WEBVTT 00:00:00.000 --> 00:00:02.080 What if your AI didn't just tell you things? 00:00:02.200 --> 00:00:04.700 What if it actually built them? Yeah, imagine 00:00:04.700 --> 00:00:08.480 that. Imagine slashing weeks of work, having 00:00:08.480 --> 00:00:11.740 an AI just deliver a completed business asset, 00:00:12.019 --> 00:00:14.660 fully formed. Not just the answers, but actual 00:00:14.660 --> 00:00:18.300 usable artifacts. It's a different level. Welcome 00:00:18.300 --> 00:00:20.940 to the Deep Dive. Today, we're really digging 00:00:20.940 --> 00:00:24.079 into a fascinating guide. It's called... Perplexity 00:00:24.079 --> 00:00:27.260 Labs Guide, Five Business Use Cases and Prompting 00:00:27.260 --> 00:00:30.739 Tips. And this isn't just about finding information 00:00:30.739 --> 00:00:33.820 faster. It's about understanding how an AI tool, 00:00:34.020 --> 00:00:36.579 specifically Perplexity Labs, can fundamentally 00:00:36.579 --> 00:00:39.280 shift, shift from being just a research assistant 00:00:39.280 --> 00:00:42.840 into what the guide calls a research -to -creation 00:00:42.840 --> 00:00:45.280 engine. Which sounds pretty powerful. It does. 00:00:45.500 --> 00:00:48.659 So we'll explore its unique spot in the whole 00:00:48.659 --> 00:00:51.299 AI landscape, its quirks, its limitations, too. 00:00:51.460 --> 00:00:54.079 Yeah, those are important. really impactful business 00:00:54.079 --> 00:00:56.159 applications. Then we'll get into the precise 00:00:56.159 --> 00:00:58.179 art of actually telling it what you need, the 00:00:58.179 --> 00:01:00.859 prompting. The secret sauce. Our goal for you 00:01:00.859 --> 00:01:04.180 today, to really uncover how this tool can save 00:01:04.180 --> 00:01:08.280 massive amounts of time and honestly reshape 00:01:08.280 --> 00:01:11.780 how businesses approach intelligence. And output. 00:01:11.980 --> 00:01:14.640 It really feels like a significant leap, doesn't 00:01:14.640 --> 00:01:17.379 it? Beyond just like finding information. How 00:01:17.379 --> 00:01:20.180 so? Well, it's like moving from having this brilliant 00:01:20.180 --> 00:01:22.480 librarian who, you know, gives you all the right 00:01:22.480 --> 00:01:26.219 books to one who then takes those books and builds 00:01:26.219 --> 00:01:29.519 you a completely finished functional model right 00:01:29.519 --> 00:01:31.680 there based on her research. That's a fantastic 00:01:31.680 --> 00:01:35.150 analogy. So this building layer. What does that 00:01:35.150 --> 00:01:36.930 actually look like for someone using it? What's 00:01:36.930 --> 00:01:38.730 the practical difference? The core difference 00:01:38.730 --> 00:01:41.750 really is that it creates tangible stuff, functional 00:01:41.750 --> 00:01:44.810 business assets, not just text. Got it. Tangible 00:01:44.810 --> 00:01:47.549 assets. And to really grasp where Perplexity 00:01:47.549 --> 00:01:50.310 Lab shines, it helps to understand where it fits 00:01:50.310 --> 00:01:53.269 within the whole perplexity ecosystem. Totally. 00:01:53.310 --> 00:01:56.650 Think of it like three tiers, a three -tiered 00:01:56.650 --> 00:01:58.549 system. Each one's designed for a different level 00:01:58.549 --> 00:02:00.209 of depth. Okay. Break that down for us. Tier 00:02:00.209 --> 00:02:02.609 one. Right. So at the base, you have tier one, 00:02:02.709 --> 00:02:05.250 basic search. This is probably the perplexity 00:02:05.250 --> 00:02:08.270 most people know. Quick answers. Exactly. Simple 00:02:08.270 --> 00:02:10.710 queries, you need an immediate answer, ideally 00:02:10.710 --> 00:02:13.129 with sources. It's super fast, really efficient, 00:02:13.229 --> 00:02:15.870 like asking that librarian for just one specific 00:02:15.870 --> 00:02:18.449 fact. Boom. Okay, makes sense. Then tier two. 00:02:18.710 --> 00:02:21.210 Then you step up to tier two. Deep research mode. 00:02:21.409 --> 00:02:23.550 This is for when you need to really dig into 00:02:23.550 --> 00:02:26.750 something complex. It performs a much more in 00:02:26.750 --> 00:02:29.430 -depth analysis, pulls from a wider range of 00:02:29.430 --> 00:02:31.870 sources, and then it generates a pretty detailed 00:02:31.870 --> 00:02:34.009 research report. So like asking the librarian 00:02:34.009 --> 00:02:36.710 to gather all the books on a subject and then 00:02:36.710 --> 00:02:39.030 write up a summary? You got it. It's fantastic 00:02:39.030 --> 00:02:43.330 for understanding a topic deeply. But the output, 00:02:43.469 --> 00:02:45.990 it's still, you know, a static document, text 00:02:45.990 --> 00:02:48.430 -based. Right. Which brings us to tier three. 00:02:48.610 --> 00:02:51.189 And that brings us to tier three, Perplexity 00:02:51.189 --> 00:02:54.409 Labs. This is the real game changer we're talking 00:02:54.409 --> 00:02:56.830 about today. Okay. Labs takes that powerful research 00:02:56.830 --> 00:03:00.150 engine from the other tiers and just adds this 00:03:00.150 --> 00:03:02.889 building layer right on top. So it doesn't just 00:03:02.889 --> 00:03:05.569 give you a report. It uses its research findings 00:03:05.569 --> 00:03:08.270 to create something, a tangible, functional, 00:03:08.430 --> 00:03:10.509 often interactive business asset. So it goes 00:03:10.509 --> 00:03:13.530 beyond summarizing. Exactly. It's the only part 00:03:13.530 --> 00:03:15.689 of their ecosystem that takes you all the way 00:03:15.689 --> 00:03:18.659 from research, to a ready -to -use deliverable. 00:03:18.780 --> 00:03:21.080 Okay, so choosing the right tier seems absolutely 00:03:21.080 --> 00:03:24.419 crucial then. How should someone decide? When 00:03:24.419 --> 00:03:27.419 do you use labs versus sticking with deep research? 00:03:28.009 --> 00:03:30.210 It's actually pretty straightforward. Use labs 00:03:30.210 --> 00:03:32.870 when you need something like an interactive dashboard, 00:03:33.210 --> 00:03:35.330 maybe a functional tool. Like a calculator or 00:03:35.330 --> 00:03:37.050 something. Yeah, or a professionally designed 00:03:37.050 --> 00:03:38.909 presentation, you know, with charts and graphs 00:03:38.909 --> 00:03:41.830 built in. Or even a working web asset like a 00:03:41.830 --> 00:03:44.689 landing page. Basically anything that needs data 00:03:44.689 --> 00:03:47.610 visualization combined with that underlying research. 00:03:47.789 --> 00:03:50.389 And deep research. Stick with deep research when 00:03:50.389 --> 00:03:52.849 all you need is the analysis itself. A literature 00:03:52.849 --> 00:03:55.310 review, a text -based report, just gathering 00:03:55.310 --> 00:03:57.069 information without needing a finished thing. 00:03:57.189 --> 00:03:59.330 So boil it down for us, the main takeaway for 00:03:59.330 --> 00:04:02.270 choosing. Labs builds, deep research analyzes. 00:04:02.389 --> 00:04:05.030 Simple as that. Okay. But before you jump into 00:04:05.030 --> 00:04:07.270 building all this cool stuff, there are some 00:04:07.270 --> 00:04:10.650 critical things to know first. Prerequisites, 00:04:10.669 --> 00:04:14.289 limitations. Right, absolutely. First off, labs 00:04:14.289 --> 00:04:17.149 isn't free. It requires a perplexity pro plan. 00:04:17.589 --> 00:04:20.970 And importantly, that pro plan comes with a usage 00:04:20.970 --> 00:04:24.290 limit. Currently, it's 50 labs queries per month. 00:04:24.470 --> 00:04:27.930 50? Okay, that's not unlimited. No. Which really 00:04:27.930 --> 00:04:30.829 emphasizes the need to be thoughtful, strategic 00:04:30.829 --> 00:04:33.269 with your prompts. You can't just mess around 00:04:33.269 --> 00:04:35.790 endlessly. Right. Each query counts. Yeah. And 00:04:35.790 --> 00:04:38.990 this next point, honestly, it's a big one. A 00:04:38.990 --> 00:04:42.069 real security consideration. Oh. As of right 00:04:42.069 --> 00:04:45.589 now, there is no way to revoke a shared link 00:04:45.589 --> 00:04:48.089 to a lab's asset once you've created it. Wait, 00:04:48.129 --> 00:04:50.670 no way at all? None that's documented. So if 00:04:50.670 --> 00:04:52.790 you build, say, a dashboard with sensitive info 00:04:52.790 --> 00:04:55.709 and you share that link, that link provides permanent 00:04:55.709 --> 00:04:59.089 public access. Permanent public access. You absolutely 00:04:59.089 --> 00:05:01.769 have to plan your sharing strategy very, very 00:05:01.769 --> 00:05:03.509 carefully. It's like a permanent digital footprint 00:05:03.509 --> 00:05:06.189 you can't erase. That 50 query limit and the 00:05:06.189 --> 00:05:08.209 unrevocable link, I mean, that fundamentally 00:05:08.209 --> 00:05:10.110 changes how you'd approach using this, doesn't 00:05:10.110 --> 00:05:12.089 it? It really does. You know, I still wrestle 00:05:12.089 --> 00:05:14.810 with prompt refinement myself sometimes, especially 00:05:14.810 --> 00:05:18.490 knowing a credit is on the line beat. It forces 00:05:18.490 --> 00:05:20.670 you to be super precise right from the start. 00:05:20.790 --> 00:05:22.949 That makes sense. So how do you actually access 00:05:22.949 --> 00:05:25.269 labs? Are there different ways? Yeah, there are 00:05:25.269 --> 00:05:28.620 two main ways. First is direct labs access. Pretty 00:05:28.620 --> 00:05:30.180 simple. You just click the little light bulb 00:05:30.180 --> 00:05:33.060 icon and the perplexity interface takes you straight 00:05:33.060 --> 00:05:35.639 there. Then you write your prompt telling it 00:05:35.639 --> 00:05:37.800 what to research and what kind of asset to build. 00:05:38.480 --> 00:05:40.839 Labs just handles both parts automatically. And 00:05:40.839 --> 00:05:43.079 the second way? The second is called the research 00:05:43.079 --> 00:05:46.379 to labs workflow. And this one's quite powerful, 00:05:46.459 --> 00:05:48.000 actually. How does that work? You start your 00:05:48.000 --> 00:05:50.360 session in the standard deep research mode first. 00:05:50.860 --> 00:05:52.839 Do your initial information gathering there. 00:05:53.259 --> 00:05:55.579 Okay, so you refine the research first. Exactly. 00:05:55.699 --> 00:05:57.579 Once you're happy with that research foundation, 00:05:57.899 --> 00:06:00.540 then you switch over to labs mode. The cool part 00:06:00.540 --> 00:06:03.839 is the system keeps all the context from your 00:06:03.839 --> 00:06:07.339 research phase and uses that information to build 00:06:07.339 --> 00:06:10.139 your asset. So it gives you a bit more control 00:06:10.139 --> 00:06:12.759 over that initial research part. Okay, that sounds 00:06:12.759 --> 00:06:15.579 useful. But what about those limitations you 00:06:15.579 --> 00:06:17.779 mentioned beyond the permanent links? Right. 00:06:17.879 --> 00:06:20.019 Well, there are iteration challenges, unlike 00:06:20.019 --> 00:06:22.279 some other AI tools where you can kind of chat 00:06:22.279 --> 00:06:24.579 back and forth to refine something. Yeah, like 00:06:24.579 --> 00:06:28.120 tweak this, change that color. Exactly. Perplexity 00:06:28.120 --> 00:06:31.180 Labs tends to treat each significant change as 00:06:31.180 --> 00:06:34.800 a whole new request. Oh. Which often means running 00:06:34.800 --> 00:06:38.079 a new query and, yep, using another one of your 00:06:38.079 --> 00:06:40.180 precious monthly credits if you want to alter 00:06:40.180 --> 00:06:42.350 your dashboard after it's made. It's not really 00:06:42.350 --> 00:06:44.670 designed for that conversational sculpting. Gotcha. 00:06:44.790 --> 00:06:48.910 Less iterative, more one shot. Pretty much. And 00:06:48.910 --> 00:06:51.009 it's worth repeating the security issue. Seriously, 00:06:51.170 --> 00:06:54.089 that inability to revoke shared links is a huge 00:06:54.089 --> 00:06:56.569 concern. So definitely don't use it for. Right. 00:06:56.610 --> 00:06:58.910 Don't use perplexity labs for anything with private 00:06:58.910 --> 00:07:01.250 client data, personal financial info, internal 00:07:01.250 --> 00:07:04.490 secrets, basically any sensitive content you 00:07:04.490 --> 00:07:06.050 wouldn't want potentially out there forever. 00:07:06.209 --> 00:07:08.709 So not for highly confidential internal reports, 00:07:08.870 --> 00:07:11.360 at least not yet. Not right now, no. It's just 00:07:11.360 --> 00:07:13.959 too risky with those permanent links. So summing 00:07:13.959 --> 00:07:16.660 that up, what's the biggest single caution you'd 00:07:16.660 --> 00:07:19.779 give new users? For me, without a doubt, it's 00:07:19.779 --> 00:07:21.660 the permanent public links for sensitive data. 00:07:21.839 --> 00:07:24.660 Think very carefully before sharing. Okay. Good 00:07:24.660 --> 00:07:28.839 advice. Mid -roll sponsor read. okay let's get 00:07:28.839 --> 00:07:31.019 into the exciting part the actual applications 00:07:31.019 --> 00:07:34.959 the guide highlights five specific ways this 00:07:34.959 --> 00:07:37.339 tool can really transform business workflows 00:07:37.339 --> 00:07:40.040 this is where it gets really interesting yeah 00:07:40.040 --> 00:07:43.000 let's dive in first up use case one the prospect 00:07:43.000 --> 00:07:46.040 research dashboard okay anybody in sales or marketing 00:07:46.040 --> 00:07:49.980 knows prospect research can be Well, soul crushing 00:07:49.980 --> 00:07:52.279 sometimes. Chuckle softly. Tell me about it. 00:07:52.399 --> 00:07:55.240 It's tedious, fragmented. You're juggling expensive 00:07:55.240 --> 00:07:58.300 databases, manually scraping LinkedIn, setting 00:07:58.300 --> 00:08:00.879 up news alerts, trying to jam it all into some 00:08:00.879 --> 00:08:03.579 messy spreadsheet. It's just incredibly inefficient. 00:08:03.879 --> 00:08:06.240 Been there. So how does labs change that? With 00:08:06.240 --> 00:08:08.399 labs, you use a single detailed prompt. Yeah. 00:08:08.540 --> 00:08:10.339 Basically, you tell it to act as your personal 00:08:10.339 --> 00:08:12.819 research team. Okay. And it builds this comprehensive, 00:08:12.959 --> 00:08:15.519 interactive prospecting dashboard in minutes. 00:08:15.860 --> 00:08:19.240 Minutes. Seriously. Yeah, minutes. Imagine prompting 00:08:19.240 --> 00:08:22.519 it like this. Research high -growth DTC that's 00:08:22.519 --> 00:08:25.079 direct consumer e -commerce brands in fashion 00:08:25.079 --> 00:08:27.850 and beauty. Look for growth signals like new 00:08:27.850 --> 00:08:30.410 product launches or recent funding rounds. Then 00:08:30.410 --> 00:08:32.909 create an interactive dashboard with lead scoring, 00:08:33.250 --> 00:08:36.549 detailed company profiles, contact info if possible, 00:08:36.710 --> 00:08:39.610 and an industry breakdown. Aim for maybe 20 to 00:08:39.610 --> 00:08:41.929 30 prospects. And what do you actually get back 00:08:41.929 --> 00:08:43.950 from a prompt like that? You get a professional 00:08:43.950 --> 00:08:46.309 -grade interactive dashboard. Usually in about 00:08:46.309 --> 00:08:49.210 10 minutes. Wow. Yeah. It includes summary metrics, 00:08:49.409 --> 00:08:51.529 a searchable database of prospects with company 00:08:51.529 --> 00:08:54.889 details, maybe estimated revenue, contacts, detailed 00:08:54.889 --> 00:08:57.409 profiles with source links. With sources. Nice. 00:08:57.690 --> 00:08:59.750 Yep. Plus data visualization, like where they're 00:08:59.750 --> 00:09:02.389 located or funding rounds. And often it even 00:09:02.389 --> 00:09:04.649 throws in some AI -generated outreach templates 00:09:04.649 --> 00:09:07.070 to get you started. That's incredible. The business 00:09:07.070 --> 00:09:10.610 impact. Huge. It replaces easily 8 to 15 hours 00:09:10.610 --> 00:09:13.830 of manual grunt work. That's like a 95%, maybe 00:09:13.830 --> 00:09:16.470 more, time -saving. Goodness. Plus, it's way 00:09:16.470 --> 00:09:18.909 better organized. It frees up sales teams to 00:09:18.909 --> 00:09:21.389 actually sell, you know, do the high value stuff. 00:09:21.549 --> 00:09:24.190 OK. Any pro tips for this one? Definitely. Always 00:09:24.190 --> 00:09:26.350 double check the key data points. Maybe spot 00:09:26.350 --> 00:09:29.139 check like. 10, 15 % of the entries, especially 00:09:29.139 --> 00:09:31.440 contacts or revenue. Right. AI isn't perfect. 00:09:31.639 --> 00:09:34.840 Exactly. And labs seems to work best data accuracy 00:09:34.840 --> 00:09:37.320 wise for mid -market companies right now. So 00:09:37.320 --> 00:09:39.700 be specific in your prompts about industries 00:09:39.700 --> 00:09:42.399 and growth indicators. Got it. Okay. What's use 00:09:42.399 --> 00:09:45.379 case two? Use case two, the high impact research 00:09:45.379 --> 00:09:48.360 driven landing page. Ooh, interesting. Landing 00:09:48.360 --> 00:09:50.919 pages are tough. They are. Traditional builders 00:09:50.919 --> 00:09:53.399 give you templates, sure, but they don't give 00:09:53.399 --> 00:09:55.320 you the messaging, the words that actually sell. 00:09:55.659 --> 00:09:57.779 True. That usually needs deep market research, 00:09:57.879 --> 00:10:00.440 customer research, which most businesses just 00:10:00.440 --> 00:10:03.100 don't have time for. Right. But labs can uniquely 00:10:03.100 --> 00:10:05.799 combine that deep research, competitive research, 00:10:06.039 --> 00:10:09.220 customer research with the actual landing page 00:10:09.220 --> 00:10:12.039 creation. So it writes the copy based on research. 00:10:12.360 --> 00:10:16.220 Exactly. Resulting in a data backed asset. Picture 00:10:16.220 --> 00:10:19.019 this prompt. OK, we're launching an email marketing 00:10:19.019 --> 00:10:21.639 software as a service targeting small business 00:10:21.639 --> 00:10:24.899 owners who run online courses. Research the messaging 00:10:24.899 --> 00:10:27.299 of competitors like ConvertKit and ActiveCampaign. 00:10:27.559 --> 00:10:30.519 Also, dig into customer feedback on sites like 00:10:30.519 --> 00:10:34.460 G2, Capterra, maybe Reddit. Okay, so real user 00:10:34.460 --> 00:10:36.980 opinions. Then create a professional one -page 00:10:36.980 --> 00:10:39.399 landing page. Needs a data -driven headline, 00:10:39.679 --> 00:10:42.759 maybe an interactive ROI calculator, a feature 00:10:42.759 --> 00:10:45.259 comparison chart against those competitors, and 00:10:45.259 --> 00:10:47.620 customer testimonials pulled directly from your 00:10:47.620 --> 00:10:50.440 research findings. Keep the design modern, clean, 00:10:50.620 --> 00:10:52.659 maybe blue and white. What makes that different 00:10:52.659 --> 00:10:54.740 from just using a template builder? It's the 00:10:54.740 --> 00:10:57.070 research integration. The AI isn't guessing. 00:10:57.110 --> 00:10:59.750 It's analyzing real user reviews, looking at 00:10:59.750 --> 00:11:02.309 how competitors position themselves, and crafting 00:11:02.309 --> 00:11:04.490 messaging based on that. That's powerful. And 00:11:04.490 --> 00:11:07.049 the output is professional. Interactive elements, 00:11:07.169 --> 00:11:09.070 those data -backed testimonials. It can even 00:11:09.070 --> 00:11:11.350 generate extra assets like those comparison charts 00:11:11.350 --> 00:11:13.690 or pricing analysis on the side. And the business 00:11:13.690 --> 00:11:16.350 impact. You're creating landing pages where the 00:11:16.350 --> 00:11:18.730 messaging is almost guaranteed to resonate better. 00:11:19.129 --> 00:11:21.269 because it's based on your target audience's 00:11:21.269 --> 00:11:24.649 own words and a real market analysis. It takes 00:11:24.649 --> 00:11:26.830 so much guesswork out of copywriting. Yeah, that's 00:11:26.830 --> 00:11:29.610 huge. Okay, use case three. Use case three, the 00:11:29.610 --> 00:11:33.230 real -time social media trend tracker. Ah, for 00:11:33.230 --> 00:11:35.610 content creators, keeping up with trends is relentless. 00:11:36.090 --> 00:11:38.409 Totally. You struggle to spot emerging trends, 00:11:38.529 --> 00:11:40.629 and by the time you jump on one, it's often already 00:11:40.629 --> 00:11:42.950 peaked or fizzling out. You miss the window. 00:11:43.370 --> 00:11:47.029 So labs can create a live dashboard. Think of 00:11:47.029 --> 00:11:48.889 it as your personal content strategy intelligence 00:11:48.889 --> 00:11:51.690 hub. It's constantly monitoring the social media 00:11:51.690 --> 00:11:54.230 landscape for you. Live dashboard? How does that 00:11:54.230 --> 00:11:56.730 work? You'd prompt something like, create an 00:11:56.730 --> 00:12:00.009 AI industry trend tracker. Monitor social media 00:12:00.009 --> 00:12:02.269 discussions from the last 30 days covering AI 00:12:02.269 --> 00:12:04.870 tools and startup news. Check across Twitter, 00:12:05.049 --> 00:12:07.850 LinkedIn, Reddit. The dashboard needs to show 00:12:07.850 --> 00:12:10.809 top trending topics, engagement metrics, maybe 00:12:10.809 --> 00:12:13.129 sentiment analysis, which platform is leading 00:12:13.129 --> 00:12:15.690 the discussion, and critically list common... 00:12:15.720 --> 00:12:17.659 questions people are asking. OK, what does that 00:12:17.659 --> 00:12:21.000 dashboard actually show you? It's dynamic. You 00:12:21.000 --> 00:12:23.100 see topics gaining momentum, maybe with little 00:12:23.100 --> 00:12:25.379 indicators. You see the overall sentiment, positive, 00:12:25.539 --> 00:12:29.860 negative. And this is gold, a question mining 00:12:29.860 --> 00:12:33.279 feature. It literally extracts the actual questions 00:12:33.279 --> 00:12:36.200 real users are asking online. Oh, wow. That's 00:12:36.200 --> 00:12:39.059 a content goldmine right there. Isn't it? The 00:12:39.059 --> 00:12:41.360 business impact is transforming your content 00:12:41.360 --> 00:12:43.639 strategy from just reacting to being proactive. 00:12:44.220 --> 00:12:46.940 You spot trends early, jump on them intelligently, 00:12:47.059 --> 00:12:49.460 and create content that directly answers the 00:12:49.460 --> 00:12:52.360 questions your audience actually has. Data -driven 00:12:52.360 --> 00:12:54.799 content planning. Very cool. What's number four? 00:12:55.240 --> 00:12:58.379 Use case four. Competitive intelligence, but 00:12:58.379 --> 00:13:00.559 with brand sentiment. Okay. Competitive analysis 00:13:00.559 --> 00:13:03.879 usually feels a bit dry. Features, pricing. Exactly. 00:13:04.080 --> 00:13:05.980 It's often just feature checklists. It misses 00:13:05.980 --> 00:13:07.779 how the market feels about your competitors. 00:13:08.019 --> 00:13:10.500 And that feeling, that sentiment, is where the 00:13:10.500 --> 00:13:12.960 really big strategic opportunities often hide. 00:13:13.159 --> 00:13:15.799 So Labs adds the feeling part. Precisely. It 00:13:15.799 --> 00:13:17.919 combines that traditional product analysis, features, 00:13:18.100 --> 00:13:20.559 pricing, et cetera, with real -time brand sentiment 00:13:20.559 --> 00:13:22.779 pulled from social media, forums, review sites, 00:13:22.940 --> 00:13:24.960 all in one place. Give me an example prompt. 00:13:25.299 --> 00:13:28.110 Sure. Develop a competitive intelligence dashboard 00:13:28.110 --> 00:13:31.889 for, say, Booking .com. Focus on Expedia and 00:13:31.889 --> 00:13:34.509 Airbnb. Research and compare their features, 00:13:34.730 --> 00:13:38.230 pricing models, loyalty programs. Also, track 00:13:38.230 --> 00:13:40.350 brand sentiment for all three over the last 60 00:13:40.350 --> 00:13:43.370 days. The dashboard should feature a side -by 00:13:43.370 --> 00:13:46.070 -side comparison table, a sentiment score for 00:13:46.070 --> 00:13:47.970 each brand with examples of positive and negative 00:13:47.970 --> 00:13:50.230 comments, and maybe a market share visualization. 00:13:50.710 --> 00:13:53.529 And the outcome of that? You get this rich, multi 00:13:53.529 --> 00:13:56.139 -layered view. You don't just see that, you know, 00:13:56.139 --> 00:13:59.000 competitor A has feature X. You also see that 00:13:59.000 --> 00:14:01.340 competitor A's customers are constantly complaining 00:14:01.340 --> 00:14:04.259 online about hidden fees. Ah, so you see their 00:14:04.259 --> 00:14:07.039 weaknesses, their pain points. Exactly. It reveals 00:14:07.039 --> 00:14:09.299 not just feature advantages, but perception advantages 00:14:09.299 --> 00:14:12.700 and disadvantages, the business impact. It allows 00:14:12.700 --> 00:14:14.919 for a much more sophisticated competitive strategy. 00:14:15.419 --> 00:14:17.399 You can position your product not just based 00:14:17.399 --> 00:14:19.279 on what it does, but on how it makes customers 00:14:19.279 --> 00:14:21.919 feel, maybe directly addressing the pain points 00:14:21.919 --> 00:14:24.399 driving competitive customers crazy. It's like 00:14:24.399 --> 00:14:26.259 knowing the emotional battlefield. That's a great 00:14:26.259 --> 00:14:28.320 way to put it. Okay, the final one, use case 00:14:28.320 --> 00:14:31.240 five. And this one. This one's pretty impressive. 00:14:31.399 --> 00:14:35.059 Use case five. The go -to -market strategy presentation. 00:14:35.710 --> 00:14:38.870 OK. GTM strategies. Those take ages to put together. 00:14:39.009 --> 00:14:42.269 Ages. Creating a comprehensive GTM strategy deck 00:14:42.269 --> 00:14:44.750 that's your whole plan for launching a new product 00:14:44.750 --> 00:14:47.309 or service, right? It's usually a massive undertaking. 00:14:48.029 --> 00:14:51.250 Weeks, sometimes months of research, data analysis, 00:14:51.470 --> 00:14:53.710 financial modeling, slide design. A huge amount 00:14:53.710 --> 00:14:56.659 of work. Labs. Well, Labs does almost all the 00:14:56.659 --> 00:14:59.259 heavy lifting. It condenses potentially weeks 00:14:59.259 --> 00:15:01.740 of that work into minutes. Okay. I need to hear 00:15:01.740 --> 00:15:04.179 this prompt. Right. So something like, develop 00:15:04.179 --> 00:15:06.779 a GTM strategy presentation for a new online 00:15:06.779 --> 00:15:09.440 learning platform. Let's say it focuses on AI 00:15:09.440 --> 00:15:12.100 skills for professionals. Research the market 00:15:12.100 --> 00:15:32.539 size and growth potential. And what does Labs 00:15:32.539 --> 00:15:36.960 actually produce from that? professionally designed 00:15:36.960 --> 00:15:40.179 PowerPoint or Google Slides presentation, usually 00:15:40.179 --> 00:15:42.559 in about 15 to 20 minutes. 15 to 20 minutes for 00:15:42.559 --> 00:15:45.799 a whole GTM deck. Yeah. It's polished. It has 00:15:45.799 --> 00:15:48.899 charts, graphs, data -driven content, all backed 00:15:48.899 --> 00:15:50.740 by the sources it used. It includes all the key 00:15:50.740 --> 00:15:53.279 GTM elements you'd expect, market opportunity 00:15:53.279 --> 00:15:57.000 analysis, you know, TAM, SAM, MRM, total addressable 00:15:57.000 --> 00:15:59.059 market, serviceable available market, serviceable 00:15:59.059 --> 00:16:01.889 obtainable market. Basically, how big the pie 00:16:01.889 --> 00:16:04.350 is and how much you can realistically get. Right. 00:16:04.549 --> 00:16:06.470 Gives you a competitive differentiation matrix, 00:16:06.690 --> 00:16:09.330 detailed customer personas based on its research, 00:16:09.490 --> 00:16:12.509 even a suggested KPI framework, key performance 00:16:12.509 --> 00:16:15.269 indicators to track success. Honestly, it might 00:16:15.269 --> 00:16:17.870 be the most powerful example of pure efficiency 00:16:17.870 --> 00:16:21.110 gain here. Whoa, moment of wonder. I mean, imagine 00:16:21.110 --> 00:16:23.870 creating a professional GTM deck in 20 minutes. 00:16:23.970 --> 00:16:26.610 That's practically a superpower, especially for 00:16:26.610 --> 00:16:29.029 startups or even established businesses trying 00:16:29.029 --> 00:16:32.450 to move faster. It really is. Condensing two, 00:16:32.610 --> 00:16:35.149 three weeks of intense work into a 20 -minute 00:16:35.149 --> 00:16:38.549 automated process. The agility that gives you. 00:16:38.990 --> 00:16:41.970 Being able to test and refine strategies at speeds 00:16:41.970 --> 00:16:45.009 that were just unimaginable before. It's kind 00:16:45.009 --> 00:16:47.090 of mind -blowing. Looking at all five, which 00:16:47.090 --> 00:16:49.149 one seems like it would have the most immediate 00:16:49.149 --> 00:16:51.450 widespread impact for many businesses listening? 00:16:51.769 --> 00:16:55.730 Oof, tough call. They're all strong. But I think 00:16:55.730 --> 00:16:59.480 probably the GTM strategy. Just because the time 00:16:59.480 --> 00:17:02.740 savings are so enormous on such a critical, complex 00:17:02.740 --> 00:17:05.200 task. Yeah, hard to argue with saving weeks of 00:17:05.200 --> 00:17:08.480 work. Okay, so we've seen these amazing use cases. 00:17:08.720 --> 00:17:11.220 But the guide stresses that the secret to unlocking 00:17:11.220 --> 00:17:13.799 these high -quality outputs, it isn't just the 00:17:13.799 --> 00:17:15.880 tool itself, it's the instructions you give it. 00:17:16.000 --> 00:17:18.279 Absolutely. Professional results demand an investment 00:17:18.279 --> 00:17:20.720 in crafting your prompt. It's not magic. It's 00:17:20.720 --> 00:17:22.880 instruction. So how do you craft a great prompt 00:17:22.880 --> 00:17:25.119 for labs? The guide mentions three pillars. That's 00:17:25.119 --> 00:17:27.180 right. The three pillars of a great prompt. First 00:17:27.180 --> 00:17:29.470 is comprehensive context setting. Meaning? Meaning 00:17:29.470 --> 00:17:31.589 you need to clearly tell the AI what role it 00:17:31.589 --> 00:17:33.890 should play. Like you are a senior market research 00:17:33.890 --> 00:17:36.150 analyst working for a venture capital firm. Give 00:17:36.150 --> 00:17:38.630 it a job title. Exactly. And provide specific 00:17:38.630 --> 00:17:41.509 business context, your industry, your target 00:17:41.509 --> 00:17:44.049 market, your goals for this asset. Give it a 00:17:44.049 --> 00:17:46.529 persona and a clear mission. Okay. Pillar one, 00:17:46.569 --> 00:17:49.069 context. What's pillar two? Pillar two is detailed 00:17:49.069 --> 00:17:51.349 output specification. You have to be crystal 00:17:51.349 --> 00:17:53.789 clear about the deliverable. Be specific. Super 00:17:53.789 --> 00:17:56.589 specific. Don't just say, make a report. Say, 00:17:57.400 --> 00:18:00.500 Create an interactive dashboard or generate a 00:18:00.500 --> 00:18:02.720 10 -slide executive presentation in PowerPoint 00:18:02.720 --> 00:18:05.279 format. Specify the functionality you need. The 00:18:05.279 --> 00:18:07.740 dashboard must include a searchable table filtered 00:18:07.740 --> 00:18:10.599 by industry. Even specify design requirements. 00:18:10.839 --> 00:18:14.000 Use a consulting -grade professional design or 00:18:14.000 --> 00:18:16.480 adhere to our brand color scheme of blue and 00:18:16.480 --> 00:18:19.039 white. The more detail, the less guesswork for 00:18:19.039 --> 00:18:21.980 the AI. Got it. Context -specific output. Pillar 00:18:21.980 --> 00:18:23.799 three. Pillar three is clear research focus. 00:18:24.000 --> 00:18:26.819 You need to guide its research. Specify timeframes. 00:18:27.200 --> 00:18:30.400 Only consider data from the last 30 days. Indicate 00:18:30.400 --> 00:18:33.299 source preferences, focus on social media conversations, 00:18:33.759 --> 00:18:38.039 or prioritize reviews from G2 and Keptera. And 00:18:38.039 --> 00:18:40.079 definitely name the specific competitors you 00:18:40.079 --> 00:18:43.160 wanted to analyze. This helps it zero in on precisely 00:18:43.160 --> 00:18:45.400 the information you need. rather than boiling 00:18:45.400 --> 00:18:48.359 the ocean. Okay, context, output specs, research 00:18:48.359 --> 00:18:50.799 focus. And the guide mentioned something about 00:18:50.799 --> 00:18:52.779 demanding quality. Yeah, this is a great tip. 00:18:52.900 --> 00:18:55.019 You can dramatically improve the output quality 00:18:55.019 --> 00:18:57.940 by including specific quality indicators right 00:18:57.940 --> 00:18:59.700 there in your prompt. Instead of just hoping. 00:18:59.920 --> 00:19:02.980 Right. Don't hope, demand quality. Use phrases 00:19:02.980 --> 00:19:05.420 like, create a deliverable with a consulting 00:19:05.420 --> 00:19:08.799 -grade professional design. Or, the final output 00:19:08.799 --> 00:19:10.980 should be of executive presentation quality. 00:19:11.420 --> 00:19:14.549 Ah, setting the standard up front. Exactly. The 00:19:14.549 --> 00:19:16.569 dashboard must have full interactive functionality. 00:19:17.029 --> 00:19:19.450 The landing page should follow modern web design 00:19:19.450 --> 00:19:22.430 standards. All claims must be supported by data 00:19:22.430 --> 00:19:24.789 -driven visualizations. You're telling it the 00:19:24.789 --> 00:19:27.890 benchmark it needs to hit. Why is being so detailed 00:19:27.890 --> 00:19:30.430 with prompting especially crucial here, maybe 00:19:30.430 --> 00:19:33.089 more than with other AI tools, given the credit 00:19:33.089 --> 00:19:35.130 limits and iteration challenges? It's really 00:19:35.130 --> 00:19:37.890 about demanding those specific, high -quality 00:19:37.890 --> 00:19:40.549 deliverables up front. You want to maximize the 00:19:40.549 --> 00:19:41.910 chance of getting what you need on the first 00:19:41.910 --> 00:19:44.569 try without burning through those limited credits 00:19:44.569 --> 00:19:47.930 on revisions. Precision minimizes waste. Makes 00:19:47.930 --> 00:19:51.609 perfect sense. So let's talk optimization and 00:19:51.609 --> 00:19:54.750 the bottom line, the business impact. How do 00:19:54.750 --> 00:19:57.650 you get the absolute maximum value from labs? 00:19:58.029 --> 00:20:00.089 Well, building on the prompting, number one is 00:20:00.089 --> 00:20:02.890 strategic prompt planning. Seriously, spend those 00:20:02.890 --> 00:20:04.849 extra few minutes crafting a really detailed 00:20:04.849 --> 00:20:06.930 prompt before you hit generate and use a credit. 00:20:07.269 --> 00:20:10.130 That upfront investment pays off massively in 00:20:10.130 --> 00:20:12.650 the output quality. Measure twice, cut once, 00:20:12.789 --> 00:20:15.890 or prompt once, ideally. Duckles, exactly. Precision 00:20:15.890 --> 00:20:18.890 in, perfection out, hopefully. Second, always 00:20:18.890 --> 00:20:21.250 maintain a data verification workflow. Fact checking. 00:20:21.529 --> 00:20:23.710 Yep. Especially for key data points, numbers, 00:20:23.910 --> 00:20:26.170 contacts, anything you're going to use externally 00:20:26.170 --> 00:20:29.009 or base major decisions on, always sanity check 00:20:29.009 --> 00:20:31.069 it. Good practice for any AI output, really. 00:20:31.230 --> 00:20:33.549 Absolutely. And third, have an asset management 00:20:33.549 --> 00:20:36.150 strategy. When labs create something, a presentation, 00:20:36.529 --> 00:20:39.920 dashboard code, beta files. Download everything. 00:20:40.019 --> 00:20:42.980 Local copies. Yeah, for offline access, backup, 00:20:43.259 --> 00:20:45.420 maybe integrating it into other workflows or 00:20:45.420 --> 00:20:48.000 tools. And yes, I know we sound like a broken 00:20:48.000 --> 00:20:50.980 record, but to be extremely strategic about who 00:20:50.980 --> 00:20:53.579 gets those permanent share links, maybe don't 00:20:53.579 --> 00:20:55.759 share them at all if possible, just use the downloaded 00:20:55.759 --> 00:20:58.579 files. Right. Manage the assets, manage the links. 00:20:58.779 --> 00:21:01.660 Okay, the business impact. Let's quantify this 00:21:01.660 --> 00:21:05.529 ROI. The time saving seems central. They're staggering. 00:21:05.630 --> 00:21:07.589 It's not just incremental efficiency. It's a 00:21:07.589 --> 00:21:09.630 fundamental shift in how quickly you can operate. 00:21:09.789 --> 00:21:11.670 Give us those numbers again. Okay. Prospect research. 00:21:12.289 --> 00:21:15.049 Traditionally, maybe 8 to 15 hours, right, with 00:21:15.049 --> 00:21:17.769 labs, potentially 10 minutes. That's easily a 00:21:17.769 --> 00:21:22.490 95 % time saving or more. 95%. Wow. And the go 00:21:22.490 --> 00:21:25.390 -to -market strategy. Two, three weeks of intense 00:21:25.390 --> 00:21:27.670 work. Labs can draft that initial comprehensive 00:21:27.670 --> 00:21:30.630 deck in maybe 20 minutes. That's like a 98 % 00:21:30.630 --> 00:21:32.390 plus time saving. That's almost unbelievable. 00:21:32.809 --> 00:21:35.170 It allows businesses to just test more ideas. 00:21:35.730 --> 00:21:38.910 Respond to market shifts way faster. And crucially, 00:21:38.990 --> 00:21:41.490 it lets you focus your human talent, your expensive 00:21:41.490 --> 00:21:44.349 creative people on high value strategic thinking, 00:21:44.569 --> 00:21:48.410 interpretation, decision making. Instead of just 00:21:48.410 --> 00:21:51.230 low value data gathering and slide formatting. 00:21:51.640 --> 00:21:53.920 Leveraging human creativity where it actually 00:21:53.920 --> 00:21:56.400 counts the most. That's the goal. So stripping 00:21:56.400 --> 00:21:58.559 it all back, what's the single biggest benefit 00:21:58.559 --> 00:22:02.140 of this incredible speed? I'd say unprecedented 00:22:02.140 --> 00:22:05.180 business agility coupled with a much sharper 00:22:05.180 --> 00:22:07.839 strategic focus. Move faster, think smarter. 00:22:08.240 --> 00:22:10.099 Okay, let's wrap this up. Big picture takeaway 00:22:10.099 --> 00:22:12.519 for everyone listening. Perplexity Labs isn't 00:22:12.519 --> 00:22:14.400 just another AI that answers your questions. 00:22:14.559 --> 00:22:16.839 It's fundamentally different. How so? It's an 00:22:16.839 --> 00:22:19.700 AI that builds. It creates professional, actionable 00:22:19.700 --> 00:22:22.480 business assets for you. It truly bridges that 00:22:22.480 --> 00:22:25.039 gap, doesn't it? Between just raw research and 00:22:25.039 --> 00:22:26.920 actually having something tangible, something 00:22:26.920 --> 00:22:29.759 created. Automating potentially weeks of work 00:22:29.759 --> 00:22:32.599 into mere minutes. It really does. And the key, 00:22:32.720 --> 00:22:35.519 the absolute key to unlocking all that power. 00:22:35.640 --> 00:22:38.779 Let me guess, the prompt. Precise, well -crafted 00:22:38.779 --> 00:22:41.339 prompts. It really is all in the ask. So here's 00:22:41.339 --> 00:22:43.579 a final thought to leave you with. This profound 00:22:43.579 --> 00:22:46.819 shift we're seeing, moving from AI doing just 00:22:46.819 --> 00:22:51.160 analysis to doing active creation, it implies 00:22:51.160 --> 00:22:53.579 something significant. What's that? It means 00:22:53.579 --> 00:22:55.759 the competitive advantage, certainly in the coming 00:22:55.759 --> 00:22:58.720 years, will likely go to those individuals and 00:22:58.720 --> 00:23:01.980 those businesses who master these new research 00:23:01.980 --> 00:23:05.140 to creation workflows first. So the question 00:23:05.140 --> 00:23:07.730 becomes... Will you be leading that transformation 00:23:07.730 --> 00:23:10.190 or will you be trying to catch up to it? It's 00:23:10.190 --> 00:23:11.829 definitely an exciting time to be building things. 00:23:11.970 --> 00:23:14.750 Lots to think about. Absolutely. If this deep 00:23:14.750 --> 00:23:16.950 dive sparked your curiosity, maybe showed you 00:23:16.950 --> 00:23:19.049 some new ways to approach your work, remember 00:23:19.049 --> 00:23:21.349 to explore our other deep dives on similar topics. 00:23:21.630 --> 00:23:24.369 Lots more to uncover. Until next time, keep learning. 00:23:24.609 --> 00:23:26.549 And keep building. Outro music.