WEBVTT

NOTE Sentence-level transcript of https://www.youtube.com/watch?v=l0FLhNqBOic

NOTE One cue per sentence. Cue ids are the line anchors on /transcripts/l0FLhNqBOic.html. A cue ends where the next begins, or 2 s after its last word.

s1
00:00:01.309 --> 00:00:03.309
[music]

s2
00:00:12.720 --> 00:00:14.800
All right, first and foremost, thanks so much for being here.

s3
00:00:14.800 --> 00:00:22.360
Um it's uh been a great experience, you know, obviously chatting amongst uh other industry giants like Cursor and Factory and Anthropic.

s4
00:00:22.360 --> 00:00:26.280
I'm sure you guys are mostly here for them, but thanks for sticking around for this talk.

s5
00:00:26.280 --> 00:00:27.000
Uh my name is Varick.

s6
00:00:27.000 --> 00:00:28.200
I'm the CEO of Varick Agents.

s7
00:00:28.200 --> 00:00:34.440
We work with some of the largest companies on the planet transforming them from the inside out with uh AI and agents.

s8
00:00:34.440 --> 00:00:39.960
Um and because of the nature of our work, which is highly bespoke, we go very deep into our clients.

s9
00:00:39.960 --> 00:00:42.400
It requires a lot of forward-deployed engineering.

s10
00:00:42.400 --> 00:00:55.360
And this conversation is around why that's so important, how we approach it at Varick, uh and some of the internal tooling that we've created internally to allow us to scale that forward-deployed motion without, you know, increasing head count exponentially.

s11
00:00:55.360 --> 00:01:04.320
And it's titled the next bottleneck because I fundamentally believe the next bottleneck is how deep can you go into a customer without scaling head count uh exponentially?

s12
00:01:04.320 --> 00:01:07.440
How can AI do that job for you?

s13
00:01:07.440 --> 00:01:11.760
So, as stated previously, AI is solving the execution of work.

s14
00:01:11.760 --> 00:01:23.880
Um if you were to look back a couple of years ago before uh you know, thinking agents and reasoning agents were a widespread phenomenon, uh and I had asked you, "How many of you have used AI to solve an end-to-end task?"

s15
00:01:23.880 --> 00:01:26.120
The answer would be slim to none.

s16
00:01:26.120 --> 00:01:31.200
But if I ask that same question to everyone today saying, "Has AI solved an end-to-end task for you today?"

s17
00:01:31.200 --> 00:01:33.120
I'm sure every single one of you would raise your hands.

s18
00:01:33.120 --> 00:01:36.600
So, clearly execution is no longer the core bottleneck.

s19
00:01:36.600 --> 00:01:46.760
The models are improving to the point where intelligence is no longer the constraint, and harnesses are being built in a way that allow us to use, whether it's browser use tooling or API tooling,

s20
00:01:46.760 --> 00:01:52.880
uh with very robust MCPs that allow us to execute work uh with near perfection.

s21
00:01:52.880 --> 00:01:59.640
The difference and the bottleneck that is still here is how much can you understand the business?

s22
00:01:59.640 --> 00:02:03.000
Because every business, every consumer is different.

s23
00:02:03.000 --> 00:02:09.479
Uh one sales department for a healthcare company, for example, operates completely differently than the sales department for a SaaS company.

s24
00:02:09.479 --> 00:02:11.800
And this is something that we see with our work today at Verkada agents.

s25
00:02:11.800 --> 00:02:20.560
And if there are any business operators in the room, you know exactly how hard it is to wrangle uh the latest models to solve for your specific use cases.

s26
00:02:20.560 --> 00:02:25.640
It's very difficult to extract that context from your you know, employees and from your team.

s27
00:02:25.640 --> 00:02:32.680
And it's very difficult to feed that into an API call or a simple model call uh that doesn't break down very quickly.

s28
00:02:32.680 --> 00:02:38.840
So, the bottleneck is how much can you process re-engineer, how much can you process understand?

s29
00:02:38.840 --> 00:02:40.840
And that's the job that we do here at Verkada.

s30
00:02:40.840 --> 00:02:44.959
So, right now operations are fundamentally centered around the human.

s31
00:02:44.959 --> 00:02:50.840
Uh right now, the work that you do today is done by humans, whether it's on top of software or completely agnostic to software.

s32
00:02:50.840 --> 00:02:53.840
Uh but in the future, operations will be centered around AI.

s33
00:02:53.840 --> 00:03:06.480
And this means not only, you know, providing companies with AI tooling like a cursor, Claude Code, Codex, like a factory, like any of the other brilliant AI tools that you for your experiencing here today,

s34
00:03:06.480 --> 00:03:10.320
but also changing the operations and the processes themselves.

s35
00:03:10.320 --> 00:03:13.600
And fundamentally, that is the role of a forward deployed engineer.

s36
00:03:13.600 --> 00:03:20.640
It's going into the company, understanding how things run today, and re-envisioning what it could look like tomorrow.

s37
00:03:20.640 --> 00:03:26.840
And we believe that is our job at Verkada and why forward deployed engineering is such a core part of what we do.

s38
00:03:27.560 --> 00:03:31.360
So, a forward deployed agent, what does that really mean?

s39
00:03:31.360 --> 00:03:32.560
So, why do we need FDEs?

s40
00:03:32.560 --> 00:03:34.640
As I stated previously, I'm not going to, you know, go into this too much.

s41
00:03:34.640 --> 00:03:36.360
I'm sure you've been hearing a lot of this today.

s42
00:03:36.360 --> 00:03:38.519
Uh FDEs are responsible for a few different things.

s43
00:03:38.519 --> 00:03:40.959
One is they map the way that humans are doing their work today.

s44
00:03:40.959 --> 00:03:45.640
So, how we do that at Verkada is several forward deployed engineers will be embedded directly with the customer.

s45
00:03:45.640 --> 00:03:51.680
You can imagine it's a enterprise company with thousands of employees, but we'll scope it down to a single department.

s46
00:03:51.680 --> 00:04:06.800
In a finance department, for example, we'll have them sit down with the process leads for AP, AR, card reconciliation, banking, billing, FP&A, etc. So, interviewing every single one of these process leads to understand not only how are things running today,

s47
00:04:06.800 --> 00:04:10.200
but more importantly, when things go wrong, what happens?

s48
00:04:10.200 --> 00:04:18.320
You know, a lot of the documentation that you have at companies about the golden path and maybe an edge case or two, but this is still fundamentally not the reality.

s49
00:04:18.320 --> 00:04:28.040
Where when we talk to customers, it's it's a lot of, you know, Sarah in AP handles the workflow today in this way, but when things go wrong, she actually sends it over to Chris,

s50
00:04:28.040 --> 00:04:34.240
who then takes four days of cycle time to handle reconciliations between a purchase order and an invoice.

s51
00:04:34.240 --> 00:04:49.280
Those are the realities that are one unique to every single company, the way that they handle things is different from one company to the next, and two, the real bottleneck for why AI can't just run amok and handle end-to-end processes without the hand-holding that you see today in the enterprise.

s52
00:04:49.280 --> 00:04:53.560
Um so, this is the first section, which is mapping how humans do the work.

s53
00:04:53.560 --> 00:04:57.160
The second is really re-engineering the process around AI.

s54
00:04:57.160 --> 00:04:58.640
So, what does this mean?

s55
00:04:58.640 --> 00:05:07.800
You know, there's a lot of talk being given today in terms of slapping AI onto broken processes, and that's fundamentally why you don't see the ROI across the industry today.

s56
00:05:07.800 --> 00:05:16.919
There's a lot of, you know, semi-outdated, but still very relevant statistics like the MIT review saying that 95% of generative AI pilots fail to reach production.

s57
00:05:16.919 --> 00:05:27.160
Or the other statistic, which is 87%, very similar thing, that most AI pilots don't produce measurable ROI or they don't ship to production, period.

s58
00:05:27.160 --> 00:05:37.360
And the reason for that is a lot of the time, AI is being slapped on top of broken processes in a way that the AI doesn't actually understand how to do things.

s59
00:05:37.360 --> 00:05:47.360
Um you see that this at the simplest level with coding, where as an engineer, it's very difficult, even with gold loops and the the latest technology there, uh to just

s60
00:05:47.360 --> 00:05:54.960
say, "Go and solve this for me and have it run off and refactor entire code bases without some degree of human input.

s61
00:05:54.960 --> 00:06:12.880
Now, if you extrapolate that to a business context, these are very non-technical operators in finance, sales, marketing, procurement, logistics, uh etc. So, giving them this AI tooling will not allow them to receive the same ROI that a software engineer might be able to to uh produce or create.

s62
00:06:12.880 --> 00:06:19.520
Um so, what this means is you need forward-deployed engineers to help them re-engineer their current process around AI.

s63
00:06:19.520 --> 00:06:24.320
It needs to be not too different to where they don't understand, you know, how to operate the system.

s64
00:06:24.320 --> 00:06:34.640
For example, if they're used to an 11-step workflow and you come in and change that with a one-step, they might be taken aback, the adoption rates might suffer, etc. As was alluded to in previous presentations.

s65
00:06:34.640 --> 00:06:39.800
Um but at the same time, it needs to be different enough to where you're actually capturing the ROI.

s66
00:06:39.800 --> 00:06:45.680
Meaning, you do say, all right, four out of these eight steps will be handled completely autonomously.

s67
00:06:45.680 --> 00:06:49.440
The other three will be handled with some human-in-the-loop intervention.

s68
00:06:49.440 --> 00:06:53.560
And one step of that process will be handled by a human, period.

s69
00:06:53.560 --> 00:07:02.680
Either because the risk is too high or because, you know, it's it's not unique enough for an agent to produce measurable value in that specific step of the process.

s70
00:07:02.680 --> 00:07:04.880
So, that's the second major step of a forward-deployed engineer.

s71
00:07:04.880 --> 00:07:06.760
That's why we need them.

s72
00:07:06.760 --> 00:07:15.160
And third and finally, and this is what I want to uh bring one of my heads of engineering to discuss in just a moment, is actually deploying these agents on top of existing systems.

s73
00:07:15.160 --> 00:07:21.640
So, fundamentally at Verity, we believe that the AI wave left a lot of enterprise behind.

s74
00:07:21.640 --> 00:07:23.800
A lot of enterprises are married to their systems of record.

s75
00:07:23.800 --> 00:07:26.200
Not everybody, but most of them are.

s76
00:07:26.200 --> 00:07:32.320
Uh they've migrated to NetSuite, they've migrated to Dynamics, uh they migrated to SAP and Salesforce.

s77
00:07:32.320 --> 00:07:39.560
And when you pitch them AI solutions that live completely disparate from these systems, uh you're ignoring the reality of enterprise.

s78
00:07:39.560 --> 00:07:47.080
One of the quotes from our clients said that they spent $5,000,000 and 5 years migrating to uh NetSuite.

s79
00:07:47.080 --> 00:07:48.160
That's a real quote.

s80
00:07:48.160 --> 00:07:54.320
So, if you're telling them, "Hey, I have this fancy AI tooling, but by the way, you have to migrate off of NetSuite," they're going to tell you to get out.

s81
00:07:54.320 --> 00:07:56.880
They don't have any appetite for that.

s82
00:07:56.880 --> 00:08:03.120
So, what we believe in Varick believe in at Varick is we'll build the agents on top of your systems of record.

s83
00:08:03.120 --> 00:08:04.880
And the way that we do that is quite uh unique.

s84
00:08:04.880 --> 00:08:12.200
We have our own Varick OS platform that allows us to spin up agents, monitor them, uh etc. with the full governance and eval suite baked in.

s85
00:08:12.200 --> 00:08:14.560
But at the same time, it lives on top of your systems of record.

s86
00:08:14.560 --> 00:08:20.160
So, if you are on a Salesforce or a NetSuite or a Dynamics or an SAP, we will not ask you to migrate off of that.

s87
00:08:20.160 --> 00:08:23.840
And that is where enterprise needs AI the most.

s88
00:08:23.840 --> 00:08:27.320
Uh because they're too large to move up.

s89
00:08:28.200 --> 00:08:31.040
So, why build the FDE agent in the first place?

s90
00:08:31.040 --> 00:08:38.159
As mentioned previously, I think everyone is saying that 2026 and onwards is the year of the forward deployed engineer.

s91
00:08:38.159 --> 00:08:40.240
And to some extent, we believe that's completely correct.

s92
00:08:40.240 --> 00:08:48.040
There's never been more of a need to go deep into customers and understand their business use cases and help them adopt the latest in AI tooling.

s93
00:08:48.040 --> 00:08:59.520
But at the same time, we realize that it's actually very difficult to find forward deployed engineers who are both the technical like top 1% who are really able to understand and speak AI

s94
00:08:59.520 --> 00:09:10.560
uh 10,000 times better than the average, you know, enterprise customer, but also have the communication and human skills needed to be, you know, as alluded to previously, very high IQ, high EQ,

s95
00:09:10.560 --> 00:09:14.839
extracting the information from the customer and meeting them where they are in real time.

s96
00:09:14.839 --> 00:09:24.120
You know, typically, you'll have consultants that you then train on the technical side or engineers that you then kind of train on the soft skill side, but it's very hard to find people who are,

s97
00:09:24.120 --> 00:09:25.839
you know, the best of both.

s98
00:09:25.839 --> 00:09:32.640
Um so, the FDE agent is our effort to bolster the existing forward deployed engineers that we do have.

s99
00:09:32.640 --> 00:09:41.480
So, for example, allowing one forward deployed engineer or forward deployed strategist to manage and maintain several client communications.

s100
00:09:41.480 --> 00:09:48.560
I know that most of the folks in the room are technical, but it's very easy to misunderstand how deeply involved you have to be with the client.

s101
00:09:48.560 --> 00:09:51.320
They're emailing you 24/7.

s102
00:09:51.320 --> 00:09:54.560
They're sending you hundreds of pages of documentation.

s103
00:09:54.560 --> 00:09:57.720
And every single process they will pull you in a different direction.

s104
00:09:57.720 --> 00:10:05.240
AP relies on AR relies on reconciliation, relies on FP&A, and they each have their own version of what they think is the most important.

s105
00:10:05.240 --> 00:10:15.480
So, being able to manage that context and being able to serve them all equally while also not hiring 50 people to do so is fundamentally very important and it's how we at Verek avoid being,

s106
00:10:15.480 --> 00:10:25.800
you know, a traditional consultancy while also offering that handheld hand-holding and like very human experience human-centered approach of consulting that we think we do think is very valuable.

s107
00:10:25.800 --> 00:10:32.160
Um so, in the past in 2024 and around that time, execution work was still the bottleneck.

s108
00:10:32.160 --> 00:10:39.200
This was before the models gained the intelligence and the harnesses gained the integration abilities that allowed them to move past the execution bottleneck.

s109
00:10:39.200 --> 00:10:43.040
Um now the AI models are trained to solve the execution of knowledge work.

s110
00:10:43.040 --> 00:10:48.400
I will go as so far as to say that knowledge work is almost entirely solved.

s111
00:10:48.400 --> 00:10:56.440
The difference is and what we're realizing now is that designing how work gets completed around AI is the next bottleneck.

s112
00:10:56.440 --> 00:11:08.839
It's the ability to go deep with the customer, redesign the workflows, deciding what should be automated versus shouldn't, and building this in a robust and scalable way on a platform that moves the needle for our clients.

s113
00:11:08.839 --> 00:11:16.680
You know, as opposed to doing a point solution, which promises to transform just one part of your sales process for example, uh maybe it's prospecting.

s114
00:11:16.680 --> 00:11:20.680
Uh that ROI might deliver 5 to 10% ROI for you as a sales function.

s115
00:11:20.680 --> 00:11:21.600
Same thing on finance.

s116
00:11:21.600 --> 00:11:27.160
If you're just doing AP and no other part of your department, you might have a 5 10% ROI.

s117
00:11:27.160 --> 00:11:32.560
But, at Verek we deliver department-wide transformations, holistically transforming the entire department at a time.

s118
00:11:32.560 --> 00:11:43.200
And that's how we get the ROI that we see for our clients which is 25% 50% 75% truly giving them back you know the three things which is revenue uplift

s119
00:11:43.200 --> 00:11:48.040
cost savings and risk mitigation as was so eloquently stated previously.

s120
00:11:48.040 --> 00:11:53.120
And an AI FDE is trained to re-engineer these necessary tasks around AI.

s121
00:11:53.120 --> 00:12:05.520
So I want to invite my head of engineering JD Pruitt to come up and and share some more of the deep technical stuff on our FD agent because I haven't written a line of production code in a while.

s122
00:12:05.520 --> 00:12:06.360
So here's JD.

s123
00:12:06.360 --> 00:12:07.200
Okay.

s124
00:12:07.200 --> 00:12:09.000
Thanks Voss.

s125
00:12:09.000 --> 00:12:12.680
And maybe we can get his mic going.

s126
00:12:15.200 --> 00:12:16.600
All right, thank you.

s127
00:12:16.600 --> 00:12:17.120
Thanks Voss.

s128
00:12:17.120 --> 00:12:23.320
So this project to give tools to our FDEs basically started with me.

s129
00:12:23.320 --> 00:12:25.840
I lead the platform team and we're over on one side of the office.

s130
00:12:25.840 --> 00:12:28.080
We're hanging out, we're chilling, having a great time.

s131
00:12:28.080 --> 00:12:29.640
Codex, Claude, we're all hanging out.

s132
00:12:29.640 --> 00:12:31.680
And then I look over at the FDE side of the room.

s133
00:12:31.680 --> 00:12:33.080
They look stressed.

s134
00:12:33.080 --> 00:12:34.280
They are sleep deprived.

s135
00:12:34.280 --> 00:12:35.160
They're extremely miserable.

s136
00:12:35.160 --> 00:12:36.920
They've got clients emailing them 24/7.

s137
00:12:36.920 --> 00:12:38.640
I'm like, "Oh my god, you guys haven't slept at all."

s138
00:12:38.640 --> 00:12:42.160
So I go and I start talking to them and I'm like, "Okay, what is your guys' process like right now?

s139
00:12:42.160 --> 00:12:43.800
How are you actually engaging with these clients?"

s140
00:12:43.800 --> 00:12:53.880
They're like, "Well, we you know upload about 150 pages of documentation to Claude and then we prompt Claude and then we wait like 2 minutes and then we get analysis and then it's verbose and incorrect

s141
00:12:53.880 --> 00:12:54.680
and it kind of sucks."

s142
00:12:54.680 --> 00:12:56.960
And I was like, "All right, we got to fix this."

s143
00:12:56.960 --> 00:13:02.560
So we have been working on an FD agent which is the Codex for our FDEs basically.

s144
00:13:02.560 --> 00:13:04.640
And there is three stages of it.

s145
00:13:04.640 --> 00:13:07.800
The last of which is certainly still in in development.

s146
00:13:07.800 --> 00:13:10.280
The first is what we call the engagement.

s147
00:13:10.280 --> 00:13:13.640
The first function is the is the engagement agent.

s148
00:13:13.640 --> 00:13:18.040
And essentially this is a better version of Claude built just for our FDEs.

s149
00:13:18.040 --> 00:13:19.000
It's their assistant.

s150
00:13:19.000 --> 00:13:21.240
They have it pulls in their granola notes.

s151
00:13:21.240 --> 00:13:22.720
It synthesizes documentation.

s152
00:13:22.720 --> 00:13:23.920
It reads PowerPoint slides.

s153
00:13:23.920 --> 00:13:27.480
It allows them to query and say, "Who is responsible for this process?

s154
00:13:27.480 --> 00:13:33.760
I got an email that mentioned Sarah spelled, you know, this way and I have a, you know, a Slack message with a different way.

s155
00:13:33.760 --> 00:13:34.960
Are these the same people?

s156
00:13:34.960 --> 00:13:41.880
Cuz these are the questions that our FDEs are asking all day, every day, and they waste a ton of time just waiting on on Claude to respond.

s157
00:13:41.880 --> 00:13:47.800
So, the engagement agent is their way of It's their It's their assistant to build um to build the workflow.

s158
00:13:47.800 --> 00:13:49.440
Then there's the workflow agent.

s159
00:13:49.440 --> 00:13:58.040
And what we did is we took our engagement agent and we embedded it inside of our platform so that when our FDEs go and they actually build the workflow,

s160
00:13:58.040 --> 00:14:02.880
the FDE agent is right there saying, "Oh, you forgot about this edge case.

s161
00:14:02.880 --> 00:14:08.960
Um you should probably ask me uh you know, who owns this process so I make sure the email goes to the right place."

s162
00:14:08.960 --> 00:14:18.840
And it lives It's basically a I can talk a little more of the details on the next slide, but Whoops.

s163
00:14:18.840 --> 00:14:21.800
It lives uh it lives on side of our inside of our platform.

s164
00:14:21.800 --> 00:14:35.120
It works next to Claude or Codex, whatever model you're using, um and make sure that the workflow that the FDE is constructing is actually uh it correctly shadows the process that we want to engineer.

s165
00:14:35.120 --> 00:14:49.440
And then there's the final stage, which we're not at yet, which is a um an autonomous assistant for FDEs where it's receiving emails from clients who say, "Actually, I want to change, you know, where my QC report goes to.

s166
00:14:49.440 --> 00:14:58.280
I want to, you know, change it to a different email, etc." And our agent is able to process that information, query the understanding of the company that we currently have,

s167
00:14:58.280 --> 00:15:11.560
ship an autonomous change to the workflow on top of our platform, and then our FDE never has to get involved, saving their time for the much more high-value work of sitting down, interviewing with the clients, really understanding what their process is,

s168
00:15:11.560 --> 00:15:20.640
um and not dealing with all of the small little minutia that anyone who has been an FDE can tell you uh takes up a lot of their time.

s169
00:15:21.920 --> 00:15:23.800
So, how do we how do we build this?

s170
00:15:23.800 --> 00:15:30.440
Um the first thing is we need some single source of truth, some representation of a company's uh functioning.

s171
00:15:30.440 --> 00:15:32.040
There's a lot of different ways to do this.

s172
00:15:32.040 --> 00:15:40.000
If you were at the booths downstairs this morning, there was, you know, five companies trying to sell you a graph DB, and you can just use Postgres, whatever it is.

s173
00:15:40.000 --> 00:15:41.440
Yes, I'm looking at you.

s174
00:15:41.440 --> 00:15:42.555
Um

s175
00:15:42.555 --> 00:15:42.560
[clears throat]

s176
00:15:42.560 --> 00:15:46.000
doesn't really matter what you use, but the point is we use a dependency graph.

s177
00:15:46.000 --> 00:15:49.520
Um most of these workflows inside of enterprise are remarkably linear.

s178
00:15:49.520 --> 00:15:56.880
They just have a lot of cycles in them, but at the at the end of the day, the process owners want things to be as dependency-driven as possible.

s179
00:15:56.880 --> 00:16:02.120
They don't want person C in the process to have to deal with something before A and B have approved it.

s180
00:16:02.120 --> 00:16:06.240
So, a dependency graph is a very nice representation of this.

s181
00:16:06.240 --> 00:16:09.120
Um then we do our own model training.

s182
00:16:09.120 --> 00:16:12.520
Um and there's really two parts to this problem that we're trying to solve.

s183
00:16:12.520 --> 00:16:20.400
The first is, given extracted context for the FTE, do we get a good high-quality output?

s184
00:16:20.400 --> 00:16:24.160
And the answer is, with Claude, honestly, no, which is kind of surprising.

s185
00:16:24.160 --> 00:16:35.280
But the really um and sure you guys have experienced this, when you are trying to do a long analysis, frontier models are extremely verbose, and they lack the um

s186
00:16:35.280 --> 00:16:42.440
you know, I would say the um I only started believing in consultants once we started hiring them at Verek.

s187
00:16:42.440 --> 00:16:50.440
And the reason is they're so good at figuring out what is the part of the detail the client actually cares about, and what is the part that can get glossed over.

s188
00:16:50.440 --> 00:16:53.040
And frontier models have absolutely no concept of this.

s189
00:16:53.040 --> 00:16:57.600
So, we started post-training our own models on top of on top of open-source models.

s190
00:16:57.600 --> 00:17:07.560
We're a fan of Kimiko 26, but a lot of these would do a lot of these would do fine to really get that nice balance between detailed and uh clarity

s191
00:17:07.560 --> 00:17:10.319
that the frontier models often often lack.

s192
00:17:10.319 --> 00:17:16.000
So, that's a that's a bit about writing a good normalized process flow from extracted context.

s193
00:17:16.000 --> 00:17:24.079
But there's the second half of the challenge, which is getting good at traversing at extracting the right context.

s194
00:17:24.079 --> 00:17:33.120
So, we might have this huge knowledge graph, but it's remarkably difficult to traverse this knowledge graph in a reliable way that finds us the right context.

s195
00:17:33.120 --> 00:17:42.600
So, once we have our post-trained model, we create an RL environment where we have exposed our own custom tools specifically designed to traverse our knowledge graph.

s196
00:17:42.600 --> 00:17:53.480
These tools are things like make sure person A and person B are actually the same person because a lot of you know, there's a lot of Mikes in every company we work with and Claude gets very confused by this.

s197
00:17:53.480 --> 00:18:04.560
Um a second thing might be something like um uh identifying identifying redundancy cycles or uh like violations of your DAG inside of your knowledge graph.

s198
00:18:04.560 --> 00:18:11.720
And so, in our RL environment, we train really good tools to do a good job of traversing this graph to extract the right context.

s199
00:18:11.720 --> 00:18:13.880
So, that's how we solve the two problems.

s200
00:18:13.880 --> 00:18:18.360
Writing good analysis from the context and extracting the correct context in the first place.

s201
00:18:18.360 --> 00:18:33.080
And then the third part, which we are um still building towards is an agent that operates autonomously to do the kind of uh workflow management on the small things that the FDA doesn't have to waste their time on.

s202
00:18:33.080 --> 00:18:33.880
I think that's all I've got.

s203
00:18:33.880 --> 00:18:35.520
I'll hand it back over to Vass.

s204
00:18:35.520 --> 00:18:37.360
Thanks, Petey.

s205
00:18:37.360 --> 00:18:39.000
So, where does that leave us?

s206
00:18:39.000 --> 00:18:44.440
And I want to share a little bit more about Verica uh because obviously we're not the cursor and throttle of the open eye of the world.

s207
00:18:44.440 --> 00:18:55.200
Um when we started this company, we fundamentally believed that the way the puck was moving, you had to get ahead of it and you had to start learning how a business runs

s208
00:18:55.200 --> 00:18:57.280
and building with that in mind.

s209
00:18:57.280 --> 00:19:04.680
I think a lot of Silicon Valley starts to go product product product, but what we're building for cannot be solved for with just a product.

s210
00:19:04.680 --> 00:19:13.480
We start off every single engagement with an audit where we actually send our forward deployed engineers, strategists into a company to learn how it works from the inside out.

s211
00:19:13.480 --> 00:19:16.040
That is what I believe is the biggest bottleneck.

s212
00:19:16.040 --> 00:19:18.200
And after that, we go into implementation.

s213
00:19:18.200 --> 00:19:20.280
We build agents on top of our platform.

s214
00:19:20.280 --> 00:19:27.440
And yes, you do still need all the bells and whistles and the fancy technology that allows us to, you know, really automate work in the future.

s215
00:19:27.440 --> 00:19:36.480
But, again, the bottleneck is the forward deployed motion, which is why we are so bullish here at Verica on our AIFDE.

s216
00:19:36.480 --> 00:19:44.920
And if you're interested in learning more about it, or if you're interested in joining one of the fastest-growing startups in Silicon Valley working with some of the largest clients on the planet,

s217
00:19:44.920 --> 00:19:48.800
come find us after and we'll have a chat because we are aggressively hiring.

s218
00:19:48.800 --> 00:19:58.760
And if you are a company looking to understand how AI can really move the needle for you internally instead of just slapping a frontier model on top of everything and watching it break in production,

s219
00:19:58.760 --> 00:20:00.160
come find me after as well.

s220
00:20:00.160 --> 00:20:00.960
Thank you all so much for the time.

s221
00:20:00.960 --> 00:20:01.760
I really appreciate it.

s222
00:20:01.760 --> 00:20:03.903
And cheers.

s223
00:20:03.903 --> 00:20:05.903
[applause]

s224
00:20:20.278 --> 00:20:22.278
[music]