Meeting Title: Contextual Platform Updates and Demos Date: 2025-10-29 Meeting participants: Samuel Roberts, Uttam Kumaran, Mustafa Raja, Rajiv Shah, Hannah Wang, Casie Aviles, Holly Condos, Henry Zhao, Awaish Kumar
WEBVTT
1 00:00:20.870 ⇒ 00:00:21.630 Samuel Roberts: Hello.
2 00:00:21.630 ⇒ 00:00:22.210 Mustafa Raja: Hey.
3 00:00:22.810 ⇒ 00:00:23.570 Uttam Kumaran: Hey.
4 00:00:26.720 ⇒ 00:00:29.560 Samuel Roberts: Yeah, I think this one will go to the platform, right? Yeah.
5 00:00:33.290 ⇒ 00:00:34.460 Rajiv Shah: Alright.
6 00:00:36.190 ⇒ 00:00:38.730 Rajiv Shah: Let me know when I should start.
7 00:00:40.510 ⇒ 00:00:42.610 Uttam Kumaran: Yeah, is everyone here on our side at Zoom?
8 00:00:44.470 ⇒ 00:00:45.950 Samuel Roberts: Holy…
9 00:00:50.320 ⇒ 00:00:56.210 Samuel Roberts: Moving here over the waist… 20…
10 00:01:00.970 ⇒ 00:01:03.379 Samuel Roberts: Holly is also on mute?
11 00:01:04.349 ⇒ 00:01:06.309 Uttam Kumaran: Let’s get started, we don’t have much time, so…
12 00:01:06.310 ⇒ 00:01:07.040 Samuel Roberts: Okay.
13 00:01:08.060 ⇒ 00:01:08.840 Rajiv Shah: Alright.
14 00:01:10.220 ⇒ 00:01:25.460 Rajiv Shah: So, I’m gonna go through… I got a couple slides to, like, highlight some of the new, kind of, updates and things that are coming inside the platform. If I have time, I can show a couple of them in the app as well, so… and feel free to ask your questions as well as we go through this.
15 00:01:25.740 ⇒ 00:01:26.590 Rajiv Shah: Okay.
16 00:01:27.100 ⇒ 00:01:41.400 Rajiv Shah: I think a big piece and kind of change over the last year is moving away from our earlier, like, RAG 2.0, if you first saw Contextual, to now much more of, like, a realization that really where we fit is this context layer.
17 00:01:41.630 ⇒ 00:01:56.649 Rajiv Shah: And so this is a layer between, kind of, the data, where it sits, your snowflakes, your Databricks, SharePoint, wherever that data sits, and then all the great AIs, right? There’s gonna be lots of LLM models, AI models, the Anthropics, the OpenAIs.
18 00:01:56.650 ⇒ 00:02:07.709 Rajiv Shah: But those things by themselves don’t solve problems. You need to kind of integrate that, orchestrate that with data to solve problems. And that’s where kind of contextual is fitting in like that.
19 00:02:09.880 ⇒ 00:02:11.529 Rajiv Shah: Hopefully that makes sense.
20 00:02:11.700 ⇒ 00:02:16.980 Rajiv Shah: Yes. So, if you think about that journey that we’re on, Oops.
21 00:02:17.860 ⇒ 00:02:20.060 Rajiv Shah: My slides are going backwards here.
22 00:02:21.090 ⇒ 00:02:28.059 Rajiv Shah: We started off by really focusing on extraction, cleanly getting that information out, traditional RAG,
23 00:02:28.100 ⇒ 00:02:41.899 Rajiv Shah: of being able to ask questions, and over time, we added some other pieces where, right, lots of multimodal inputs. We’ve added connectors now for things like SharePoint, for example. I think Confluence is coming up soon, we have Box.
24 00:02:42.630 ⇒ 00:02:45.760 Rajiv Shah: A lot of those RAG tools, like multi-turn.
25 00:02:45.840 ⇒ 00:03:02.689 Rajiv Shah: We added groundness score that you can see in the platform, right? An instruction following re-ranker that we trained, like, all this great stuff that really helps that use cases of those single-shot rags, or where you’re trying to extract some facts, some reasoning out of a PDF, for example.
26 00:03:04.050 ⇒ 00:03:12.480 Rajiv Shah: But one thing that’s changed over the last year is we’ve had all these great new AI models that can reason and take advantage of tools.
27 00:03:12.580 ⇒ 00:03:23.179 Rajiv Shah: So now what we want to do is incorporate those into the platform, but also be able to find ways to orchestrate all of those as well to solve problems.
28 00:03:23.410 ⇒ 00:03:26.050 Rajiv Shah: My slides, I don’t know if I…
29 00:03:26.460 ⇒ 00:03:31.219 Rajiv Shah: Somebody copy these and put them in the wrong order, like that. Okay, sorry.
30 00:03:31.900 ⇒ 00:03:33.559 Samuel Roberts: My build is not…
31 00:03:33.560 ⇒ 00:03:35.320 Rajiv Shah: coming out as magical as I want it.
32 00:03:35.510 ⇒ 00:03:41.439 Rajiv Shah: And so, kind of what we’ve added is this new agent orchestration layer.
33 00:03:41.700 ⇒ 00:03:53.699 Rajiv Shah: So that way, now, it’s not just kind of a RAG1 agent, but we can have much more dynamic agents, even multi-agents as well, that collaborate together to help solve problems.
34 00:03:54.180 ⇒ 00:04:09.740 Rajiv Shah: And then we’re pairing that with even kind of more reasoning applications, whether it’s reasoning on the extraction side. So we’ve recently added the ability to have an agentic table extraction, where when it’s doing table extraction, it spends a little bit more time thinking about it.
35 00:04:09.900 ⇒ 00:04:19.600 Rajiv Shah: looking at that table, making sure it’s right, it can give us a little bit of boost in that. Table extraction, some of the agentic data analysis I’ll show you here.
36 00:04:19.600 ⇒ 00:04:24.179 Uttam Kumaran: in a bit as well. But this is, in terms of capabilities.
37 00:04:24.180 ⇒ 00:04:27.980 Rajiv Shah: This is where we’re going with the platform. I realize this is really hard to read.
38 00:04:28.310 ⇒ 00:04:40.970 Rajiv Shah: slide, but just to give you a sense, like, right, we’re adding lots of tools, the text is SQL, right, the MCP, all that kind of stuff inside the platform as well. So just expanding the kind of the capabilities like that.
39 00:04:42.930 ⇒ 00:04:43.660 Rajiv Shah: Now…
40 00:04:44.000 ⇒ 00:04:51.229 Rajiv Shah: How’s it going to manifest itself? I think there’s some couple early pieces that I’ll be able to show you inside the platform like that.
41 00:04:52.590 ⇒ 00:05:00.319 Rajiv Shah: So, one is, we can build much more complex workflows, and a good example of this is Agentic search.
42 00:05:00.730 ⇒ 00:05:08.199 Rajiv Shah: So, instead of single-shot RAG, where we ask a question, right, go find the information, give you the results.
43 00:05:08.450 ⇒ 00:05:19.469 Rajiv Shah: Now we can put that LM inside the middle there to look and see, hey, were the results good enough? Maybe I should reframe my question, go do another round of tool calling.
44 00:05:19.620 ⇒ 00:05:28.119 Rajiv Shah: Now, that of course can take a little extra time, so we can get much better performance using a Gentic rag in terms of accuracy.
45 00:05:28.250 ⇒ 00:05:44.230 Rajiv Shah: But now you have a little bit of latency, so this is where you have to have the trade-offs for what’s best for the kind of problem that you’re solving. But we can get much more accurate, solve problems a lot better by adding the sigentic piece like that.
46 00:05:45.430 ⇒ 00:05:58.549 Rajiv Shah: And just to show you, kind of, what that looks like in the platform, so I have my traditional, kind of, RAG agent here, but if I do even, like, one of the traditional searches I had something like this.
47 00:05:58.990 ⇒ 00:06:13.999 Rajiv Shah: what you’ll see is we’ve added the ability now to follow all those steps in here. So, I’m doing different tool calls, like a semantic lexical search. You can see it’s looked at some of the results here, and decided, hey, I need to do another search.
48 00:06:14.080 ⇒ 00:06:20.340 Rajiv Shah: Around that, and then puts together the complete answer like that. So, much more, you know.
49 00:06:20.510 ⇒ 00:06:23.309 Rajiv Shah: Capable, powerful kind of search like that.
50 00:06:23.550 ⇒ 00:06:27.239 Rajiv Shah: And of course, comes with all the attributions, all the stuff that you’re used to as well.
51 00:06:29.480 ⇒ 00:06:30.160 Rajiv Shah: Okay.
52 00:06:32.700 ⇒ 00:06:37.350 Rajiv Shah: Next piece, because we’re on the speedrun here, is around extraction.
53 00:06:38.130 ⇒ 00:06:50.340 Rajiv Shah: A lot of what we use our platform is, you’ve ingested those PDFs, you know we do a really good job in terms of being able to identify what are figures, what are tables, being able to give descriptions of that.
54 00:06:50.720 ⇒ 00:07:00.660 Rajiv Shah: Some of you might remember kind of demos earlier, where I kind of show that we can take a bunch of unstructured PDFs, right? I can take a bunch of PDFs like this.
55 00:07:01.050 ⇒ 00:07:04.440 Rajiv Shah: I have some standard questions that I want to run across them.
56 00:07:04.970 ⇒ 00:07:09.100 Rajiv Shah: I can use our platform to then ask questions across
57 00:07:09.780 ⇒ 00:07:15.590 Rajiv Shah: Across this, fill out something like a spreadsheet with this, kind of in a very automated way to do this.
58 00:07:16.230 ⇒ 00:07:22.689 Rajiv Shah: So, this is a very familiar application, I’m sure you guys are doing it, where you have unstructured information, you need to put structure out of it.
59 00:07:23.570 ⇒ 00:07:25.460 Rajiv Shah: Everyone following me at this point?
60 00:07:25.830 ⇒ 00:07:26.420 Uttam Kumaran: Yes.
61 00:07:26.420 ⇒ 00:07:26.930 Rajiv Shah: Okay.
62 00:07:27.000 ⇒ 00:07:45.280 Rajiv Shah: So, this is a great example here of what to do, but it’s kind of hard to implement it the way I have here, right? We’re using our API, we’re putting it inside Excel, in this case, to do that. So, this is where we’ve generated a new endpoint that’s just focused on this task.
63 00:07:45.320 ⇒ 00:07:50.329 Rajiv Shah: It’s really geared for more the data scientist, the data engineer.
64 00:07:50.630 ⇒ 00:07:52.980 Rajiv Shah: Persona.
65 00:07:53.180 ⇒ 00:07:58.199 Rajiv Shah: And the idea here is, I have a PDF, a very long PDF,
66 00:07:59.230 ⇒ 00:08:07.110 Rajiv Shah: And we’re gonna be able to extract information out of it, We’re going to… Use a schema here?
67 00:08:07.860 ⇒ 00:08:23.300 Rajiv Shah: And the schemas we can use are going to be much more complex. This is a… this is a very much more of a power tool than what I showed you earlier. So, in this schema, for example, you’ll see I have 102 different fields that we want to request information out of.
68 00:08:24.250 ⇒ 00:08:30.620 Rajiv Shah: What is the filing date, the fiscal year, description, document, all of this information we want to pull out.
69 00:08:30.860 ⇒ 00:08:35.809 Rajiv Shah: And the idea is you’re gonna have this endpoint where you’re gonna be able to pass a PDF like this.
70 00:08:36.280 ⇒ 00:08:42.399 Rajiv Shah: what your schema is, and we’re gonna have tools to help people come up with schemas if you don’t know how to write up the schema like that.
71 00:08:42.960 ⇒ 00:08:49.460 Rajiv Shah: You’ll pass that through, and then you’ll get some extraction results that look like this, where you’ll be able to see
72 00:08:49.590 ⇒ 00:08:50.980 Rajiv Shah: All the answers?
73 00:08:51.180 ⇒ 00:09:10.939 Rajiv Shah: Each of these answers here come… will come with a confidence score, so that way you know, hey, was the extraction good or not? We’ll also put a little bit of, context for where we found the information as well, so if you’re visually going and checking, you can see what part of the document that you would want to use like this as well.
74 00:09:13.470 ⇒ 00:09:19.060 Rajiv Shah: So this should be online in a couple of weeks. You should be able to test this out, so if you have problems where you have
75 00:09:19.270 ⇒ 00:09:32.190 Rajiv Shah: It’s complex schemas, you want to take advantage of the latest kind of vision language models. This is also going to be reasonably priced, too. It’s not going to be crazy expensive in terms of the optimizations the folks have built to kind of run this as well.
76 00:09:35.650 ⇒ 00:09:36.220 Uttam Kumaran: Okay.
77 00:09:36.220 ⇒ 00:09:38.329 Rajiv Shah: Looking good? Does this make sense? Yeah. Okay.
78 00:09:38.330 ⇒ 00:09:39.330 Uttam Kumaran: Yeah, this is awesome.
79 00:09:39.330 ⇒ 00:09:39.980 Rajiv Shah: Yeah.
80 00:09:40.780 ⇒ 00:09:47.250 Rajiv Shah: And so the last one here I want to show you is… goes more into this multi-agent world.
81 00:09:49.020 ⇒ 00:09:58.439 Rajiv Shah: Now, I think you all know that a bunch of our early customers were folks like Qualcomm, right, out in telecommunications, semiconductors, those sorts of things, and…
82 00:09:58.670 ⇒ 00:10:10.250 Rajiv Shah: as we worked with them, the first set of use cases were under customer support, right? Like, let’s take their technical manuals, make them into PDF, make them… take… make those manuals available, people can ask questions on them.
83 00:10:10.600 ⇒ 00:10:23.310 Rajiv Shah: Well, the next set of use cases we have is they also have a lot of log data, and they’re always trying to figure out, like, what went wrong. Can I do a root cause analysis on this data to figure out what the problem was?
84 00:10:23.990 ⇒ 00:10:28.960 Rajiv Shah: This is a widespread problem through telecommunications, through anybody that has logs.
85 00:10:28.960 ⇒ 00:10:29.540 Samuel Roberts: Hmm.
86 00:10:29.920 ⇒ 00:10:38.809 Rajiv Shah: But it’s also very much… a much more generic problem of, like, I have massive amounts of data, I need to be able to find an answer to understand what’s going on.
87 00:10:39.880 ⇒ 00:10:59.260 Rajiv Shah: Now, the way we approach this is we use a multi-agent orchestration approach, where we’re gonna have multiple agents that work through this, where we have an agent that looks and figures out how to parse all that semi-structured log data, another one that puts that into an easy database, so then we can query it.
88 00:10:59.320 ⇒ 00:11:02.170 Rajiv Shah: Then we have a multi-agent… a reasoning
89 00:11:02.480 ⇒ 00:11:13.710 Rajiv Shah: a root cause analysis agent that uses Python and multiple calls to the database, and time series analysis to look through all that information to figure out what the faults are.
90 00:11:13.710 ⇒ 00:11:29.440 Rajiv Shah: And this reasoning here is assisted by any extra data that the users give us, as well as any domain knowledge rules that we know. So this isn’t, like, totally anomaly detection we want to take advantage of.
91 00:11:29.540 ⇒ 00:11:37.600 Rajiv Shah: past reasoning folks have used to be able to do that. But then, this will output a nice report that folks can use as well.
92 00:11:38.250 ⇒ 00:11:44.169 Rajiv Shah: So, I’m really excited about this, it’s kind of our first entry into kind of these multi-agent workflows.
93 00:11:44.310 ⇒ 00:11:56.289 Rajiv Shah: This stuff is, right now, it’s very driven very hard by our customer machine learning engineers, so it’s literally, like, services work where we’re using a lot of our domain knowledge engineers to build this.
94 00:11:56.510 ⇒ 00:12:05.689 Rajiv Shah: But I wanted to tease this to let you know we’re gonna give this out, or not give this out, but you’ll be able to use this, access this more as we move along.
95 00:12:06.120 ⇒ 00:12:16.889 Rajiv Shah: to get a sense of, like, how it works right now, I got a quick little demo I can do. So this is kind of a web app that we’ve built to illustrate how this log analysis works.
96 00:12:17.200 ⇒ 00:12:20.300 Rajiv Shah: I’m gonna go ahead and select my file.
97 00:12:20.750 ⇒ 00:12:25.309 Rajiv Shah: So let me go, I have a nice long… Log that we can grab.
98 00:12:26.700 ⇒ 00:12:33.780 Rajiv Shah: So I upload the log, I’m gonna use that dynamic agent, I can use some general Rules here.
99 00:12:34.220 ⇒ 00:12:35.350 Rajiv Shah: That we can have.
100 00:12:36.070 ⇒ 00:12:42.630 Rajiv Shah: I’m gonna apply those rules, and then I can ask it, like, hey, Tell me about the call.
101 00:12:44.170 ⇒ 00:12:45.210 Rajiv Shah: Failure.
102 00:12:45.730 ⇒ 00:12:47.650 Rajiv Shah: And it’s off and running?
103 00:12:50.070 ⇒ 00:13:05.319 Rajiv Shah: to do this. It’s gonna go through, perform all the tasks, it’s… this is a nice app for engineers, because there’s a lot of details in exactly what it’s going through to do. But it’s basically… it’s taken this really complex, super long file here.
104 00:13:05.720 ⇒ 00:13:25.010 Rajiv Shah: Right? Like, it’s not easy to parse or read all this stuff, and it’s going through and doing that root cause analysis, and kind of at the end of the day, what you’ll end up with is something like this, where we have the root cause analysis, why was the call so long? It describes what it is, what the errors were, we can even generate some artifacts.
105 00:13:25.100 ⇒ 00:13:38.120 Rajiv Shah: So, we’ve got a few customers now we’re kind of testing this out with, and we’re getting good results like that, but hopefully it gives you an idea of, like, the crazy complex use cases that we’re going to be able to solve, based on our platform coming up here fast.
106 00:13:40.100 ⇒ 00:13:42.499 Uttam Kumaran: No, I say the multi-agent use case is great.
107 00:13:43.850 ⇒ 00:13:49.770 Rajiv Shah: So, yeah, I think this is going to be a wave for a lot of these complex use cases we have inside of enterprises.
108 00:13:49.930 ⇒ 00:14:00.960 Rajiv Shah: Because you can break it, break it up, you can take advantage of parallelization, right? There’s only so much you can get out of one model, because you hit context limits and thinking, so it’s a natural pattern to follow.
109 00:14:04.100 ⇒ 00:14:07.890 Rajiv Shah: Anything else that I should show, or you’re curious about, kind of where we’re going? Like, I…
110 00:14:08.460 ⇒ 00:14:11.820 Rajiv Shah: kind of hit the… hit the biggest highlights that I could think of, like this.
111 00:14:13.890 ⇒ 00:14:15.490 Uttam Kumaran: Yeah, Wish, what’s your question?
112 00:14:17.480 ⇒ 00:14:20.129 Awaish Kumar: Yeah, I’m just asking that, like.
113 00:14:20.450 ⇒ 00:14:24.000 Awaish Kumar: I’m not sure if I missed anything before I came, but…
114 00:14:24.140 ⇒ 00:14:27.160 Awaish Kumar: Like, here in the demo, we saw that
115 00:14:27.310 ⇒ 00:14:46.780 Awaish Kumar: we uploaded a file and got some results, so is it possible to connect it with any external tools? Like… like, for example, I’m running GitHub Actions, and it fails sometimes, and it has logs there. Does it connect with tools directly to read their logs, and…
116 00:14:46.950 ⇒ 00:14:57.590 Awaish Kumar: send us the summary of why… what is the root cause of that failure, instead of engineer going into getting the logs and uploading to contextual.
117 00:14:57.810 ⇒ 00:15:15.860 Rajiv Shah: Yeah, so I think there’s two elements there. One is just, like, what are the input-output interfaces that we’re gonna have for this? And here, we just have that simple, kind of, app that somebody probably vibe-coded up, where the user’s just putting it in like that, because we use this for demos. In a real situation, it could be
118 00:15:15.860 ⇒ 00:15:23.790 Rajiv Shah: integrated easily with GitHub Actions, it could be integrated with another enterprise application to automatically pull that logs over.
119 00:15:23.790 ⇒ 00:15:36.279 Rajiv Shah: So I think that’s one element of your question, and this is where there’s a lot of flexibility in coding that piece. I think the other is, what kinds of use cases are we going to solve this for? And so this is where we’re tightly controlling this at this point, because
120 00:15:36.820 ⇒ 00:15:48.589 Rajiv Shah: We know that to get this to work well, we need that combination of understanding how best to build that reasoning agent, as well as taking advantage of the domain knowledge.
121 00:15:48.970 ⇒ 00:16:02.329 Rajiv Shah: If you just put this out there, you’ll get what… if you use something like a Crew AI or some generic thing to do troubleshooting, and we really want this to be special and highly accurate, and bring in those details like that to really solve problems.
122 00:16:05.570 ⇒ 00:16:07.120 Rajiv Shah: Does that help with that piece?
123 00:16:11.700 ⇒ 00:16:12.689 Uttam Kumaran: That makes sense to me.
124 00:16:12.690 ⇒ 00:16:13.849 Rajiv Shah: Yeah, and…
125 00:16:14.290 ⇒ 00:16:20.239 Rajiv Shah: Look, if you guys have anything that you think is close to this, let us know. I mean, we’re open, we’re happy to kind of work with you.
126 00:16:20.620 ⇒ 00:16:33.579 Rajiv Shah: I just like the level set for this log analysis, because everybody has some type of logs, right? And everybody, right, wants to find what exactly went wrong, and like, there’s only so much we’re going to be able to do, but it’s an exciting kind of path we’re going down.
127 00:16:38.390 ⇒ 00:16:38.990 Rajiv Shah: Any questions?
128 00:16:38.990 ⇒ 00:16:39.850 Uttam Kumaran: Questions, guys?
129 00:16:40.970 ⇒ 00:16:44.050 Rajiv Shah: I’m assuming, like, that extraction piece, that this would be…
130 00:16:44.330 ⇒ 00:16:48.730 Rajiv Shah: I’m sure you guys do that now, and you could probably be up for testing that one pretty soon.
131 00:16:48.730 ⇒ 00:16:49.980 Samuel Roberts: Yeah, I could see that.
132 00:16:51.220 ⇒ 00:16:57.050 Uttam Kumaran: Yeah, I guess, I mean, there’s only a couple of AI… Casey and Mustafa, you guys have any questions?
133 00:17:06.800 ⇒ 00:17:12.460 Samuel Roberts: I have a… I have a quick question. I’m curious about that, like, generating these schemas for…
134 00:17:12.760 ⇒ 00:17:21.849 Samuel Roberts: actually, like, getting the structured output. It seems like a decent amount of work, potentially, to… or, like, prior knowledge, potentially, some of this, so I’m curious…
135 00:17:22.150 ⇒ 00:17:25.450 Samuel Roberts: You mentioned that there might be some, like, things to help put that together.
136 00:17:25.630 ⇒ 00:17:26.380 Samuel Roberts: Wondering what that…
137 00:17:26.380 ⇒ 00:17:34.670 Rajiv Shah: I know in the first iteration of this, we did have helper tools, like, you could just essentially say, hey, these are the elements I want out of it.
138 00:17:34.670 ⇒ 00:17:53.110 Rajiv Shah: and it could build that JSON schema. What we want to do is give kind of the advanced users the flexibility that if they, for example, know it’s a certain type, for example, to be able to bring that information with those schemas like that, so… Okay. But yeah, I mean, it’s a pretty straightforward schema, and we can kind of easily help with.
139 00:17:53.280 ⇒ 00:17:54.600 Samuel Roberts: Yeah, yeah, okay.
140 00:17:56.310 ⇒ 00:17:57.660 Samuel Roberts: Great. Thank you.
141 00:17:57.660 ⇒ 00:17:58.210 Rajiv Shah: Yeah.
142 00:18:00.680 ⇒ 00:18:17.749 Rajiv Shah: So, in terms of availability, the multi-agent let us know. We should have this extraction available for testing in the next couple of weeks, and then we’ll also be rolling in this Agentic search that I showed you in the next couple of weeks as well. Internally, we call it ACL, but,
143 00:18:18.180 ⇒ 00:18:22.840 Rajiv Shah: I think that is also going to come with a node-based UI as well.
144 00:18:22.840 ⇒ 00:18:23.390 Samuel Roberts: It was in the.
145 00:18:23.390 ⇒ 00:18:40.060 Rajiv Shah: but I didn’t mention it, like an N8N type of approach as well. So… but, you know, that stuff, we’re building it, right? Like, N8N’s had a long, long advantage to it, but you’ll be able to do that stuff inside of our platform as well now, and keep those workflows inside the platform.
146 00:18:45.000 ⇒ 00:18:45.810 Rajiv Shah: Alright.
147 00:18:47.860 ⇒ 00:18:56.340 Rajiv Shah: Well, thanks! Reach out if you guys have any questions, we have the Slack channel set up, happy to kind of do this, and keep you up to date on all this.
148 00:18:56.700 ⇒ 00:18:57.410 Uttam Kumaran: Perfect.
149 00:18:57.640 ⇒ 00:18:58.929 Uttam Kumaran: Thank you for taking the time.
150 00:18:58.930 ⇒ 00:18:59.600 Rajiv Shah: Yeah, no, thanks.
151 00:18:59.600 ⇒ 00:19:00.060 Awaish Kumar: Appreciate it.
152 00:19:00.060 ⇒ 00:19:05.210 Rajiv Shah: I’m excited, yeah. No, if you guys have anything that’s close, let us know. Definitely. So we’ll find a way to work with you.
153 00:19:05.210 ⇒ 00:19:17.899 Uttam Kumaran: Yeah, we have… we’re gonna start using some of the Agentix search for a couple things, because we’re already planning on doing that. Yeah. So, yeah, we’ll definitely let you know once we… once we have some… when that gets out, and we can kind of flip up a demo, like, we’d love to show you guys.
154 00:19:18.420 ⇒ 00:19:20.330 Rajiv Shah: Yep, that’d be great.
155 00:19:21.630 ⇒ 00:19:24.220 Uttam Kumaran: Okay, perfect. Alright, thank you.
156 00:19:24.690 ⇒ 00:19:25.500 Samuel Roberts: Yeah, thanks for the pass.
157 00:19:25.990 ⇒ 00:19:26.550 Holly Condos: See ya.
158 00:19:26.550 ⇒ 00:19:27.210 Rajiv Shah: See ya.