you run back to the first bank, write a check for $1,000 from bank B.

101
00:06:26,000 --> 00:06:29,000
Now you've got $2,000 in this bank because this one hasn't hit yet.

102
00:06:29,000 --> 00:06:35,000
And you go back to the bank A, now you write a $2,000, now you've got a $3,000 balance in this bank.

103
00:06:35,000 --> 00:06:38,000
And really, and so you're kind of making--

104
00:06:38,000 --> 00:06:39,000
It's just a matter of the timing.

105
00:06:39,000 --> 00:06:42,000
Yes, by exploiting the timing differences.

106
00:06:42,000 --> 00:06:46,000
It's getting harder to do, but it's still done.

107
00:06:46,000 --> 00:06:50,000
You know, then there's just plain old--

108
00:06:50,000 --> 00:06:54,000
One thing we're seeing a lot of, actually, is you can get check printing software and all that kind of stuff.

109
00:06:54,000 --> 00:06:57,000
So if you can get someone's--

110
00:06:57,000 --> 00:07:01,000
Well, when I say someone, all fraud happens on commercial accounts.

111
00:07:01,000 --> 00:07:07,000
You know, there is the identity theft stuff isn't really the big loss for banks.

112
00:07:07,000 --> 00:07:15,000
I mean, it's terrible when it happens to a person, but in the industry as a whole, 98% of it is all on corporations.

113
00:07:15,000 --> 00:07:16,000
It's not through identity theft.

114
00:07:16,000 --> 00:07:21,000
They just-- There's so much money moving through it, and it's so big that no one notices the $50,000 check.

115
00:07:21,000 --> 00:07:22,000
Right.

116
00:07:22,000 --> 00:07:27,000
And so what we find are actually checks, and we're just flagging them for the banks.

117
00:07:27,000 --> 00:07:33,000
When they-- After we flag them and they catch them and they pull it up and do a little bit of research, often it's laughable.

118
00:07:33,000 --> 00:07:43,000
It'll be the wrong size check that was printed on an inkjet printer that, you know, has some random serial number on it because they don't know, you know, those things about the checks.

119
00:07:43,000 --> 00:07:44,000
Right.

120
00:07:44,000 --> 00:07:49,000
And it's ludicrous, but no one has time to look at them all.

121
00:07:49,000 --> 00:07:50,000
Right.

122
00:07:50,000 --> 00:07:58,000
So someone might find a, I don't know, a check from big company XYZ, and they duplicate the check and then write a check out to themselves.

123
00:07:58,000 --> 00:07:59,000
Yeah.

124
00:07:59,000 --> 00:08:02,000
Or they print a check out to themselves or whoever and try to deposit that check.

125
00:08:02,000 --> 00:08:03,000
Yeah.

126
00:08:03,000 --> 00:08:05,000
And steal money from the commercial company.

127
00:08:05,000 --> 00:08:06,000
Yeah.

128
00:08:06,000 --> 00:08:07,000
By virtue of a fake check.

129
00:08:07,000 --> 00:08:08,000
Yeah.

130
00:08:08,000 --> 00:08:09,000
I mean, you get more sophisticated things.

131
00:08:09,000 --> 00:08:22,000
I mean, you get-- Certain companies get set up as scam companies and they do forms of kiting or, you know, they'll pretend like they have a lot of, you know, they'll like, they'll deposit a bunch of banks, you know, false checks.

132
00:08:22,000 --> 00:08:24,000
They'll actually get a good history with the bank.

133
00:08:24,000 --> 00:08:25,000
Yeah.

134
00:08:25,000 --> 00:08:27,000
And then pretend like they got a bunch of income.

135
00:08:27,000 --> 00:08:34,000
They'll try to withdraw it all right away before the bank learns that all the, you know, all these checks that they deposit are really just fake.

136
00:08:34,000 --> 00:08:35,000
Oh, right, right, right.

137
00:08:35,000 --> 00:08:41,000
So that's kind of a limited kiting, but you have to set the company to do that and, you know, take the time to establish credentials.

138
00:08:41,000 --> 00:08:42,000
Right.

139
00:08:42,000 --> 00:08:51,000
But, I mean, we can spot things like that and we have, I mean, all of our, most of our banks have found something large that, you know, more than pays for the system at the time.

140
00:08:51,000 --> 00:08:52,000
Right.

141
00:08:52,000 --> 00:08:59,000
But the legendary rounding error fraud, does that ever happen, you know, as most recently in office space, right?

142
00:08:59,000 --> 00:09:00,000
Right.

143
00:09:00,000 --> 00:09:02,000
To do that, that's more of an interest system thing.

144
00:09:02,000 --> 00:09:08,000
So there, that would be someone committing, from within the bank committing fraud, which isn't.

145
00:09:08,000 --> 00:09:10,000
We're looking at external actors.

146
00:09:10,000 --> 00:09:11,000
Right.

147
00:09:11,000 --> 00:09:12,000
Perpetuating fraud in the bank.

148
00:09:12,000 --> 00:09:18,000
And at that point, there's really no opportunity for that kind of, for that kind of fraud.

149
00:09:18,000 --> 00:09:29,000
We did, I'm trying to remember if we actually implemented it or not, there is some theories about how numbers should group in actual payments.

150
00:09:29,000 --> 00:09:39,000
And so you can do analysis on numbers and a lot of times if people are, like, making up reimbursements for themselves, the numbers will be too random.

151
00:09:39,000 --> 00:09:43,000
Be like, you know, very often price is in the 99, right?

152
00:09:43,000 --> 00:09:44,000
Right.

153
00:09:44,000 --> 00:09:51,000
So that means you expect 99 or you expect some, you know, certain things depending on the taxes that were being charged.

154
00:09:51,000 --> 00:09:52,000
Right.

155
00:09:52,000 --> 00:09:55,000
But you don't expect 41 to come up a lot.

156
00:09:55,000 --> 00:10:02,000
So a lot of people, when they're just, like, making up reimbursement expenses will, like I said, the numbers end up being too random.

157
00:10:02,000 --> 00:10:06,000
We actually do have some tests that can analyze account history for that kind of thing.

158
00:10:06,000 --> 00:10:12,000
It sounds like there's a little crossover as far as spam filtering and as far as just building up a historic record.

159
00:10:12,000 --> 00:10:13,000
Oh, absolutely.

160
00:10:13,000 --> 00:10:14,000
And just matching patterns.

161
00:10:14,000 --> 00:10:20,000
And when something's out of the ordinary, then you just, you know, in this case, you know, in the case of spam filter, you stick it in the spam filter.

162
00:10:20,000 --> 00:10:23,000
In this case, you ask a human to intervene and make a decision.

163
00:10:23,000 --> 00:10:24,000
Right.

164
00:10:24,000 --> 00:10:25,000
Yeah.

165
00:10:25,000 --> 00:10:29,000
And a big part of the program is actually learning more about the customer.

166
00:10:29,000 --> 00:10:35,000
So we'll learn who writes big checks regularly, what to expect, and then start accepting those out of the tests.

167
00:10:35,000 --> 00:10:36,000
Right.

168
00:10:36,000 --> 00:10:40,000
So if the customer writes a huge mortgage check every month and that's their only big check, we do a couple things.

169
00:10:40,000 --> 00:10:46,000
One, we allow them to say, you know, don't, if you see this once a month, that's fine, just pass it.

170
00:10:46,000 --> 00:10:48,000
If you see it more, obviously, flag that.

171
00:10:48,000 --> 00:11:01,000
And also, you know, like, the system's smart enough to then take that out of the big item calculation so that we don't get skews for these one-off things.

172
00:11:01,000 --> 00:11:02,000
Right.

173
00:11:02,000 --> 00:11:09,000
So, I mean, there is a lot of sophistication kind of underneath, but at the end of the day, it's really just kind of doing a lot of reasonable checks

174
00:11:09,000 --> 00:11:15,000
on every single item that the bank has no way of doing and then just saying, hey, I think you should look at this.

175
00:11:15,000 --> 00:11:18,000
Then it's up to a person to come in and actually, I mean, we can't tell them it's fraud.

176
00:11:18,000 --> 00:11:19,000
Right.

177
00:11:19,000 --> 00:11:20,000
There's no way we can know that.

178
00:11:20,000 --> 00:11:26,000
But the person will come in, look at the original check image, maybe call the customer if they think, ask them if they actually sent it,

179
00:11:26,000 --> 00:11:28,000
and do the later investigation.

180
00:11:28,000 --> 00:11:34,000
But what we're doing is we're pointing out the items that they would never notice, and we're only pointing out those items.

181
00:11:34,000 --> 00:11:35,000
Right.

182
00:11:35,000 --> 00:11:40,000
Yeah, it's also similar to everyone experiences this with their credit card company every now

183
00:11:40,000 --> 00:11:45,000
and then where, you know, you can't charge something and you call them up and they just tell you there was some weird activity.

184
00:11:45,000 --> 00:11:46,000
Right.

185
00:11:46,000 --> 00:11:47,000
You have buying behavior that's out of the ordinary.

186
00:11:47,000 --> 00:11:48,000
Exactly.

187
00:11:48,000 --> 00:11:51,000
Like, buy three refrigerators in one day and then some gas, right?

188
00:11:51,000 --> 00:11:52,000
Yeah.

189
00:11:52,000 --> 00:11:53,260
Or something crazy like that.