Product Demo: State of GTM AI Tech 2025: Adoption, Impact & Future Trends

1671
05:05:58.190 –> 05:06:05.449
Julia Nimchinski: of course, and we are in for a real trip here. Leehon Hicken, welcome to the show again.

1672
05:06:07.190 –> 05:06:08.010
Lihong Hicken: Julia.

1673
05:06:08.410 –> 05:06:12.979
Lihong Hicken: thank you for having me. I think this will. Mine’s the last session, right? It’s gonna be the last.

1674
05:06:12.980 –> 05:06:14.780
Julia Nimchinski: Yeah, yeah.

1675
05:06:14.780 –> 05:06:15.450
Lihong Hicken: Awesome.

1676
05:06:15.670 –> 05:06:34.869
Julia Nimchinski: It’s a good it’s a good transition, because, as as you could hear Gainsight community talk, and you’ve been part of the community roundtable. So now we are gonna focus on the survey results. Put it together. The state of Gtm AI tech

1677
05:06:35.080 –> 05:06:38.090
Julia Nimchinski: 2025. Let’s dive into it.

1678
05:06:39.040 –> 05:06:58.680
Lihong Hicken: Yeah, all right, let’s get into it. So all right. Friends of hard skill exchange in the AI Led Growth Summit. What an epic three-day event on the topic of AI-led growth. You probably have heard a lot from the panelists and the experts talking about AI, and finally, it is your time to hear

1679
05:06:58.800 –> 05:07:15.459
Lihong Hicken: your voice. Right? Survey is your voice, your opinion. What you think about go to market is going. Where is AI is going to contribute? So. Thank you, Julie, Julia, for giving me the opportunity to present the hard work of the survey results. I think you’re putting a lot of effort in

1680
05:07:15.460 –> 05:07:31.394
Lihong Hicken: building and designing this. So later on I can. I might invite you to talk about some of your thoughts in this but before we started that. Let me just share my screen here. Want to do a small introduction of who is they said, and what we do, and then,

1681
05:07:32.760 –> 05:07:40.329
Lihong Hicken: go into the the idea of the survey design. And what kind of result we’re looking for sounds good, Suya.

1682
05:07:40.330 –> 05:07:41.570
Julia Nimchinski: Amazing yeah.

1683
05:07:41.960 –> 05:08:11.430
Lihong Hicken: All right. Sounds good. So they said, if you look at the woes of kind of the universe, of getting feedback from your customers, you have what customers say on the top, and you also have what customer do on the bottom, the quantitative side on the right, and the qualitative on the left hand side and then. Now we break this down into the research method of knowing your customers. If you want to know what customers say

1684
05:08:11.430 –> 05:08:30.359
Lihong Hicken: you do surveys right, send out a survey is pretty scalable. But if you want to understand why customers say certain things, why they do this really go deeper, understanding, the motivation and the intent buying, intent on that. You do interviews, or sometimes you call customer conversations, or sometimes a fancy way is focus group

1685
05:08:30.360 –> 05:08:54.280
Lihong Hicken: right? And then vice versa. On the action side, you either do user testing to test out why customer do things or use user analytic tools to understand customers behavior. All right. So our funding team comes from a company called a usertesting.com, we user testing went public in 2021. So

1686
05:08:54.490 –> 05:08:57.930
Lihong Hicken: growing and building user testing. We know how powerful

1687
05:08:58.170 –> 05:09:24.290
Lihong Hicken: the value and how how powerful the qualitative insights are. But so is interview. So so problem with qualitative insight is that it’s super valuable, but it’s not very scalable. It’s very manual. It’s kind of a 1 off. It takes forever for you to schedule calls, analyze the results and take action. Right surveys, on the other hand, is very scalable.

1688
05:09:24.520 –> 05:09:30.759
Lihong Hicken: but it’s surveys are designed to count things like S number of percent of people like

1689
05:09:31.140 –> 05:09:37.169
Lihong Hicken: Y and Z. So what without the context of the Y, it’s not very helpful.

1690
05:09:37.350 –> 05:10:04.310
Lihong Hicken: So it’s and all these different things live in different system which cause more painful siloed results. Siloed customer insight that cannot be put it together. So what we’re trying to do is build a feedback OS system that can combine all these feedback into one and give you kind of a chat. Gtv. Style to build on your customers feedback. They can ask the bot about what does my customer like? You know what is my customer’s concern and pain point?

1691
05:10:04.310 –> 05:10:23.630
Lihong Hicken: But as of today, we are able to do. The AI surveys AI interviews and AI user testing which allow you to get qualitative insight at scale. This is kind of our unique point. All right. So so much about what this has done, and let’s dive into a few examples.

1692
05:10:23.870 –> 05:10:50.210
Lihong Hicken: In our survey, which we, Julia, we went out about 5 business days ago. We got about 74 people responding to it like, for example, one of the questions we asked is like, How do we envision AI transforming your go to market strategy in the next 12 months. Now this is a traditional survey question, and the majority of people say the will say it will. Automate routines.

1693
05:10:50.380 –> 05:10:58.249
Lihong Hicken: Do all the routine tasks like, improve your efficiency. Okay, so what does that mean? What kind of tasks can we improve?

1694
05:10:58.370 –> 05:11:24.780
Lihong Hicken: Is it sales task? Is it marketing task? Is it Sdr work, or is it a post sales work? We don’t know. Like, we only know this is this is what a traditional survey on the quantitative side give you the result. Now let me show you another concept in the AI conversational survey. A person can answer this one, and you, AI, is instruct to follow up on this and asking them why

1695
05:11:24.780 –> 05:11:36.220
Lihong Hicken: you believe automated routine tasks will be a key transformation for your go-to-market strategy in the next 12 months. So this person say, Oh, AI can do it at scale.

1696
05:11:36.220 –> 05:11:46.939
Lihong Hicken: No Sdr. Can do so. His belief is, Sdring is a number again hitting a number day by day. So AI power assistant will step in entirely. So

1697
05:11:47.500 –> 05:12:13.169
Lihong Hicken: in his mind the automated routine task is, he wanted to automate routine tasks of Sdr work, and other people might have different tasks. So you can see here it give you additional layer of insights, the traditional, so they wouldn’t be able to get to you and the AI will ask these questions dynamically based on what they are asking. So this is kind of a conversation level. What you can do in one on one conversation, you can do

1698
05:12:13.410 –> 05:12:15.169
Lihong Hicken: right, going into deeper.

1699
05:12:15.520 –> 05:12:30.970
Lihong Hicken: What are some of the interesting insights? Same same question on that which is going to the most transformative go to market strategy in the next 12 months. So majority of the people answer that.

1700
05:12:31.070 –> 05:12:50.749
Lihong Hicken: you know, automating routine tasks, including data gathering this building research and administrative tasks. A lot of the sales related tasks can be replaced by AI, and this is a majority of people saying it. And then we also have

1701
05:12:51.140 –> 05:13:07.559
Lihong Hicken: another 2 3rd of people believe in, you know, more personalized outreach, better decision making in customer health score. So this is some of this is still in prospecting in new sales side.

1702
05:13:07.680 –> 05:13:35.210
Lihong Hicken: and some of that is in the customer. Existing customer management manage the customer’s health. Scores, manage predict customer care, and then retention side. So you can see here overwhelmingly. Most people talk about AI in go to market for acquiring new businesses, and then a significant chunk, maybe a quarter of them talk about

1703
05:13:35.360 –> 05:13:39.349
Lihong Hicken: better decision making process in the existing customer flow.

1704
05:13:39.560 –> 05:13:52.749
Lihong Hicken: And then we also don’t forget about some negative boys. So they are a significant worry and concerns about adopting AI. And some people just believe AI is going to be

1705
05:13:53.070 –> 05:14:12.369
Lihong Hicken: too overused, and that will lose the sincerity in human really have a real impact in go to market. So that’s kind of a base view in terms of this question. So I’m going to pause here before we go in there. So, Julia, when you are designing this survey question.

1706
05:14:12.637 –> 05:14:25.749
Lihong Hicken: what what would you? What was in your head? And what would what did you want to find out? And was this result a bit surprising to you? And what was your experience? So tell me a little bit about that. If that’s sorry, it’s a lot of questions.

1707
05:14:26.170 –> 05:14:33.659
Julia Nimchinski: Yeah, happy to share. I actually want to focus on the technology itself. On, they said.

1708
05:14:34.454 –> 05:14:45.739
Julia Nimchinski: I’m not the easiest customer as you as you experience, I’m a design freak. Yeah. And I use a very popular alternative normally today said

1709
05:14:46.100 –> 05:15:03.489
Julia Nimchinski: so it was really hard to get used to the new user interface, and just generally, you know, to just to start using the tech. But then, once I tried it, I I if you can just showcase the experience itself, how it shows the insights.

1710
05:15:03.860 –> 05:15:04.510
Lihong Hicken: Yeah.

1711
05:15:04.880 –> 05:15:24.399
Julia Nimchinski: So it’s I mean super dynamic. It’s really it’s it’s the opposite of static, by all means, and you can dive in every conversation and see the depths of insights you can share this link with, like, I mean, I have just, automatically, a lot of sales use cases on my mind.

1712
05:15:24.770 –> 05:15:34.630
Julia Nimchinski: So this is a really cool way to expand like, I don’t know if any research into a meaningful conversation in your outreach.

1713
05:15:35.600 –> 05:15:36.999
Julia Nimchinski: Yeah, I’ll pause here.

1714
05:15:38.050 –> 05:15:58.735
Lihong Hicken: Yeah, it’s amazing. Yeah. So what I want to say, you know you cannot you? You need to be a platform agnostic. So you cannot say. Ss, survey is the history. But you can say they said, it’s the future, right? Yeah. So 1 1 of the things I was we were when we are designing the survey.

1715
05:15:59.130 –> 05:16:16.960
Lihong Hicken: for the questions we really want to understand why we allow we turn on AI insight which you can tell the AI to follow up one question before moving on, or do 2 plus follow up before moving on. So you have a full control of how deep

1716
05:16:17.010 –> 05:16:31.610
Lihong Hicken: you want AI to dig deeper in in terms of the specific topic, and the whole conversation experience make it a little bit more natural. Typically most of the days. That conversational survey will have less questions. But

1717
05:16:31.670 –> 05:16:47.990
Lihong Hicken: you like you, said Julie. You want a lot of information, and you wanted to gather a lot of information. I totally understand it. So thank you for those who went through this process, and, you know, give us a lot of valuable insight.

1718
05:16:48.190 –> 05:16:49.360
Lihong Hicken: I actually

1719
05:16:49.530 –> 05:17:17.080
Lihong Hicken: had a fortune of looking at several of these kind of large scale survey results, and there’s some interesting threats and findings I will share in the following few minutes. Let’s let’s go on here so earlier, I show, I think, Julia, you mentioned one of the things you really really like is that you can see each conversation. And my favorite thing is, actually, I really like the action item, because the

1720
05:17:17.130 –> 05:17:24.210
Lihong Hicken: a lot of people might say, Yeah, we don’t do surveys, because survey result is really bad. Bad. You know, we don’t have really high response rates

1721
05:17:24.220 –> 05:17:37.799
Lihong Hicken: based on Mckenzie study. The number. One reason why people don’t fill out surveys is that they think the organization would not take action on the feedback they give. So think about this.

1722
05:17:37.800 –> 05:17:56.330
Lihong Hicken: If my feedback is going to a place where dreams dies. I’m not going to give you feedback. Okay, straight. So that’s why we actually build out action items where AI will suggest, based on the conversation. Do this, do this. So this is my own account, where I use this for AI code prospecting

1723
05:17:56.330 –> 05:18:22.230
Lihong Hicken: which I taught in. I think it’s a last event, Julia, we talk about using how to use AI survey to do co-prospecting and book a hundred meetings in about 30 min time. So I’ve been running that campaign I recorded a step-by-step guide where you can use that to do that as well, and you can use, they said, platform for free to do that.

1724
05:18:23.520 –> 05:18:45.310
Lihong Hicken: So if you’re interested, ping me in the slack channel in the Hsc slack channel, I’m there. The Hsc group is my kind of community of go to market leaders who are really into AI. So sometimes I hang out there. So if you are interested in hearing that recording and that

1725
05:18:45.440 –> 05:19:12.439
Lihong Hicken: I’m happy to share a copy. But you know, taking action, I cannot tell you how important it is. I’m using this to engage with my potential customers, existing customers and take action. I usually do it within 24 h of time, that being, said Julia. So for those of people who have done the survey, and they said, they do want a copy of the result. We need to probably send it to them.

1726
05:19:12.907 –> 05:19:20.292
Lihong Hicken: Do you want me to send it, or you want to engage with the audience, sing it. Luckily there are not many of them.

1727
05:19:21.080 –> 05:19:27.860
Lihong Hicken: I think it would be a great step for you to just get acquainted with the community and just take the next step.

1728
05:19:28.680 –> 05:19:40.230
Lihong Hicken: Sounds good. So these people, I’ll be sending you guys a copy of your own survey results. And if people want the entire survey result, happy to provide that as well. Okay.

1729
05:19:40.360 –> 05:20:00.420
Lihong Hicken: so that’s going on next. So if you’re interested in using AI survey like just to understand what AI conversational survey is. You can try it with the QR code, scan it. Benefit of that is, you can actually talk to the AI using your voice to speech function. And that’s

1730
05:20:00.530 –> 05:20:22.630
Lihong Hicken: speech text function on the phone. I personally like that better because I don’t need to type. I’m lazy, so give it a try. It’s free. And then, if you pay it with my AI code prospecting session that actually will help you generate some more leads specifically from people who said, No, thank you. Those are the ones that you can get more attention.

1731
05:20:22.890 –> 05:20:23.730
Lihong Hicken: All right.

1732
05:20:24.550 –> 05:20:40.480
Lihong Hicken: I know you guys have sit through 3 days of a boring presentation on that. So I want to make it a bit more fun. We use the AI survey result and created a podcast that’s generated by AI. This podcast is about

1733
05:20:40.480 –> 05:20:56.680
Lihong Hicken: 15 min long. So if you later on get this presentation just click on that, it’s kind of pretty hilarious. It’s a male AI and female AI. Talking about the things and argue about the different opinions because the survey results shows like there is.

1734
05:20:56.680 –> 05:21:15.519
Lihong Hicken: There’s some trends. But there’s also conflict in people’s opinion about AI, so they will be like, I almost feel like they’re talking in the mind of my thoughts. So it’s pretty fun. So go ahead and listen to this. Podcast. It’s about 15 min long. Now, if you are

1735
05:21:15.590 –> 05:21:20.510
Lihong Hicken: too time, crunch and don’t have time to listen to the 15 min, podcast

1736
05:21:21.100 –> 05:21:27.630
Lihong Hicken: I will spend the remaining 10 min to show you some of the interesting funding I have so

1737
05:21:28.220 –> 05:21:46.829
Lihong Hicken: interesting finding here is, remember, we asked about the company size we see here. It looks like the hard skill community or AI growth community is majority smaller companies. And the interesting thing we found is like

1738
05:21:47.270 –> 05:22:13.089
Lihong Hicken: in terms of budget contribution. Here the most common is 0 to 5% and 6 to 10%. What we found here is like overwhelmingly smaller companies are more willing to spend more budget on AI driven solutions, and the larger company are not as much, I mean, in terms of budget. It’s kind of hard to tell, because larger company might have a bigger budget, which percentage is low. But

1739
05:22:13.090 –> 05:22:29.820
Lihong Hicken: this is just a perception. Smaller companies are more eager to embrace, and trying AI, and more willing to pay for it. So that’s some of the interesting things I found, Julia. Were you surprised to find this, or you think this is as expected.

1740
05:22:30.130 –> 05:22:38.672
Julia Nimchinski: Oh, that’s a funny thing about data, Levon. It’s it’s never quite objective without the context correct. So

1741
05:22:39.100 –> 05:22:39.989
Julia Nimchinski: bye, bye.

1742
05:22:40.230 –> 05:22:52.660
Julia Nimchinski: actually share it on my linkedin feed, and we made it crazy hilarious. Because these days it’s really hard to make people contribute to any survey, and we were running out of time.

1743
05:22:52.910 –> 05:22:59.920
Julia Nimchinski: So yeah, I’m assuming that the data reflects on the people that basically got into.

1744
05:23:00.690 –> 05:23:22.429
Lihong Hicken: Yeah, I guess if we run the survey a bit 2 more weeks, and after the event I will keep the survey. Keep going, and then we can keep collecting the result. And when I’m curious to see like for the attendees who are joining this, if you haven’t done the survey, and you want to contribute that things might change in the next 2 weeks, and we’re happy to send you the updated result afterwards.

1745
05:23:23.120 –> 05:23:27.989
Lihong Hicken: Cool? All right. Another interesting thing here is

1746
05:23:28.820 –> 05:23:58.291
Lihong Hicken: We asked, like, okay, what works the best? Right? Well, what are some of the tips that you can share with people? So this one lucky thing is, we actually turn on AI inside, on, on this one. So we do know the why, and people give detailed instructions and tips that they work so overwhelming. The number one people have chosen is like sales again, maybe because our target audience, our community is mostly sales, but it’s a sales is the number one in terms of,

1747
05:23:58.670 –> 05:24:04.830
Lihong Hicken: have already seen the most measurable impact followed by marketing and followed by

1748
05:24:04.920 –> 05:24:30.690
Lihong Hicken: people who didn’t do anything. Okay? In terms of some cool AI tips, including, like analyzing large volume of closed loss data, using elephant AI creating assets without hiring designers. So those are the the notable mentions that people about, you know, 6 people mentioned this. I want to say.

1749
05:24:31.265 –> 05:24:47.850
Lihong Hicken: I used to be in the space of win loss analysis which help you find out why new business people, you know why new sales, you know. They talk to you, and they went cold, and they never buy like. Why did they not buy? And and you?

1750
05:24:48.040 –> 05:24:49.850
Lihong Hicken: I’m actually a little bit

1751
05:24:50.260 –> 05:25:08.579
Lihong Hicken: a little bit skeptical about this method. Because when you say analyzing large volume of closed loss data is basically data that the sales people enter like when you enter close loss, and then the sales force will pop up asking why and then users typing. So it’s not direct

1752
05:25:08.630 –> 05:25:36.730
Lihong Hicken: customer data is more indirect customer data. But in order to get direct customer data, a lot of companies will hire agencies, and they’ll do weight loss and analysis interviews, and it costs like a thousand dollars per interview to get that. But it is very valuable. It is amazingly valuable. And you don’t need that many you need like 20 interviews. You’re good for a year. Right? So we actually are lowering our AI win loss that allow you to

1753
05:25:36.730 –> 05:26:02.240
Lihong Hicken: automatically send out the AI interview, asking them why they didn’t buy as soon as the deal is closed. So there’s a fresh. So if you ask someone like 3 months later, why did they didn’t buy that? Well, I don’t remember. I forgot. So make it fresh and easy to do, and then also a fraction of the cost of the traditional ways. So those, I think, will be really interesting way for sales team if you wanted to learn, grow, increase your win rate.

1754
05:26:02.240 –> 05:26:19.360
Lihong Hicken: trying AI win loss because it’s garbage in and garbage out. It’s about the quality of data you get. So step number one is, analyze the data you already have. Number 2 is, get a higher, more direct, fresh data from your customer. That’s actually the golden rail. If you can do that.

1755
05:26:21.100 –> 05:26:41.013
Lihong Hicken: I would say. In here we’ll see overwhelmingly. A lot of people say they do use AI, but there’s nothing they will claim. It’s like a secret sauce like the same use case where everyone use use AI to write content linking posts like nothing unique. So I think

1756
05:26:41.570 –> 05:26:55.750
Lihong Hicken: You know, the the people who feel like they’re using AI as a secret sauce is not the the majority. So people are. Still. That’s maybe that’s why people are coming to AI customer like growth to learn about other people’s secret sauce. Okay.

1757
05:26:56.290 –> 05:27:00.529
Lihong Hicken: anything you want to add to Julia before we move on here.

1758
05:27:02.170 –> 05:27:04.939
Julia Nimchinski: Nothing I could think of. Yeah, this is great.

1759
05:27:05.890 –> 05:27:22.010
Lihong Hicken: Cool. Let’s move on to that. Okay, this one was really interesting. Julia. You asked which AI power solution are most integrated with your go to market tech stack. That one actually kind of surprised me. I thought, the

1760
05:27:22.010 –> 05:27:35.750
Lihong Hicken: it will be because the previous answers is all sales sales. And this one is our marketing automation tools and a few mentions of the tools are, of course, chat, Gdp

1761
05:27:35.750 –> 05:28:00.110
Lihong Hicken: design AI gamma. I think that’s the the sales deck AI sales deck, icon, and elephant. AI. So I actually check out these ones after I see the result. Just kind of check it out, see if they work, and I actually signed up gamma as well. So it’s really cool tools. Are there any tools you have seen, Julia, that you think should add to this list?

1762
05:28:01.690 –> 05:28:07.830
Julia Nimchinski: Wow a plenty. I’m trying to be Switzerland when it comes to tools. So yeah,

1763
05:28:09.990 –> 05:28:23.813
Lihong Hicken: okay, Miss Switzerland, I will not make you choose a vendor because your vendor specific yeah, vendor, agnostic, right, cool, okay, let’s move on here. So

1764
05:28:24.780 –> 05:28:44.998
Lihong Hicken: so this is a a lot of this is a very sneaky buyer. Intent question, Julia. You ask which AI brand or vendor do you currently use or are consider using for your go to market initiative. So of course, besides, all the Lrms out there like Chat Gdp, you know,

1765
05:28:45.620 –> 05:29:05.559
Lihong Hicken: Google’s all these ones there are a few mentions here. Gamma is one again. Easy. Gen aws, 6th sense, fathom napkin. So there’s a lot of AI tools that out in the market. So check it out. If you haven’t heard about that might be something useful for your toolkit.

1766
05:29:06.370 –> 05:29:11.679
Lihong Hicken: All right. I won’t ask more question on this, Julia, because I know you’re vendor agnostic.

1767
05:29:11.780 –> 05:29:38.840
Lihong Hicken: Let’s move on. This one is actually one of the more open, ended question. We kind of want to ask people like, what AI powered innovation do you think will revolutionize? Go to market in 2030? So that’s like 5 years from now. So it’s a bit, I mean, in AI years this is like a decade or even longer. So people have

1768
05:29:39.100 –> 05:29:59.080
Lihong Hicken: have fun with this like, talk about AI native operators. Hyper personalizations enhance account based marketing, you know, automate sales and marketing. So real time comes. Customer conversation. I want to say, some of this is actually

1769
05:29:59.270 –> 05:30:29.080
Lihong Hicken: already happening, like AI customer conversations. We already doing it, they said, AI content creation. Well, which blog post is not created by AI these days. You can also tell right linking posts. I think some part of the sales and marketing has been automated. I think some people mentioned about clay, you know, and a bunch of tools together to automate this stuff. One thing I think.

1770
05:30:29.520 –> 05:30:53.389
Lihong Hicken: Julia, a lot of us will talk about is the hyper personalization. Will be. I think that would be really interesting trend. Because right now, what we can achieve is we can achieve personalization during a single interaction. But how do we get personalization at the customer journey level, the entire customer journey level? That would be something very interesting, because

1771
05:30:53.460 –> 05:31:10.880
Lihong Hicken: then the AI have a memory of your past behavior, your previous complaints, your worries. You don’t need to repeat again, so it can truly be marvelous if we can achieve that. So that’s my wish as well, and

1772
05:31:11.300 –> 05:31:14.179
Lihong Hicken: and most people would will think

1773
05:31:14.990 –> 05:31:23.160
Lihong Hicken: You know, most people’s attitude toward the future of AI is pretty positive on that. So

1774
05:31:23.440 –> 05:31:36.099
Lihong Hicken: I think that’s it. We wrap it up with big vision for 2,030, and this wrap up the results of the AI go to market

1775
05:31:36.200 –> 05:31:39.430
Lihong Hicken: 2025. AI. Toolings.

1776
05:31:40.040 –> 05:31:49.469
Julia Nimchinski: Thank you so much, Leehan. Real treat. Everyone just tried these set excited to hear your feedback in our slack channel, and

1777
05:31:49.670 –> 05:31:58.500
Julia Nimchinski: that wraps up day 3 of the conference. Again. Just don’t hesitate to reach out to me directly.

1778
05:31:58.710 –> 05:32:04.130
Julia Nimchinski: The any ideas, any feedback. What conversation you’d like to have.

1779
05:32:04.340 –> 05:32:16.250
Julia Nimchinski: And we’re gonna be back on March 20th with AI practice sessions focused on most innovative marketing methods and demand Gen. So

1780
05:32:16.710 –> 05:32:23.389
Julia Nimchinski: spoiler alert. We’ll keep you posted, and thanks so much for being part of this

1781
05:32:24.010 –> 05:32:27.180
Julia Nimchinski: excited thanks. Leon again. Bye, bye.