Fireside Chat with Olivia Nottebohm & Erik Charles — Becoming AI-First: Redesigning the Org for Agentic Advantage

1240
03:32:18.740 –> 03:32:33.150
Julia Nimchinski: Olivia Nodembaum, COO of Box, and Eric Charles, fractional CMO and CRO at Hana Operators, and former CMO, exactly. Welcome! Our community’s super excited, can’t wait to get into this fireside Chat! How are you doing?

1241
03:32:34.670 –> 03:32:39.350
Erik Charles: I’m do- I am doing well. I’m glad to say we all made it in here on time.

1242
03:32:40.810 –> 03:32:47.410
Julia Nimchinski: Yeah, Olivia, what’s new with Box? We’re seeing it making a lot of waves in all of the communities.

1243
03:32:48.610 –> 03:32:54.160
Olivia Nottebohm: First, it’s just lovely to be here, so thank you for inviting me on, Julia, and look forward to chatting with you, Eric.

1244
03:32:54.610 –> 03:32:55.620
Olivia Nottebohm: Hi, everyone.

1245
03:32:59.320 –> 03:33:00.210
Erik Charles: Alright.

1246
03:33:00.360 –> 03:33:05.440
Erik Charles: Are you doing it? Do you want to do intros, or shall I take it on, Julia?

1247
03:33:05.440 –> 03:33:06.330
Julia Nimchinski: Take it on, Eric.

1248
03:33:06.330 –> 03:33:14.800
Erik Charles: All right, we’ll make this quick. My name is Eric Charles, former CMO of Exactly, currently working in fractional work and board advisory work.

1249
03:33:14.850 –> 03:33:19.140
Erik Charles: But actually, for… on a living space, I gotta say, I’ve been a…

1250
03:33:19.170 –> 03:33:35.799
Erik Charles: customer of Box since probably almost the very beginning, switching over to it from certain other systems and platforms. And my favorite, I think, Box memory actually is when my trusted Mac

1251
03:33:36.210 –> 03:33:53.570
Erik Charles: truly had a hard drive completely die, which Apple was not used to, and the person was so concerned, he’s like, what did you lose? I said, nothing. Everything I do is either admittedly an online email system, or all files are backed up on Box, so I just need you to put in a new drive, and I’m good to go.

1252
03:33:53.610 –> 03:34:06.380
Erik Charles: Now, that is the most basic use of the technology the company’s been providing. I’m realizing this, which is not what we’re talking to. So, for… to quickly introduce, I get the privilege of being able to talk to Olivia for the next…

1253
03:34:06.380 –> 03:34:22.409
Erik Charles: I don’t know, 25 minutes or so. She’s the COO at Box. She handles the global go-to-market organization, also, which includes partnerships, marketing, sales, and customer success. That should be a whole different conversation we should have at some point, of how do you combine all those.

1254
03:34:22.410 –> 03:34:26.510
Erik Charles: And, and manage that aspect, in that sense.

1255
03:34:26.510 –> 03:34:39.810
Erik Charles: She’s a member of the board of directors over at Appfolio and Light Matter, and a BA in Econ from Harvard, and an MBA from Stanford. So, on that side, I’d have to say go Cardinal, just because Stanford is my undergraduate alma mater.

1256
03:34:39.810 –> 03:34:54.789
Erik Charles: But to jump right into this, our conversation is all about becoming AI first. I think our fancy title, we came up with this was Redesigning the Organization for Agentic Advantage, which sounds really confusing to a lot of people.

1257
03:34:54.790 –> 03:35:03.900
Erik Charles: So, I thought I’d let you… and I had the opportunity to read a great white paper that you participated in, and I enjoyed that, but…

1258
03:35:04.030 –> 03:35:08.060
Erik Charles: if… what did I miss in the intro, or how do you want to jump into this?

1259
03:35:09.590 –> 03:35:29.000
Olivia Nottebohm: Well, first of all, thank you. Thanks for the intro. Nice to… to be with you today. I’ll just jump in. I mean, if the conversation’s on AI first, thanks for reading the white paper. Erin and I co-wrote that, and you know, we also used AI, so that was… it’s good to, like, do it on yourselves when you’re creating something.

1260
03:35:29.000 –> 03:35:53.999
Olivia Nottebohm: So yeah, so we think about AI first having five key principles. The one is that you have AI as a capability expander, which is to say the humans are doing the work, but really, at least in this first evolution of AI first, it’s all about AI enabling things… you to do things you’d never thought possible. The second is that there’s this human-to-AI collaboration.

1261
03:35:54.000 –> 03:35:57.770
Olivia Nottebohm: So you always want a human to be part of the decision-making.

1262
03:35:57.770 –> 03:36:15.440
Olivia Nottebohm: The third is AI-native design, right? So, as you go to think about processes that you’re now going to execute in an AI-first world, how do you actually start with a blank canvas, right, and think afresh, given what you know is possible with AI? And obviously, that’s constantly changing day by day, but at least having.

1263
03:36:15.440 –> 03:36:15.890
Erik Charles: I’m like.

1264
03:36:15.890 –> 03:36:31.460
Olivia Nottebohm: And then the fourth is data. You know, data is the strategic asset of all of this. If you… it’s like garbage in, garbage out. That’s like an old saying in tech land, but still applies in the land of agents. And then, of course, the last is trust and governance. You know, in all of this.

1265
03:36:31.630 –> 03:36:40.240
Olivia Nottebohm: flurry of using AI and agents and all that we think we can do if we start having data leakage and all of that, you know, we have much bigger problems on our hands.

1266
03:36:40.940 –> 03:36:53.390
Erik Charles: I was at an interesting talk in, actually, Salt Lake City a couple weeks ago, and a professor from BYU who talked about… and it was an AI panel, and he just jumped in and said, oh yeah, and by the way, quantum computing.

1267
03:36:53.390 –> 03:37:06.969
Erik Charles: is getting really close, and that’s gonna completely shift the AI landscape in and of itself, once it happens. Now, his job is to be a futurist, and there’s certain things we’re always talking about what’s coming next. I want to ask you.

1268
03:37:07.070 –> 03:37:10.330
Erik Charles: What’s your… current.

1269
03:37:10.540 –> 03:37:13.570
Erik Charles: favorite use of AI today in the workplace?

1270
03:37:15.120 –> 03:37:36.540
Olivia Nottebohm: Honestly, we have so many, we do box on box here at Box, not surprisingly, so we, basically have a thousand flowers blooming, so anyone can ask to create an agent, and then it goes for some, you know, bit of ratification, make sure that it’s all wired up correctly, and then we publish it out, and then it’s available to anyone who wants to use it.

1271
03:37:36.540 –> 03:37:44.310
Olivia Nottebohm: But an example would be, okay, someone’s going in to prepare for a customer meeting. You saw some of that in the previous one, but now this is

1272
03:37:44.310 –> 03:37:48.420
Olivia Nottebohm: all of the content that you have on Box, so it’s secure and trusted, we have integration.

1273
03:37:48.420 –> 03:37:48.800
Erik Charles: It’s like.

1274
03:37:48.800 –> 03:38:00.660
Olivia Nottebohm: Salesforce, all of that. So we call that our discovery agent, but then we also have a messaging agent, and of course, you’re uploading all of the information about Box, our value proposition, all of our product specs, all of our positioning.

1275
03:38:00.660 –> 03:38:17.509
Olivia Nottebohm: And then, you know, auto-magically, you have outreach written for you. Now, we’re still trying to keep humans in loop on that to quality check it, but that’s another great one we’re using. And then another one that’s been a real hit with our professional services team is our RFP responder agent.

1276
03:38:17.510 –> 03:38:19.589
Erik Charles: Oh, yes. Yeah, yeah, yeah.

1277
03:38:19.830 –> 03:38:27.009
Olivia Nottebohm: So, I was once a consultant, you know, boy, would I have loved to have an agent write all my proposals.

1278
03:38:27.010 –> 03:38:37.440
Erik Charles: So, are you using that if you’re… and you can just say next question, are you… because I always say the worst RFPs I ever dealt with, and even some companies I advise.

1279
03:38:37.570 –> 03:38:39.919
Erik Charles: were not actual RFPs.

1280
03:38:40.220 –> 03:38:41.490
Erik Charles: it was…

1281
03:38:41.580 –> 03:38:52.429
Erik Charles: analyst RFPs, as in, say, one of the big analyst firms is reviewing your space, and they send you… I actually, I found a folder in a backup drive

1282
03:38:52.430 –> 03:39:01.859
Erik Charles: That said… and it was just… I’ll admit, my folder was RFPs from hell, questions and answers, and I’d save them that way. Are you… are you trusting that for analysts?

1283
03:39:03.190 –> 03:39:19.200
Olivia Nottebohm: Honestly, we… yes, we just keep humans in the loop, so before anything gets sent out, someone looks it over and makes sure that all the content is okay, but yeah, I mean, at this point, I mean, quantum computing aside, those agents are performing as good as a human.

1284
03:39:20.000 –> 03:39:39.900
Erik Charles: You mentioned in your process, when someone creates an agent, and you quickly hit the validation check phase before we let this out in the public. What does that phase look like? I mean, again, I’m a company, I want to be AI first, hold it, yeah. I get, I’ve got some smart people who can create agents.

1285
03:39:40.660 –> 03:39:42.119
Erik Charles: Now, what’s that check phase?

1286
03:39:42.120 –> 03:39:57.279
Olivia Nottebohm: Yeah, so here’s what we’re solving for from a business perspective. One is, we don’t actually want, a thousand, like, thousands and thousands of agents, because we think that’ll be, frankly, just confusing and hard for people to execute on.

1287
03:39:57.280 –> 03:40:03.450
Olivia Nottebohm: And so, we love the innovation, so everyone’s coming in with their ideas.

1288
03:40:03.450 –> 03:40:28.170
Olivia Nottebohm: But, we do want to have some sort of, like, publishing mechanism, which is what we do. Now, we don’t overly constrain it. We have over 50 agents at work, in our go-to-market teams today. But there is a step where it’s like, who should… what is the process? Are they pulling data that’s appropriate, frankly? Would they be more powerful if they could have access to additional data?

1289
03:40:28.170 –> 03:40:39.689
Olivia Nottebohm: Right? And then who are we provisioning it to? So, you know, all our SDRs get access to this agent, or our people team gets access to this agent, but no one else should. You know, that kind of stuff.

1290
03:40:39.690 –> 03:40:49.730
Erik Charles: Sure. And then, not that you glossed over it, but one of the key things, because you talked about as one of the key pillars, is cleanliness of data.

1291
03:40:49.760 –> 03:40:58.910
Erik Charles: Yes. And I’ve been playing with that game for a while. I, you know, spent a lot of time in the sales compensation space, and even, like.

1292
03:40:58.910 –> 03:41:11.079
Erik Charles: you know, pulling… you know, in that case, I first started with survey data for my clients in the old, old days, and then I remember using CompuStat as a source for company performance data.

1293
03:41:11.160 –> 03:41:12.310
Olivia Nottebohm: Yep. You know…

1294
03:41:12.360 –> 03:41:27.379
Erik Charles: now we can get so much stuff slung in here, and I’ll freely admit, I will admit, you know, my name is Eric Charles, I use AI regularly, and I actually put it in for any of the work I do for companies I’m advising. I will… I’ve actually started putting in a notation, here is

1295
03:41:27.430 –> 03:41:44.510
Erik Charles: how I used AI in generating this report. Sometimes it’s just like, I let the AI summarize a bunch of stuff. Sometimes I summarized things I’ve written that I’m like, I really don’t want to re-read my writings, because I find things that have been published in major things that have typos that somehow snuck through, but…

1296
03:41:44.710 –> 03:41:47.140
Erik Charles: You know, how are you…

1297
03:41:47.870 –> 03:41:54.440
Erik Charles: checking the data. What is the data… the data cleansing or validation that you follow?

1298
03:41:55.310 –> 03:42:07.099
Olivia Nottebohm: this is the crux, honestly, of the whole thing, right? Because, and I think that this is a bit of the misfires that we saw early on, right? Which was… it wasn’t, so much that

1299
03:42:07.270 –> 03:42:22.149
Olivia Nottebohm: Yes, it was a combined two failure modes. One, agents had access to information they probably shouldn’t have had access to, and that kind of caused alarms, but then also they realized that the agent didn’t delineate between version 23 and version 25,

1300
03:42:22.150 –> 03:42:31.139
Olivia Nottebohm: Right? Because they didn’t know whether it was the later timestamp, or the previous timestamp, or which was the data that should be kind of the ground truth that they should be using, right?

1301
03:42:31.140 –> 03:42:38.399
Olivia Nottebohm: So, we obviously live and breathe unstructured data, that’s our focus. What I mean by that is, you know.

1302
03:42:38.400 –> 03:42:53.330
Olivia Nottebohm: your Google document, an Excel spreadsheet counts as well, a PDF, a photograph, a video, all of that. We are constantly creating ways in which people can curate the data. So, on Box, we have something called a hub.

1303
03:42:53.330 –> 03:43:05.260
Olivia Nottebohm: You can put as many files as you want in that hub. And, if you have a working team or a functional team, we have a go-to-market hub, and we have a sales hub, and we have a pricing hub, and, you know, you can imagine how this goes.

1304
03:43:05.260 –> 03:43:20.010
Olivia Nottebohm: But the beauty is we’re not replicating all of those documents, it’s just, you know, anchoring back into the original document, so you’re not causing, spaghetti in your overall, environment. But the important thing is there, is the ability to curate.

1305
03:43:20.010 –> 03:43:36.000
Olivia Nottebohm: Right, so you’re putting in the product spec that you want to reference, or you’re putting in the customer information that you want that’s relevant to what you’re working on. And I do think that in the world of AI, there needs to be some sort of curation step.

1306
03:43:37.670 –> 03:43:42.259
Erik Charles: Yeah, I think curation, and I’m gonna emphasize the… at least for

1307
03:43:42.510 –> 03:43:48.039
Erik Charles: for fa… in certain… I’ll call it customer-facing, but outside of the company facing, almost anything facing.

1308
03:43:48.390 –> 03:43:58.740
Erik Charles: I do think there’s going to be a place for that caveat of, here’s where I used it, and here’s where I didn’t, just for the people that are still having some questions.

1309
03:43:58.740 –> 03:44:14.069
Olivia Nottebohm: But also, you know, less people think this is a heavy lift. Like, this is if you’re… you want your agent to write a marketing blog for you. Okay, well, you better load in the right content for them to be working on, even if it’s a thousand documents, right? That’s okay. Put that in a hub.

1310
03:44:14.070 –> 03:44:21.999
Olivia Nottebohm: Different than we have customers who do insurance claims, and they process 10,000 insurance claims per week.

1311
03:44:22.000 –> 03:44:29.589
Olivia Nottebohm: And that’s just the inbound flow. There’s not a curation, in the sense that, like, whatever they receive, we’re doing metadata extraction on.

1312
03:44:29.930 –> 03:44:34.240
Olivia Nottebohm: Workflow based on what’s in that metadata, and then it progresses through.

1313
03:44:34.610 –> 03:44:39.799
Erik Charles: Right. How do you… it’s funny, because you mentioned pricing,

1314
03:44:40.120 –> 03:44:54.629
Erik Charles: And obviously, you’ve got your internal price books, you know, what… and, you know, from my experience, it’s like, here’s the price you’re allowed to give, here’s the discount your manager can approve, your director can approve, the vice president can approve.

1315
03:44:54.750 –> 03:45:04.979
Erik Charles: Are you using the agent to… to help? Like, if I’m a rep and I’m involved in a pricing negotiation, enterprise company, it’s like, alright, here’s the price book that…

1316
03:45:05.320 –> 03:45:11.510
Erik Charles: Please note, 90% of the time, this is where we end up at, or… and so you better get in front of.

1317
03:45:11.510 –> 03:45:13.159
Olivia Nottebohm: We don’t expose that to our.

1318
03:45:13.160 –> 03:45:16.300
Erik Charles: Oh, oh, not… okay.

1319
03:45:16.540 –> 03:45:23.750
Olivia Nottebohm: That’s a kind of choose your own adventure. They’ll find the floor eventually if they work hard enough, but we don’t tell them what it is.

1320
03:45:23.750 –> 03:45:48.159
Erik Charles: Yeah, well, because I was asked, you know, one of my questions that I see bandied about is at what point, you know, will sales be replaced by AI? And my response is usually when procurement is purely AI. If the entire buying, you know, if I’m no longer negotiating with executives and just negotiating with the procurement AI, I should probably have my own bot to represent me.

1321
03:45:48.160 –> 03:45:52.680
Erik Charles: And they can both go at light speed and come to that optimal

1322
03:45:52.680 –> 03:45:56.180
Erik Charles: That optimal decision point from grad school statistics.

1323
03:45:56.680 –> 03:46:08.469
Olivia Nottebohm: Yeah, I believe, and we just went through BoxWorks, we just had thousands of customers in town, and hundreds of partners, and, you know, we had our dinners, and our… our cruises in the Bay, and all of that. I…

1324
03:46:08.470 –> 03:46:21.939
Olivia Nottebohm: it still comes down to relationships and who people want to work through. I think that is the human part of sales that is irreplaceable, and you can be better prepared for meetings, you know, someone else… some agent can do your slide deck.

1325
03:46:21.940 –> 03:46:22.260
Erik Charles: Right.

1326
03:46:22.260 –> 03:46:35.460
Olivia Nottebohm: But you have to still be able to articulate the value proposition, and why it’s good to work with you as a company, and why you’re always going to be there to answer the phone, all of that. I guess if we fast forward and agents are talking to agents, all of that gets…

1327
03:46:35.460 –> 03:46:44.150
Olivia Nottebohm: you know, obscured, but for the here and now, I think there’s… humans matter a lot, and that relationship is pivotal.

1328
03:46:44.540 –> 03:46:46.999
Erik Charles: Do your partners have access to some of your agents?

1329
03:46:47.490 –> 03:47:00.909
Olivia Nottebohm: For sure, I mean, most of our partners are on Box, right? So that’s important, because they need to be kind of drinking the champagne that they’re helping sell. So absolutely, yeah, they have access to Box, they leverage Box, they spin up agents, all of that.

1330
03:47:01.340 –> 03:47:13.979
Erik Charles: Okay. How about the cost factor? As more and more agents in different systems… I’ll use Salesforce as an example, since they have publicly stated that, you know, to use their agents, there’s gonna be a…

1331
03:47:14.080 –> 03:47:18.230
Erik Charles: I feel like we’re back to the cell phone minutes age, you know?

1332
03:47:18.230 –> 03:47:20.780
Olivia Nottebohm: Yeah, they just published Flex Credits, yep.

1333
03:47:20.780 –> 03:47:21.550
Erik Charles: Yeah.

1334
03:47:21.550 –> 03:47:22.100
Olivia Nottebohm: Yeah.

1335
03:47:22.100 –> 03:47:36.979
Erik Charles: So how does an AI-first company control their costs? I have this vision, like you’ve said, you’ve got all these agents people are spinning up. That’s gotta… somewhere, someone in finance, the Office of the controller, is going, oh my gosh, we spent how much last week?

1336
03:47:38.470 –> 03:47:46.180
Olivia Nottebohm: Yeah, I mean, this is where, I mean, not to date myself, but I do feel like we’re on the maturity curve of cloud.

1337
03:47:46.180 –> 03:47:51.199
Olivia Nottebohm: Right? Where, when Amazon made that market, they realized, and I think the same is true now.

1338
03:47:51.200 –> 03:48:15.269
Olivia Nottebohm: that actually unpredictability is more scary, at least for the enterprise, right? And so that’s why they came up with the concept of reserved instances, all of that. Now, I worked at Google Cloud, we entered the market 7 years later, we had 0% market share, but 7 years into that journey, people were pissed about paying for RIs that they were not using, right? Gone were the days of, I want predictability, and now they’re like, I want

1339
03:48:15.270 –> 03:48:17.030
Olivia Nottebohm: I want to pay only for what I use.

1340
03:48:17.030 –> 03:48:39.110
Olivia Nottebohm: And that was our wedge into the market. We came out with down to per second billing of compute, which was a little overkill, I have to be honest. But, you know, it was like, okay, you’ve dialed it in, you have pattern recognition, you understand, you’ve got your budgeting figured out, like, great, you’re only going to pay for what you use for, and we might throttle you if you ask us to throttle you, but otherwise, we’re just gonna let it run and we’ll send you the bill.

1341
03:48:39.110 –> 03:48:52.490
Olivia Nottebohm: Right? We’re just not there yet with AI. I think people are terrified, frankly, of that potentially happening in their environment, and with their budgeting, and all of that. So, I do predict that people will buy in kind of, like, packaged amounts.

1342
03:48:52.490 –> 03:48:56.660
Erik Charles: Right? Because they… and I see it in the contracts coming in, which is.

1343
03:48:56.660 –> 03:49:12.799
Olivia Nottebohm: you know, I only… this is how many API calls, we call them AI units, this is how many AI units we want to consume, it’s in the millions and millions and millions, obviously, right? Per month. But then cap me at this, and only charge me for this, and then throttle me there.

1344
03:49:12.800 –> 03:49:15.610
Erik Charles: You know, so people do need to still navigate that.

1345
03:49:17.340 –> 03:49:23.199
Erik Charles: So, with all this, we’re… you know, going back to an AI-first company.

1346
03:49:23.510 –> 03:49:29.129
Erik Charles: Where would you say Box is right now, in terms of on the journey versus optimal?

1347
03:49:29.640 –> 03:49:35.040
Erik Charles: And what is optimal? And again, optimal based on what we can see, you know…

1348
03:49:35.040 –> 03:49:36.790
Olivia Nottebohm: Yeah, yeah.

1349
03:49:36.790 –> 03:49:37.340
Erik Charles: Part of the pot.

1350
03:49:37.340 –> 03:49:48.470
Olivia Nottebohm: I mean, I feel like you’re… there’s always more to be done, right? Especially in this world, because AI is changing so quickly, the change management can’t possibly keep up, right? So…

1351
03:49:48.470 –> 03:49:57.389
Olivia Nottebohm: I would say we’re actually quite far along. We define the use of AI in three stages. The first stage is, like, you’re just answering a question.

1352
03:49:57.390 –> 03:50:02.259
Olivia Nottebohm: Right? You’re asking them to retrieve some sort of answer to a question that you have.

1353
03:50:02.280 –> 03:50:12.939
Olivia Nottebohm: The second is you’re asking them to do a task, right? It’s finite, it’s obviously they are provided an objective, right? But it feels more almost like a predictable task.

1354
03:50:13.310 –> 03:50:21.140
Olivia Nottebohm: And then the third is, like, it’s an entire workflow, right? And in the RFP example I gave you, it is actually an entire workflow. Like, it’s this…

1355
03:50:21.180 –> 03:50:40.980
Olivia Nottebohm: beautiful thing where the RFP comes in, the metadata is extracted out, we capture all the key fields, what are the scopes, what are the deliverables, what are the timelines, right? And then, actually, we use our DocGen functionality to create the response to the RFP, then you have human in the loop, they check it over, and then you send it out for signature on BoxSign.

1356
03:50:41.760 –> 03:50:53.930
Olivia Nottebohm: awesome, right? Of course, all of our processes do not run that way. So, in the sense of, like, we’ve tasted stage 3, we’ve executed Stage 3, but it’s not universally true, and it’s not every single process.

1357
03:50:54.410 –> 03:51:00.260
Erik Charles: Is there a process that you’re willing to admit to that you backed away from? You thought, hold it, I don’t think…

1358
03:51:00.830 –> 03:51:05.250
Erik Charles: I don’t think we’re ready to… to go AI first on this one yet.

1359
03:51:05.760 –> 03:51:19.289
Olivia Nottebohm: Well, so what we’ve done, is actually map all the go-to-market processes, and then decided, both from a change management and from a building of agents, okay, we’re not gonna tackle all of these at once.

1360
03:51:19.290 –> 03:51:26.680
Olivia Nottebohm: Right? So we went with, kind of, we literally have a 2×2, which is embarrassingly dark to admit, but it’s…

1361
03:51:26.680 –> 03:51:35.580
Erik Charles: Hey, it’s required, it’s required. I should have… if I ever got a tattoo, it started getting ink, I think a 2×2 box would have to be on there somewhere.

1362
03:51:35.580 –> 03:51:48.049
Olivia Nottebohm: Right, right. So it’s, you know, degree of impact and amount of effort required, right? And so, things like our discovery agent for sales, okay, well, we have hundreds and hundreds and hundreds of sellers out there, boy, that’s gonna have

1363
03:51:48.050 –> 03:51:55.620
Olivia Nottebohm: fast impact quickly, right? And so we almost took, like, a Maslow’s hierarchy to it, also because we were testing into it.

1364
03:51:55.660 –> 03:52:14.320
Olivia Nottebohm: Right. So, and then more some of the sensitive stuff that we haven’t quite figured out, both where we stand on the matter, you know, like, someone writing your… someone else’s performance review, like, you know, we’ve drawn a hard line there, like, an AI agent shouldn’t be writing someone else’s performance review. But… but, you know, so I think that’s, like, kind of how we think about it, at least.

1365
03:52:17.180 –> 03:52:27.000
Erik Charles: Who should be scared for their job as we go AI-first? And an AI… I mean, let’s say I’m applying for a job at an AI-first company, or I’m at a company that says, we’re going AI-first.

1366
03:52:27.040 –> 03:52:46.930
Erik Charles: Who should be paranoid, and how can they, protect themselves? And again, I’m thinking about… I’ll use Salesforce again, you know, Mark talking about dropping 4,000 CS people, thanks to AI, or 4,000 CS roles, let’s say CS roles, I don’t want to misquote their miscite. Yeah.

1367
03:52:47.170 –> 03:52:47.900
Erik Charles: on edge.

1368
03:52:47.900 –> 03:53:03.070
Olivia Nottebohm: I don’t think so. I mean, my genuine belief is that there’s always more code to write, there are always more customers to see, there’s always a better, deeper understanding of how we could be serving our customers, reaching out to prospects.

1369
03:53:03.130 –> 03:53:12.159
Olivia Nottebohm: So it… it might prov… it might slow down some future hiring, I don’t see any scenario

1370
03:53:12.200 –> 03:53:27.589
Olivia Nottebohm: where people are being exited, to be totally honest. I think there’s, like, a very real thing, and, you know, everyone’s heard it a thousand times now, that, you know, people should just learn how to use AI, right? Because the people who will have a tougher time going forward will be the people

1371
03:53:27.610 –> 03:53:39.940
Olivia Nottebohm: who don’t know how to use AI or somehow threatened by it, right? And then they don’t embrace it and try new things and all of that. So I think it’s more that than anything else, personally. But I’m an optimist, so…

1372
03:53:39.940 –> 03:53:53.419
Erik Charles: Oh, it’s funny, I play around with things. I had a company I was advising, they asked me to take a look at something. I said, well, I just asked 3 different AIs the exact same question to see if they even came up with your company as an answer, and here’s the results, so that you

1373
03:53:53.420 –> 03:54:04.549
Erik Charles: I mean, and it was a small go-to-market thing. We were talking about, you know, you know, AI search optimization as opposed to search engine optimization. I said, well.

1374
03:54:04.850 –> 03:54:11.829
Erik Charles: When I ask people, when I ask the agent, you know, the systems, not agents, the systems, this is their answers.

1375
03:54:11.830 –> 03:54:14.800
Olivia Nottebohm: We’re in this. Look, we are doing LLM marketing.

1376
03:54:14.800 –> 03:54:15.640
Erik Charles: Scary.

1377
03:54:15.710 –> 03:54:17.330
Olivia Nottebohm: Very real. Yeah, people are.

1378
03:54:17.330 –> 03:54:21.330
Erik Charles: Any huge successes there? Sorry, just to go sideways.

1379
03:54:21.330 –> 03:54:22.490
Olivia Nottebohm: I love that.

1380
03:54:22.490 –> 03:54:44.510
Olivia Nottebohm: Yeah, we’re early in that, but we recognize that gone are… it’s just the same arc, right? Like, CIOs used to read magazines, they don’t do that. They went to look on the websites, and so your website had to be perfect, and the Google search had to be right, and now it’s like, well, actually what matters is what ChatGPT tells them when they ask them, right?

1381
03:54:44.510 –> 03:54:52.890
Olivia Nottebohm: And so, actually, like, creating the content so these LLMs pick it up, is much more important than it ever has been before.

1382
03:54:53.740 –> 03:54:55.479
Erik Charles: What would you tell…

1383
03:54:55.930 –> 03:55:08.299
Erik Charles: a growing company, say, SMB moving into the mid-market, so, you know, they’re at that beautiful, let’s say, you know, in the tech space, say, past the 25,

1384
03:55:08.330 –> 03:55:18.249
Erik Charles: million in ARR stage, what should they be doing to get the first steps to be going agent first? They’re ready to hire a couple extra people, they’ve got their basic stuff.

1385
03:55:18.510 –> 03:55:25.030
Erik Charles: they know… they’ve got some product-market fit, maybe one product back to come up with a second product. What should they be doing to be…

1386
03:55:25.180 –> 03:55:27.590
Erik Charles: an AI-first, AI-forward company.

1387
03:55:29.670 –> 03:55:54.130
Olivia Nottebohm: Well, I mean, the beauty of a company at that size, you know, I worked at Notion, it was a little… it was a little bigger than that, but we only had 60 people, you know, when I joined, so you can just, like, literally get in a room and say, okay, you know, where should we… where’s the biggest bang for our buck in deploying AI? Because often it’s bringing in third-party vendors, like, you know, obviously we’re lucky at Box that we’re using Box on Box.

1388
03:55:54.390 –> 03:56:14.780
Olivia Nottebohm: But, you know, for the most part, people are bringing in third-party vendors, and the security of this really matters. Like, it really matters. Like, I really think that, if you are a company and you serve customers, and by virtue of the AI you leverage, you create breaches or security issues for your customers.

1389
03:56:14.780 –> 03:56:38.939
Olivia Nottebohm: And that is, like, just an incredible, terrible place to be, right? So I think there has to be a level of seriousness about, okay, what are the applications that we’re going to leverage, and what is the diligence that we do on their level of security, right? Do they promise us in our T’s and C’s that they leverage models that don’t train on our data? Right? All of these questions that…

1390
03:56:38.940 –> 03:56:48.060
Olivia Nottebohm: you know, us operators have now learned to ask for are really, really important. So first, there has to be that rigor, right? But then matching it to

1391
03:56:48.210 –> 03:57:01.009
Olivia Nottebohm: okay, what are some low-hanging fruit? And I do think the areas of customer success, just, like, prime support is prime, right? Sales, much more on the prep than on the actual engagement.

1392
03:57:01.010 –> 03:57:07.559
Erik Charles: Right? Whereas with support, you can actually have, like, self-service agents helping out your customers in a way that is…

1393
03:57:07.700 –> 03:57:09.180
Olivia Nottebohm: Pretty fantastic.

1394
03:57:09.510 –> 03:57:10.050
Erik Charles: Yeah.

1395
03:57:10.170 –> 03:57:15.799
Erik Charles: Projection time, what do you see coming in the next… and we’re gonna keep it to 12 to 18 months, you know?

1396
03:57:15.800 –> 03:57:16.270
Olivia Nottebohm: Yeah, good!

1397
03:57:16.270 –> 03:57:17.109
Erik Charles: And I even think that.

1398
03:57:17.110 –> 03:57:17.710
Olivia Nottebohm: Farther.

1399
03:57:17.710 –> 03:57:26.150
Erik Charles: I feel like saying, what do you see before the end of the year, even, being announced, released, or somebody who’s doing it really well.

1400
03:57:28.840 –> 03:57:33.079
Olivia Nottebohm: So, for myself, or for the broad society at large?

1401
03:57:33.080 –> 03:57:34.039
Erik Charles: Let’s do…

1402
03:57:34.040 –> 03:57:35.250
Olivia Nottebohm: Fox, or…

1403
03:57:35.250 –> 03:57:36.809
Erik Charles: Yeah, let’s do both.

1404
03:57:36.810 –> 03:57:38.839
Olivia Nottebohm: Yeah. Okay, I mean…

1405
03:57:38.840 –> 03:57:41.269
Erik Charles: In the business world, you know, I won’t have you, you know.

1406
03:57:41.270 –> 03:58:03.580
Olivia Nottebohm: Oh, right, right, sure. I’ve got business all day long, yeah. I think for Box ourselves, I think we’re just moving a mile a minute, and we will be automating so, so many processes, and I’m so, so excited about… you know, I mean, I literally carry around a list of all the things I wish I could do, but I can’t do, because we either don’t have the resourcing or we don’t have the bandwidth, so that is super exciting.

1407
03:58:03.580 –> 03:58:15.140
Olivia Nottebohm: that we, you know, might be able to do that. I mean, I just had a problem solving, with our community leader, right? And as… it’s like a totally different game. If you can have agents supporting communities.

1408
03:58:15.140 –> 03:58:25.899
Olivia Nottebohm: you can do a much better job at community, right, community engagement as a software company, because this is really important. It’s by role, it’s by industry, it’s by what their care about, what are their pain points, right?

1409
03:58:25.990 –> 03:58:36.209
Olivia Nottebohm: So that’s just an example of, like, all of a sudden, it was only what one person could do, and now we still have that one person, but now it’s about all of these things that they can capture and harness to do.

1410
03:58:36.330 –> 03:58:52.999
Olivia Nottebohm: I would say in the broader ecosystem, you know, I go to New York, and it is just the West Coast is, like, pretty intense on this topic, right? And I think the rest of the country will take some time, to be completely honest. So, I think a lot of

1411
03:58:53.460 –> 03:59:01.329
Olivia Nottebohm: technological advances will happen over the next 12 months, but, like, humans only move so quickly, and so I think there’ll be a bit of a lag.

1412
03:59:02.140 –> 03:59:09.529
Erik Charles: Yeah, I think it’s… I think it’s gonna be fun to see the practical applications. I like how you went down.

1413
03:59:09.810 –> 03:59:20.820
Erik Charles: specific examples. I’ve seen too many announcements on LinkedIn and other places, like, we’re now Agentic AI, and I jump into doing what?

1414
03:59:20.820 –> 03:59:22.540
Olivia Nottebohm: Yeah, for what purpose?

1415
03:59:22.540 –> 03:59:36.399
Erik Charles: Yeah, I’m tired of hearing, you know, like, it’s like, draw… it’s like, we’re a cloud company. Are you? Or do you just, like, license some private servers someplace at one point in the old days? So, thank you very much. This was fun.

1416
03:59:36.900 –> 03:59:53.180
Erik Charles: much longer, and I agree with you that live meetings… I’ve been to your box events in the past, and I always liked them. I have to give you points for not only being at the forefront of technology, but citing classics like Maslow and the 2×2 box, which proves that

1417
03:59:53.470 –> 04:00:16.969
Erik Charles: Yes, the world has changed, you better get ahead of it, but the stuff you’ve learned in the past is still applicable, you have to find out how to apply it to the new models that make things a little more efficient and a little smarter, a little quicker, and I think people should have listened to your comments on trust, governance, security, and how important that is. If you’re gonna do it, because, man, you mess it up for one customer, and it gets public, you’ve got issues, so…

1418
04:00:17.000 –> 04:00:23.120
Erik Charles: Thank you very much for joining me for this one, and thank you, Julia, for inviting us for the talk.

1419
04:00:23.420 –> 04:00:31.949
Julia Nimchinski: Our pleasure, sensational session. Thank you so much, Olivia and Eric, and where can our community go to best support you? Olivia, let’s start with you.

1420
04:00:33.050 –> 04:00:34.520
Julia Nimchinski: What’s next per box?

1421
04:00:34.520 –> 04:00:43.779
Olivia Nottebohm: Oh, just come visit Box.com, we’ve got our keynote up there, you can watch it, we’ve got a lot of information about, to Eric’s point, practical applications.

1422
04:00:44.270 –> 04:00:45.980
Julia Nimchinski: Amazing, and Eric, how about you?

1423
04:00:46.370 –> 04:00:57.020
Erik Charles: If anybody wants to get ahold of me, ericharles.com makes it very easy in the old-school way of a website. Happy to work with… always interested in advising companies on where they’re on their go-to-markets.

1424
04:00:57.140 –> 04:01:03.690
Erik Charles: Journey, and what places do they lack some of the internal skills and knowledge to help out?