Say Hi To ChatSpot.ai: The All-In-One A.I. Powered Chat App For Growing Better

Greetings. I'm Dharmesh from HubSpot and I've been getting this question a lot recently from friends and family, which is, have you tried out ChatGPT? And the first thought that pops into my head is, do you even know me? Of course I've tried out ChatGPT. Along with 100 million other people, ChatGPT from OpenAI has been the fastest growing product ever, going from 0 to 100 million users in just two months. Now I'm even more excited than the average person about ChatGPT, and the reason is it makes possible something I've been obsessed with and passionate about forever, which is Chat ux. Chat UX is the ability to put a natural language chat interface on software. So traditionally, when we build software, we take what engineers would call an imperative approach. You go through this step by step process. The user clicks here, drags there, clicks that, touches this, swipes that. At the end of a series of steps, they get the outcome they were looking for, which is, let's say, a report. Now, with the power of a generative AI in ChatGPT, we can take a declarative approach. So we can just say, as humans, here's the thing I want, AI does all the work and we just get the outcome we were looking for. We can kind of skip to the end. So, based on this excitement around generative AI and ChatGPT, I've been cranking away with the HubSpot team on an app called Chat Spot and I'd like to show it to you. And before we jump in, here's sort of how Chatbot works. It takes a natural language query in Chat, just like ChatGPT, and then using the power of ChatGPT and the HubSpot CRM and Dall E2 and Google Docs and keyword research and a bunch of other things, it actually then does something with your natural language query and boom, it gives you the outcome you are looking for. So that's the idea. So rather than trying to kind of explain it further, let's just jump in and do a demo. Now, I'm going to show you three buckets of use cases in Chat Spot. We're going to start with sales and CRM use cases, then we're going to do some reporting use cases, and then we're going to do some marketing use cases. So let's jump into the demo. So I'm going to go over here to our Chat Spot homepage, which lives at chatspot AI with just a couple of quick notes. I am here in the HS demo account. So all the data we're looking at is made up, it's fictional. And we have the nice alpha label on chatspot. Just so you know, this is very early software. All right, so let's just jump right in. Here is the chat interface for Chat Spot and along the left here we have some suggested prompts, but we're just going to start typing things in. So let's say how many contacts are there? So here Chat Spot is going to take that query, it's going to figure out what we're asking, it's going to Talk to the HubSpot API and come back with an answer, which it did, which is 15,603 contacts in the HubSpot CRM. All right, so now let's say I want to add a contact. Add contact ada74velvelace.com that's Ada Lovelace. We'll say 314 scientific clip away, Cambridge, MA 02141 and follow up in four days. Now this is something I could have done in HubSpot CRM. I could have gone to the Contacts app, click the Add Contact button, found the right properties for address, then gone to add a task, pick date four days out and add the task. But instead of that series of 20 plus clicks, we were able to do it in one simple prompt in natural language. And you'll see that I parse all this information out and we can look at this information in HubSpot. So here's the HubSpot CRM. We see that Ada Lovelace, it was even recapitalized correctly. I parsed the address. Here's a task assigned to me four days out. So other things you can do sales related is we can research a company. So I happened to mention a company called OpenAI. So we can say research company OpenAI.com or I could have said OpenAI and it comes back with some high level information. How many employees when it was founded, the description of what they do, where they're based, the Twitter Link and the LinkedIn profile all in a convenient spot. And we'll be fleshing this out as, as OpenAI. I mean, as Chat Spot gets better. All right, now this is good if you know which company you're looking for, right, you can do research on it. But what if you don't? What if you're prospecting for companies, you're looking for potential customers based on the geo you're in or the industry you're in, you can do something like find company. So we're going to do find, let's say, SaaS companies. We'll start there in Massachusetts with over 1,000 employees. So now knows what a SaaS company is, knows we're talking about companies with over a thousand employees. I can come here. HubSpot's obviously on the list, as is Toast, another local Boston area SaaS company. Once we see a company that's of interest, we can do that same research and just do a lookup, see information about the company, and then with one click, we can just add that company to HubSpot. Okay, so that was us looking up companies and adding contacts and things like that. Now we're still in sales, so let's do something slightly more fancy here. So let's say I'm going to. I want to write an email, which obviously salespeople and founders and execs do all the time. And what we're going to do is we're going to draft an email. And one of the things that ChatGPT is really, really good at is writing. It's amazing how well it does. But what we do is make it even better. So based on the email address, bhubspot.com, we're going to pull contextual information about Brian, everything we know in the CRM, everything we know from research, we know the location, we know the title, what company he works for, and it knows everything about me and HubSpot because it's chatgpt. All right, so we're just going to read this real quick. Not all of it. As the executive chairman of HubSpot, you know better than anyone the importance of leveraging technology to help businesses grow. So this is a really, really good, in my opinion, first draft. And I trained this particular model, you know, well, performing kind of outreach emails, you know, based on our own sales team and feedback. So, anyway, so I think this is a really great start. Now, if we like this particular email, we can just hit this one button, it copies it to the clipboard, and we can either paste it into Gmail or Outlook, or we can paste it into a HubSpot sales sequence if we wanted to make it part of a sales sequence. All right, so that's kind of the sales and CRM use cases. Now let's kind of shift into reporting. So let's create a report of companies added in Q4 Greek by country. Let's say we're going to walk into a meeting, we're talking about geofocus and what locations and what countries are customers and leads coming from. And it comes back with a nice table of the countries and the number of companies from those countries added in Q4. And we can check up here. And it knows what Q4 means, which is it was from October through December 31st of last year and it found 243 such countries. But now here's where it gets interesting. So this information is natively in the HubSpot CRM, right? We didn't have to do anything particularly fancy. But as I mentioned earlier, we also have the power of ChatGPT. So we can say update report and only include countries in South America. Right. Now this is not part of most reporting tools, but ChatGPT knows what South America is and what Latin America is and what EMEA is and all those things, right? So it comes back with a list of here are the countries and regions that are in South America and the number of companies we added for each. So that's a nice thing. Let's do one more ChatGPT powered thing. We'll say add column for country population. Now, once again, you know, the CRM doesn't know the population of all these countries, but ChatGPT does. We can intersect the data set we have with additional information that ChatGPT has as of 2021 until they update the model. But that's pretty cool. So anytime you kind of bring a report back, you can intersect that report based on countries or industries or whatever with information that ChatGPT knows, which is pretty much everything that's on the web. All right, so that's cool. So we generated this report. Now one of the nice things and most powerful things and most useful things about Chat Spot is that once you get information as resulting from a prompt, you can do a lot of different things with it. For instance, you can say, you know what, I like this information. Send me that as an email. It'll take that exact same thing. It'll actually drop it into my email inbox. So then I can kind of forward to someone. I can shout an email, start email discussion around a particular output. So that's in my email inbox now. But let's say I'm interested not in just getting a one time email. I want to get updates monthly. By the way, if you're wondering like, why do I keep like saying the things that I'm typing? It's not, I'm trying to insult your intelligence. I know, I know you can read, but this helps me with my kind of inner dialogue that I have the inner monologue which is like, Dharmesh, you're doing fine, Dharmesh. It's okay, if you have to redo it and take a 47th take. But I digress. But that's why I'm kind of saying the prompts to myself or to you as I'm going to. Anyway, carrying on. So we sent it via email, we got it as a monthly update. Another cool thing. So let's say we like this data, that's great, but we want it in a Google sheet. So instead of downloading it as a CSV and then uploading it into Google, we kind of skip all those steps and it just comes back with the same data and with one click access, I have that same data as a Google sheet. Or I can say, hey, you know what, I'm actually walking into a meeting. So instead of a Google sheet, give me that as a Google Slide. Actually give us the full report as a Google Slide and we open that in Google Slides and then we can format it however we choose. If we drag and drop it into, into a Google deck. Okay, so once you get information, once again, it's not just reporting, it's across the entire entire application. You can send it to yourself, you can get updates on it, you know, daily, monthly, weekly, quarterly, whatever it is, you can get it as a sheet, get it as a slide, however you choose. Okay, so let's carry on the theme of reporting and let's ask a simple atomic question, which is how much revenue did we make in make last month? Let's say. So now what's interesting here is it knows that revenue kind of ties back the deal Object in the HubSpot CRM and it looks at the amount property on the deal object, adds them all up for last month and comes back with the answer of 10 million and change. Okay, so that's cool. But what about a question like how many or who were the top five reps last month? So who are top five sales reps? It knows what our rep is and that means sales rep or account manager, account executive, whatever you call them. We want the top five and we want them for last month. So it's going to go generate a report based on that information. We'll see that it picked up the correct dates because we said last month, recording this in the beginning of March. So now one thing you can do that's cool with reports is you can say, hey, I want to, I want that report inside of HubSpot so I can save report as top sales recs last month. That's the name of the report. And what this is going to do is super cool. It's going to actually create that report inside of the HubSpot CRM. And so I will click on the View Report button and, and you will see that it takes us straight into HubSpot CRM, straight into the report builder, sets the correct date range. It gives us the data we were looking for. And now I can click on, customize and use the full power of the HubSpot Report Builder, which is super cool. And I can do anything I could have done with any other report. I can add it to a dashboard, I can apply it to a particular team, I can restrict it and control who has access to it, all the things that HubSpot reporting has app provides. So that's super cool. All right, so now lost my place here. Let's go back chat spot. All right, so those were the reporting use cases. Now let's kind of shift mindsets and let's look at some marketing related things. We'll start off with a simple analytics prompt. I'm going to pick one for the suggested ones here, which is show monthly summary of web visits for last year. So this has access to, so all the analytics Data that's in HubSpot around web visits and conversions and leads and customers and all those things and it gives back that data. I can click a single button and get that as a bar chart and it gives that to me as a bar chart, which is fun. Okay, so that's cool. Now let's say we're, let's say I'm a marketer at HubSpot and I'm curious as to what keywords HubSpot is ranking for organically. So I can ask what keywords is HubSpot.com ranking for? So now if I look at keyword research data, pull the information in real time, come back and say here are some of the keywords that are high value keywords for HubSpot. It tells us how much search volume we're getting. So if we look, let's say we look at this Instagram one, it's got 30 million plus search volume and we rank number 17. Good, not great. And the CPC is $6.39. If I click on this link, it actually takes me to the webpage within the HubSpot website that ranks for the term Instagram marketing. So now let's say As a marketer, HubSpot, I want to write a blog post about Instagram and see if we can't improve on that number 17 ranking. And the way I might do that is I'm going to do a, I'm going to Draft a blog post about the impact of generative AI on Instagram marketing. Okay, so as I mentioned earlier, one of the things that ChatGPT is great at is writing right? Now we took a simple prompt here. My idea is I want to take something that we know is popular, everyone knows what Instagram is, and intersect it with something new, which is a new eternity, which is this generative AI stuff, pull the two things together and write a blog post. Now, it does a really good job. You're probably not going to want to like directly publish the thing that it outputs, but you can iterate, you can kind of modify that prompt. You can describe how long you want it to be and what style you want it to be. It comes back. But let's say we're happy with this post as it exists. We're going to just put it inside of HubSpot, not publish it, just save it as a draft inside of HubSpot. We could have also saved it in a Google Doc. And I'm going to click on view in HubSpot and it gives me that article in HubSpot and I can come in here and I can edit it just like I would any other blog post in the HubSpot CMS. Okay? Now, as any good marketer knows, a blog post is really not complete until you have an image, because images make the post do better and they're much more engaging and social and things like that. So let's say we want to generate an image. We're going to generate an image of a retro phone. And the reason is I'm trying to draw this contrast between the cutting edge with generative AI and classic old school phones. And Instagram is a phone photo taking app. All right, so I came back with this nice image of a retro phone. Now you'll notice that this is in kind of like a pencil sketch style. And you might be wondering, well, why is it, why did it pick that? Of all the possible things it could have picked. And the reason is in Chat spot, we have this visual style library, right? So we have a bunch of different styles. These are all images of John Lennon. I'm a John Lennon fan. And, and we can pick them. So the one we just have highlighted, that's the one I'm currently doing, is, is pencil drawing. But let's say we change that. Let's say just for fun, we want to do a Salvador Dali style image. So we pick that, go back and I'm going to resubmit that same prompt, which I can do by clicking on the Edit icon and then just hitting the send. And so now we're going to get an image of a retro phone done in the style of Salvador Dali. We can click on that image. We can see. And that's pretty good, I'd say, as far as a Dall E style image. All right, so let's say we tinker with that and we can regenerate the image as many times as we want. We can provide more kind of guidance to what kind of image we want, because different visual style. But let's say now we're happy with this image. With one button click, I can say add to the recent blog post. It knows what blog post we were working on. It adds it. I can view it in HubSpot. And now we can see our article, the impact of Generative AI on Instagram marketing with a Salvador Dali style image of a retro phone. Okay, I have just one more thing. And I did not plan this. This is something I just literally just got done. The cool new thing I want to show you is this little microphone right here. So let me just show you. Create a report of Companies added in Q4, summarized by industry. There you have it, folks. Not only can you chat, you can now actually talk to your CRM. All right, now I'm done for the demo. Thanks. That's a whirlwind tour of Chat Spot. And we were only able to really cover about 10% of what the application can do. So it's moving along. It's getting better every day. It's been the most fun I've had working on a project and encoding in a while. All right, so one of the big things we believe in in HubSpot is dreaming big but iterating small. And we have big dreams for Chat Spot in terms of where we can take it in Generative AI and create the world's first smart CRM that you can actually chat with. But we like to iterate small, and that's why we're getting Chat Spot out there even while it's in alpha, because I love getting kind of user and customer feedback. And we'll update the software literally every day. It'll get better and smarter. And so I invite you to come check it out. The water's warm and the wait list is open. So I'll close with one last dad joke, since I have since I have a quota. Let's see. What did Old MacDonald say when he came across ChatGPT and Dall E2 AI. AI. Oh. Anyway, please come say hi at ChatSpot. AI. Thanks for your time. And thanks for indulging me. Hopefully you'll have a good chance to check it out. Thank you.