Hello, everyone, and welcome to Breaking Ground, our quarterly launch event where we gather to celebrate social media, share strategy, and learn what's new in the Sprout platform. Today's Breaking Ground, the social-powered AI engine, is a milestone event where we'll reveal one of our biggest product innovations yet. While we're waiting for people to log in, please navigate over to the live chat and share. Outside of work, what is one task you wish AI could handle for you that it doesn't do yet today? I'll share one to get us started. I recognize this is a little bit lazy on my end, but if AI could somehow put on my duvet cover over my duvet, the thought of doing laundry would be so much less daunting. I would look forward to it so much, much more. Ok, yes, laundry. It seems like I'm not alone there.
That makes me feel a little bit less lazy. Oh, mopping the floors. That's a great one. I'm getting a lot of cleaning themes. Really, we need AI to clean for us. Ok, I see that there are some duvet fellows with me here. I appreciate that. Ooh, yard work. That's a really good one. Litter box, I saw. I have the cat litter robot. I don't know if we have any cat pet owners here, but that's not necessarily AI, but automation. It has changed my life. Taxes, meal prep. Wow, these are amazing. I am so inspired now. Whoever can get on the AI cleaning capabilities, that would be great. All right, looks like we have most of everyone in the event now, so it looks like we're ready to kick off. Hi, everyone. My name is Tazi Flory.
I'm a Product Marketing Manager here at Sprout, and I will be your host for today's event. We are so thrilled to have you join us for Breaking Ground, the social-powered AI engine. Before we get started, we need to cover just a few housekeeping items. During today's event, we're going to be making some forward-looking statements. This is just a reminder that while today's session will include some very exciting previews of what's in development, Sprout customers should make decisions based on what's already live and available today. During this broadcast webinar, we encourage you to engage with the Sprout team and with your fellow attendees. You guys all already have 10 out of 10 love seeing an active chat. Please continue to use the live chat to share your reactions and engage with each other.
As you take part in the live chat, there are two things we ask you to remember. First, please keep things respectful. We're here to create a positive and inclusive space for everyone. Second, as the chat gets active, it's likely some questions may be lost. If you have a question you'd like answered by the Sprout team or any of our guests today, please enter it in the Q&A, and we'll do our best to respond to questions live during today's event. As always, if you have a technical issue or a question specific to your account, it's best to reach out to support or your dedicated account manager. Finally, our most asked question, yes, this event is being recorded. You can expect a link to the replay before the end of the week.
We have a lot of exciting content today and big reveals in store for you. We'll kick things off with a look at how evolving AI changes the job of social. Then we'll learn agentic AI that's powered by social data with our biggest product reveal yet. Next, we'll dig into new and upcoming releases across the Sprout platform. We'll wrap things up with a crowd favorite, our enter-the-chat panel discussion with special customer guests and AI experts. Make sure you stick around after the event for additional Q&A with all of our speakers. Now, who's ready to talk about the role of social in the era of AI? Let's jump in. The universe of social data is expanding, and there is more business-critical data on social networks than ever before.
As we learned in last quarter's Breaking Ground, the customer journey has shifted to social. Social search is the new SEO. Social content is the new marketing site map. Social shopping is the new storefront, and social care is the new help desk. Social is where product discovery happens, and it's where customers want to interact with your brand. For every part of this funnel, there is a wealth of data representing real interactions from the people your brand is trying to reach. When we zoom out, we're talking about 5.4 billion people sharing hundreds of billions of messages, consuming tens of petabytes of data on social daily. This data is rich, unstructured, disorganized, and streaming in real time. Out there in this awe-inspiring universe of social data, there's information your business needs to drive decisions.
When we look across this universe, we find valuable market data, like insights from customers we need to improve our products or campaigns, customer sentiment we need to understand our brand health, competitive intelligence we need to stay ahead of the game, and audience and cultural trends we need to know so we can stay relevant and plugged in. The problem is you do not have the tools to act on the social data at scale. This is frustrating for social media teams, and it is disappointing for customers. 55% of users say they think brands are listening to them on social, but they are not taking actions often enough. There are missed opportunities. Your job as a social media leader is to solve this problem. You need to navigate the social universe and transform human interactions into insights so you can meet customers where they are and close open loops.
You cannot do this without the right tools. Here's the good news. The latest generation of AI models are incredibly well-suited to help us navigate and harness the data. New versions of large language models are faster. They're more capable and more accurate for analysis. Today, LLMs have smarter logical reasoning to break problems down step by step, larger context windows to handle more complex tasks from start to finish, and better tool calling so AI can leverage the right tools to get the job done. AI is reliable and ready to handle professional workflows in the platforms we use every day. Let's take a minute to talk about some different types of AI and how they each fit into your workflows. Generative AI is the type of AI you're probably already familiar with.
It creates new content based on patterns it has learned from massive amounts of data. It's great for writing captions, generating visuals, or summarizing posts. You make a request, and it delivers. The next type of AI you need to understand agentic AI. this is a newer, more advanced type of AI that can reason and take agentic AI goes a step further. Instead of just generating responses, it can take actions towards a goal, reasoning through steps, deciding what to do next, and even using tools to complete a task. You can think of it like a teammate who not only drafts ideas, but can also help you execute them. Agentic AI, we have technology that processes huge quantities of data and analyzes it using a step-by-step process that mimics human reasoning.
For example, your typical workflow for analyzing social data today might look like writing the right queries, running a few reports, investigating and comparing results, and eventually formatting your findings so you can share them out to stakeholders across your business. When we do this process manually, it really can feel overwhelming. You may fall down a few rabbit holes or run into some dead ends in your data. Relevant context is not always obvious. Even if you truly enjoy running and creating reports, which no judgment, we're all data nerds here, this requires a lot of time and effort that takes away from energy you could be putting towards more effective data storytelling, strategic work, campaign creation, or customer engagement. When you agentic AI into that workflow, AI can speed run through the same steps you would take and more.
It can examine the context and give you a more complete picture of your campaign performance, customer sentiments, or a viral trend that's popping off. With agentic AI on your team, you get to flex new skills. Your task becomes writing clear prompts, asking curious questions, and overseeing the results. The social media managers who embrace this new way of working with agentic AI are the leaders in this new era of social intelligence. This isn't just about a few productivity gains here and there. This is about owning an entirely new workflow that's yours to own because you are the part of your organization that is closest to your customers. Now, you can break down silos and enable every department across your business with real-time social intelligence.
With your new optimized workflow, you can send customer feedback to product research and development, know whether a PR crisis is likely to blow over or need a response, sort real human fans from the bots and offer better support. You can design compelling moment-in-time brand campaigns, distribute promos and sell more products, and enable sellers with insights to stay ahead. Every part of your business can move faster. This is your AI engine powered by your social data, generating insights at a speed that was impossible to achieve before. Now, I am very excited to pass the mic to our amazing VP of AI Product, Meg Christolin i, who's going to show you how Sprout AI is powering a new era of social intelligence. She'll be unveiling some game-changing new capabilities her team is working on and showcasing our vision for the future. Welcome to Breaking Ground, Meg.
Thank you, Tazi. Hello, everyone. I'm Meg Christolin i , and I am so excited to have the opportunity to share more about Sprout AI and introduce some of the innovation we've been working on over the past year. AI is by no means new to Sprout. We have been investing in AI to power up capabilities for over a decade. Looking at the numbers, just last year, Sprout customers sent over 14 million messages with the assistance of Sprout AI. We have over 40 custom AI models in play built specifically for social data, empowering key workflows across the Sprout platform. These models power features like optimal send times, which help teams determine the best days and times to schedule posts to maximize organic reach. They power automations, helping ensure that conversations needing a response make it into your queue with the necessary priority and context.
They also power sentiment, suggested hashtags, and alt text. These capabilities and many others make up the foundation of Sprout AI. Now, agentic AI, we have the opportunity to connect these tools and the millions of data points across the Sprout portfolio to help you solve one of the biggest challenges that you face today: making sense of all of your social data and getting to insights faster. This is not only about making things easier for your social team, but about turning social data into a source of learning and action for your entire company. This helps them make decisions and also move at the speed that you need to be moving. Over the past few months, we have been partnering with our customers to understand your workflows and bring social intelligence to life. Today, I am excited to introduce you to Trellis.
Trellis is Sprout's AI agent designed to unleash social intelligence across your organization. You can think of Trellis as a member of your team and a strategic partner within your company. We know you need to be able to reach insights faster, learn from those insights, and use them confidently to take action. Let's take a closer look at how Trellis Chat works with social listening. The conversational interface is the easiest way to explore social data and generate insights. You can chat with Trellis to complete all sorts of tasks like trend analysis, competitive research, or campaign performance reporting based on listening topics and themes. You likely know the feeling of being in a meeting or getting an email or a message where someone asks you for some type of social data. What they do not understand is everything that goes into getting that insight ready to share.
Now, you and your team can just ask Trellis. If you're not sure what to ask, we've created some conversation starters with suggestions. When you kick off a conversation, this triggers Trellis to perform a task using the whole ecosystem of AI tools available to it. Behind the scenes, Trellis is using those foundational proprietary AI models that we talked about earlier. Trellis can plug into these existing tools across Sprout. Here, I'm asking Trellis to show me what's trending in a listening topic for our pet retail brand over the last week. It looked across thousands of conversations, and you can see it revealed this week's hottest themes, including a harness issue, a cute pets campaign, and our new pet color cam. Key points about these themes are all summarized here for me.
This one about harness issues really stands out to me, and I want to learn more. We know that insights are not always just one way. Sometimes you have questions, and you need to have a back-and-forth conversation to fully understand the data. This used to mean hours of extra filtering and research, but with Trellis, you can just ask and get answers instantly. One of the challenges with social data and social insights can be that spikes are helpful, trends are useful, but they are often too high level. You need to drill into what triggered the spike or trend to get down to the specific messages and profiles at the origin. Trellis can help with this. Here, I am asking Trellis if we could take a closer look at the harness issue over the past three months.
Now I can see more information about what exactly is going on. With this closer look, I'm seeing three major spikes and persistent concerns. Now I can share this insight back with our product development team, emphasizing how big of an ongoing issue this is. This insight will guide their improvements for harness design, and hopefully, our next trending theme on social can be something way more fun, like the best holiday gifts for your pets. The chat interface for Trellis is in beta for listening today. This is already in the hands of several customers. Some of you, I know, are here today, and thank you so much for being early adopters. Trellis will continue to roll out to professional, advanced, and enterprise customers between now and the end of the year. However, I've got a special surprise.
Every qualifying social listening customer who's on this Breaking Ground call today will be getting exclusive early access to Trellis over the next month. Please watch your email inboxes, and we'll be reaching out to you when Trellis is activated for your account. In the coming months, we'll be expanding Trellis into more areas like publishing. With the help of Trellis, you'll be able to conversationally optimize your content calendar, get ideas for new content, compare content or campaign performance, and discover ways to improve your planned posts. Trellis will also work for you in the Smart Inbox. You could ask Trellis to explain trends across conversations, understand the root cause of volume changes, or collaborate with Trellis to prepare responses for customers. We've been talking a lot about chat, but Trellis is so much more than just the chat experience.
In addition, there's a growing ecosystem of Trellis agents built to provide insights, sometimes before you even know to ask. One great example of this is the NewsWhip Monitoring Agent, which is now part of the Trellis ecosystem. This always-on agent is designed to detect emerging issues proactively so comms teams can get ahead of them. It does this by detecting meaningful shifts in online news. When they occur, the agent decides when to notify your team by judging factors like brand relevance and whether activity is predicted to grow. For NewsWhip customers, this agent delivers a concise, analyst-quality update that helps leaders act at the right time. We agentic AI is only going to become more capable and that teams will use Trellis agents to handle all sorts of tasks.
With this growing number of agents, you are going to need a workspace where you can create, where you can manage, and supervise them. This will be possible in Agent Studio. With Agent Studio, you will have the ability to create custom agents to power up tasks and team workflows. You'll be able to configure agents to run in the background, like the monitoring agent I just mentioned, or have them run on a schedule, like to provide a weekly campaign report. What's more is that you won't need to be an engineer to create Trellis agents. Using our natural language builder, users will configure agents without writing a single line of code. Basically, just describe what you want the agent to do, and the agent will build itself.
We are so excited about where we're headed with AI agents, and we are so eager to get this into your hands. We know that your work doesn't stop at the edge of Sprout. You're likely already using a conversational AI tool. In this moment where everything is speeding up, you need easy access to information. This is why we're focused on connecting your social data into the tools your teams are already using. Today, we are launching our remote MCP server that enables a connection with ChatGPT. In case this concept is new for you, MCP stands for Model Context Protocol, and it's an open standard way for AI tools to communicate with one another, similar to an API. The connector allows teams to analyze and improve their drafted and upcoming content with the help of ChatGPT.
You can also explore the performance of published content on TikTok and feed this back into your content strategy. Let's take a look. Here, we're asking ChatGPT to analyze top-performing TikTok posts and how they compare to the lowest-performing posts over the past 30 days. ChatGPT connects to Sprout through the MCP server, and next, it returns a 30-day performance comparison. It summarizes the format, impressions, comments, reactions, shares. We can see our posts about mobile grooming vans with trending audio really popped off. These sort of high-energy clips did really well, but it looks like our happy Friday weekend vibes content didn't really resonate. As it breaks down these key themes, it makes recommendations for us around narrative and storytelling. ChatGPT even gives us guidance to boost performance. For example, it lists out the specific themes that worked best.
It suggests that we use more humor to improve generic posts, and it calls out where we should possibly add more relevant tags and provide some example CTAs to try. For deeper analysis, we can follow the suggested prompts that ChatGPT generates, or we can ask a new follow-up question. Here, we're asking which times and days of the week we should focus on moving forward. ChatGPT uses Sprout Publishing data to build a table of post times for the past 30 days sorted by performance. It looks like our highest-performing post was on Tuesday evening with 123 impressions. Weeknights performed best with our audience, but we should avoid Mondays and Fridays if we want to reach our customers. You can imagine how powerful these strategic recommendations can be for planning your calendar and campaigns.
It is great to have this information available in ChatGPT alongside other apps that you use to get work done. The ChatGPT connection is available today for all Sprout customers who have a ChatGPT subscription of Pro or above. So far, we have talked about two new additions to Sprout AI: Trellis, Sprout's AI agent, and the ability for Sprout to connect with AI applications in the ecosystem. Both of these new innovations sit on top of Sprout's AI data layer, powered by a growing number of proprietary models, like the ones I was talking about earlier. Our latest addition to this foundational AI layer is a new model trained to detect and auto-complete spam. We have all seen spammy messages full of handles, hashtags, or maybe some sketchy links.
Not only are these messages cluttering your inbox, they're throwing off your metrics because they take time to manually sift through. Now, when spam is identified, it will be categorized and can be automatically marked as complete, helping your team spend less time sifting through noise and making it possible to get to the most important conversations faster. Also, by pairing spam with an automation rule, you can directly hide spammy comments from view on Facebook and Instagram. Spam detection is powered by Sprout AI, and it is live in the Smart Inbox today, so I hope you'll go check it out. Another area where AI can radically improve workflows is within influencer marketing. Brands get better business outcomes when they pair with the right influencer partners.
At last quarter's Breaking Ground, we showed you brand fit and brand safety scores in influencer marketing, which let you determine whether a potential influencer is a good fit for your brand at a glance. With these AI-powered scores, teams like Sunkist have been able to take their vetting process for creators down from hours to minutes. They're not just saving time; they're also more confident in that process overall. Our goal is to provide the best influencer matchmaking service in the industry, and Sprout AI will help us get there. Here's what you can expect next. Soon, Sprout AI will be able to analyze videos and imagery to identify content themes beyond what's possible or available in post captions, revealing more attributes about creators, such as their occupation, their hobbies, and even their tone of voice.
This evolution in creator properties detection will give you even more confidence that the creator you're vetting is a perfect match for your brand or for your campaign. The possibilities here are awesome. Everything I have covered today is aimed to show you how Sprout AI and now Trellis are here to unleash social intelligence for your team and for your organization. For over a decade, we've focused on making Sprout the central command center for social. Today, the rise agentic AI meets our depth of social data, creating the perfect intersection of potential. Sprout's AI engine is here to unlock social intelligence for your entire business at the speed and scale never before possible. It brings every corner of your business closer to your customers and helps you turn that connection into unstoppable business momentum. We are so excited to get this into your hands.
With that, I'm going to hand it over to Paige, who has even more updates to share.
Huge, huge thank you to Meg for that amazing demo of Trellis and a look at what's next for Sprout AI. I'm loving all the excitement in the chat. I know we've been hearing a lot of great things from customers who have gotten their hands early on Trellis. We're all very, very eager to see the impact that it has on your and your team workflows. Hey, everybody. I'm Paige Britton. I'm a Product Marketing Manager here at Sprout. Today, I have the pleasure of walking you through some of the other new and upcoming releases that we have in store.
First up, we have made a ton of updates and care to make sure that you and your teams have everything you need to be as productive as possible and can continue delivering amazing customer experiences. With bulk action enhancements, you can now make Smart Inbox updates significantly faster, taking action on up to 500,000 messages at once. With case macros, you can apply tags, type, priority, status, or assignment without actually having to click into the case so that you can get more done faster. Auto-close lets you close out cases that do not receive a customer response within a set amount of time so that your care metrics are not negatively impacted if a customer goes radio silent on you.
With feature access controls, you can set permissions in Sprout based on job function or other criteria so that your team members only have access to what is relevant to your role, helping reduce any distractions or mitigate brand risk. Finally, goal-time reporting and agent status exports give managers more tools to measure success. We talked a bit about NewsWhip monitoring agents earlier. If you were at our Breaking Ground last quarter, you know that NewsWhip is our new predictive media intelligence solution. NewsWhip shows you where conversations are headed so that you can make the right strategic moves at the right times. Last month, we rolled out NewsWhip Social Alerts, which unify real-time alerts and scheduled digests across Reddit, TikTok, Instagram, Facebook, X, YouTube, and web articles, so the full gamut. This allows you to create your monitoring setup once and apply it everywhere.
You can also set up custom engagement thresholds, apply post-type filters, and bulk create across networks. When something important happens, you'll get notified right away where your team already works, whether that's in Slack, Microsoft Teams, or email. These social alerts work alongside AI agents and AI digests to cover alerting needs across all of your networks. You will get immediate signals, scheduled summaries, and analyst-like insights in one place so that you can catch any reputational risk early and jump on opportunities before anybody else does. In influencer marketing, we are deepening integrations with Snapchat and Instagram, two networks where stories are everything. On Snapchat, we recently launched support for Snapchat Stories and now have verified demographic data for tens of thousands of creators.
On Instagram, you can now get partnership ad recommendations, which suggest which posts will drive the highest ROI when you boost them. We're also integrating with the new Creator Marketplace API so that you'll have even more verified information about potential partners. Altogether, Sprout's influencer marketing platform really helps you scale and drive business impact. We know for regulated industries, it can especially be a challenge to balance the goals of compliance with your social strategy. Guardian by Sprout Social embeds customizable, intuitive guardrails into everyday social workflows, helping brands stay active on social while mitigating compliance risk. This quarter, we are making Guardian even stronger with three critical updates. First, secure forms will be PCI compliant, which will help our healthcare customers stay compliant while providing seamless social customer care to their patients.
Next, we're going to be introducing custom variables with our existing data masking capabilities, meaning that you can proactively identify and mask sensitive information according to your organization's specific rules or any customer categorization that you might have. The third big update, we are going to be rolling out blocked words and employee advocacy. This will ensure that your content stays on message when your team posts, and you can still achieve that massive reach and ROI without risking your brand's reputation. Now, let's talk about some publishing wins, which is always my favorite. We recently launched support for first comments on Facebook and LinkedIn. Yes, you can go ahead and delete all 52 reminders that you have on your calendar to remember to drop a link in the comments. Instead, now you can automatically schedule these straight from Sprout Compose. Big win there.
Building on more highly requested updates, we've got big, big news for brands on Instagram. Mixed media carousels are on the way next month. We know that when you combine videos and images on Instagram, engagement goes up. Not talking about small numbers here. These posts see 24% more engagement and 18% more likes than a single image post. Soon, you'll be able to take advantage of mixing images and video in one post for richer storytelling that really captivates and keeps your audience hooked while driving better results. A big theme at today's event has been around the power of social listening. Listening is a gateway to that universe of social data that we've been discussing. When you have access to more networks and more sources, you can gather more insights and activate more strategic plays.
Plus, now you will have more data to power Trellis. This is why we are glad to share that in a couple of weeks, we're launching Bluesky as a new source for social listening so that you can get comprehensive insights across this network. Up next after that, we will be adding branded hashtags for TikTok so that you can monitor your brand holistically. Soon after that, we'll be adding new listening sources for Instagram and Facebook. This is all right around the corner, so we can't wait to see what inspiring campaigns you create with these new unlocked sources. Now, before we get to our highly anticipated panel discussion, I'd love if everybody could take a moment to answer our poll question. As we just learned, social listening is more powerful than ever now with Trellis and even more network sources.
If you'd like to connect with our team to learn more about Trellis and social listening, please let us know by answering this poll question. We are here and ready to help you adopt AI in your workflow. This is truly such a great opportunity to get ahead of the competition and make sure that social insights are flowing throughout your organization. Okay, amazing. It looks like all the answers are in. If you ask for a demo, we will follow up with you after the event and make sure that we get time on your calendar. Now, on to something special, one of my favorite parts of Breaking Ground, enter the chat. This is our episodic series where marketing professionals and experts come together and share stories about their firsthand experiences. To get us started, I'd like everybody to please welcome Jamia Kenan.
Thanks, Paige.
Hello, everyone. I'm Jamia Kenan, Senior Social Media Specialist here at Sprout and your host today for the Enter the Chat panel. Before we begin, do not forget to send your questions in through the Q&A box because we'll get to as many as we can after the panel. Today, I'm super excited because we have two AI experts who have made some impressive strides within social media. Today, we have Jeff MacDonald and Bianca Shaw. Jeff, can you come on in and introduce yourself?
Absolutely. Nice to see you all, and thanks for having me back here. I'm Jeff MacDonald. I'm Director of Innovation and Technology at the social agency Movers + Shakers. My job is to make sure that we keep our clients and our agency on the forefront of everything technology and AI.
Awesome. Thank you so much for being here.
Bianca, welcome to Enter the Chat. Could you say hi and introduce yourself?
Hi, everyone. I'm excited to be here. I'm Bianca Shaw. I'm the Vice President of Enterprise Social and Digital Reputation for Caesars Entertainment, which includes 50+ destinations across the US, our award-winning loyalty program of Caesars Rewards, and our powerhouse brands of Caesars Sportsbook as well as Caesars Palace Online Casino. It's all about content and care over here and maximizing everything from an enterprise perspective.
Amazing. Thank you so much, Bianca, for being here. Okay, let's dive right in. One of my favorite generative AI eras was during like 2022 to 2023 when people were doing like those kooky AI remixes with cartoon characters like SpongeBob and Scooby-Doo to perform popular songs.
My question to you all to get us started, if we could use AI to have your favorite childhood character sing a song for the soundtrack of your life, who would you pick and what song would you choose? Anyone can jump in.
I'll go first.
Oh, you go ahead, Bianca.
I grew up in the MTV era, so I'm probably dating myself. I'm going to give an example of what's relevant to my almost 10-year-old daughter. She absolutely loves Groot from Guardians of the Galaxy. I think that one of our favorite songs together is Safe and Sound by Capital Cities. How cool would it be for Groot, even though he's been done before, to sing that song just for us?
Oh, I love that. Jeff, what about you?
For me, the kind of show that was just, I feel like in the background of my entire childhood was The Simpsons. You know, I think it would butcher my favorite songs, but I'm just going to go ahead and say probably The Smashing Pumpkins, 1979, for those that remember the episode where Lollapalooza comes through Springfield.
Oh my gosh. Thanks so much for sharing this. I know for me, it would probably be Penny Proud to do that Soraimen-themed song because I just was obsessed with them growing up. Kind of staying on the whole memory lane with AI, the presentation day really has me thinking about how AI has evolved so much over the past 10 years.
I know for me, like one of the first things that I think about when I reminisce on my first experience was when I was at AfroTech in 2022, and a speaker showed me a generative AI video of Will Smith eating spaghetti. It was so alien-like, but since then, AI images have gotten so good. Jeff, what do you remember, and can you describe your first experience with AI?
Yeah, I remember specifically it was November of 2022. My daughter was just born. I was on paternity leave, and I had gotten the ChatGPT 3 beta access for the website UI. People were really talking it up. I decided to start just like writing poems. I do not know, like I was feeling particularly poetic at the moment. Immediately, I saw the capabilities. I was like, this is writing poetry today.
Tomorrow, it's creating decks for us. When I got back from my paternity leave, I immediately did an agency-wide presentation to everyone to show them kind of just like what AI could do today and then kind of thought starters on where it could be in a couple of years and how we should start going ahead and retooling what we were doing. Behold it, this is now like my full-time job.
Oh, I love that. I love how you both kind of have like memories connected to AI in your family. Like that's so sweet. Bianca, what about you? What was your first experience like?
I would not say it's my first experience, but it's definitely one of my fondest experiences. I think it was in 2023, Adobe just started delving into the generative AI capabilities.
Much like many of you, we work on small but mighty teams, so budgets are not necessarily the biggest. My team was experimenting with natural language processing functionalities of the AI generation. We took an empty restaurant photo, plugged it into Adobe, and decided to ask Adobe to add people to it. In 2023, which does not seem so long ago, it produced something hilarious. The faces were okay. The teeth were questionable, but the hands were just something alien-like. It is amazing to see how much it has progressed. Now we have to have disclosures and filters to tell us what is AI and what is not.
Oh, that is so real. Early AI was definitely otherworldly. There were so many struggles with details like fingers and teeth. Like why are there humans with 10,000 teeth?
Earlier in the show, one of the things we talked about was generative AI and how it compares agentic AI. how are each of you thinking about these different types of AI in your work flows? Let's start with Jeff.
Yeah, for me, like I'm really in that transition period from AI as an assistant to AI as a coworker. And how I'm doing that and like agentic AI is simply I just got another monitor. Like there's so many ways you can set a task off with AI and have it completed on its own, a lot of the things that were shown today in the preview.
I kind of like started predicting that I'm going to need a space where I'm going to have that human-in-the-loop moment where I'll see it needs the next step done for it, or I need to see where it's completed. Then I can do like my more human tasks in my main browser, where I can do like the things that are like more creative, and then it can be analyzing spreadsheets over in another tab.
Right. I know I'm always looking for more time. Anything agentic AI can do, I'm always looking to do that. Bianca, what are your thoughts?
As a Sprout customer, generative AI has become a core function and just overall tool for me on a daily basis along with my team.
Just being able to have the workflow efficiencies as a complex organization as large as ours, you can imagine that if we're rolling out a campaign, we want to make sure that we diversify. The AI Assist is clutch when it comes to that. Secondly, what I love about being a Sprout customer is that my customer success team listens. On an annual basis, I do have a connect with our DEI team, and they're asking, what are you doing from a social media perspective for those customers that are looking for accessibility? I brought this to my customer success team and asked them, hey, if it's ever something, would love to see it on the roadmap. One day they came back and said, we have alt text generative descriptions now.
What a game changer, just because this has been something that I've been asking for, and it would be nice to have. Now we're able to utilize the function to generate alt text, not just from an SEO perspective, but once again, from an accessibility reader perspective.
That's so true. That's definitely a focus for us here at Sprout. It's actually one of my favorite features too. Like in my role, I support our social search strategy. As you mentioned, it's definitely a must for an accessible experience. Using alt text also extends your reach because of indexing within social search. When your content uses those keywords or phrases, people are going to discover your content so much more. Now, Bianca, I have to ask, I know you had a chance to explore Trellis, Sprout's agentic AI agent within listening.
Can you share how Trellis has impacted your team so far?
Absolutely. I'm a data geek like many of us. When I was approached by my customer success team and Meg to ask if I wanted to participate in this alpha, I absolutely said, no brainer, count me in. I'm a veteran within the industry, so I've worked with tools for those of you that know Radiant 6 back in the day. Heavy hitter social listening program that required so much manual effort and just overall the ability to set up the queries, setting up the Boolean search, optimizing it based off of the findings, and specifically with the insights, that was a manual process.
When I got my hands on Trellis, holy crap, like the ability to ask it questions, all these different prompts, what I would typically spend hours, days, even weeks making the analysis because I was trying to carve out moments in my schedule to be able to focus. It has provided so much ease and being able to extract a lot of the information that would normally take me so much time and bandwidth. It is a starting point, but it is a strong starting point because of the questions that I can ask it and the ability to drill down on sentiment as well as sources of where the detractors and overall contributors are coming from.
Thank you so much for that. Like you already knew that I am team Trellis, and I am so glad to hear that you have already made some incredible strides with that.
That is exactly what we're talking about when we mention the social-powered AI engine, right? Getting those insights, empowering your team across the entire org, experimenting and seeing how you can use this new technology. We have also talked a lot about how the customer journey lives on social. Let's speak to AI's role at each stage of the customer journey. Jeff, let's start with you.
Yeah, for us, you know one of the things that we're doing at Movers + Shakers is trying to make brands popular on social and get their story out there ahead of everyone. We are really looking at agentic AI can kind of be that always-on social listening expert that is looking and finding those new opportunities for us. It is really around awareness.
Like we're looking for stories that we can build around trends, around awareness, and kind of build tools around that. We think that there's a huge opportunity agentic AI to help us with that because, you know, we can only scroll so much of our FYP to find trends. At some point, we need somebody to step in and help us agentic AI is a perfect way to do that. If it can help us build a creative brief that then keeps us on brand and on trend much sooner than we would any other way, then that's a win as well for us.
Definitely. Bianca, I know you've kind of shared your thoughts on the customer journey. I'd love for you to speak to that as well.
Sure.
As social media marketers, everybody understands that we are part of the entire customer journey. For hospitality and gaming, you can only imagine, to Jeff's point, the awareness is something essential, especially from organic teams. The retention, the consideration part is something that we are deep in as well. AI has a part in all of that, as well as the conversions. Where I really want to hone in on is really the social listening. Just because once again, with Trellis and the ability to leverage AI to work smarter, not harder, we have the ability to listen into conversations and pick up on conversations that we were not even aware of, to be able to grab those insights and understand the granular data that is happening. I will give an example for Reddit specifically.
It's a little bit of a scary place for us brands to play in, but so precise and just a great source of information to really drill down on whether it's crisis-related or if it's just the deep ambassadors of Redditors having conversations about things that you're not even thinking about today.
Yes. When you mentioned Reddit, I definitely was just like, yeah, that's definitely a wild west there. I know for like at Sprout, that's one platform that we're experimenting on, which has me curious about how each of you are exploring AI. What's one AI win or flop that blossomed from curiosity or iteration? What are some takeaways from those learning moments that other teams should consider?
Yeah, for me, the big win that AI has had is us getting away from just like the surface level what AI can do and actually starting to find those like personalized use cases. It's a lot of things that are on the back end. It's a lot of things behind the scenes that AI is making a change on. It's helping us be more efficient. It's helping us get away from spreadsheets and get more in using our human intuition and our creativity. You know, kind of what I always tell my team is if it seems like computers will be better at a task, it probably should be a computer doing that task. You should be doing something that's uniquely human to you, like using your knowledge around insights and about culture and about human feelings. That's the things that we should be doubling down on.
We should be finding ways that we can automate tasks in places where humans were never meant to be. Sorry for people who love spreadsheets.
I'm definitely a team. Let's not use a spreadsheet. Let's pass along that work to the AI. Bianca, what's one of your AI wins or flops?
I would agree with everything that Jeff has said. Just once again, workflow efficiencies. AI is one of my best friends. I want to address the elephant in the room that AI, in speaking with some of my customer advisory board members, it's not going to replace us. It's going to be our team member. It's going to be our tools to help us once again work smarter, but not harder. Really, just overall, what we've discussed is that customers don't want to see a full AI approach content strategy.
I think that people are on the other side of it and just wanting to see the human aspect of being able to see that content that really resonates. There might be, once again, the ability to leverage AI to help us create different things that we do not have the tools or budget for, but it is not replacing us from a care perspective, content, et cetera.
Yes, definitely. It is more so like a sparring partner. And it is not going to replace us. It is something that we should look forward to saving more time, right, as we talked about during this entire conversation. Jeff, Bianca, thank you so much for this conversation. I know I am definitely walking away feeling inspired about what I can test and explore. I am sure the audience is as well.
Talking about feeling inspired, what a perfect segue into my favorite part of Enter the Chat, the Q&A. I know we have a lot of great questions coming in, and I'm sure there's some more brewing. Remember, if you have a question, please add it to the Q&A box. The chat is absolutely on fire, which is amazing, but those questions might get lost in the chat. Make sure you're using the Q&A box. I think all of our speakers are able to stick around. We're going to bring them back to the stage and try to answer as many questions as we can live. Ok, let's get started. This first one is for Jeff. In the past, you've said, when AI does everything, it does nothing. Can you elaborate more on what this means?
Absolutely. Thank you for bringing that one up.
I love that phrase, when AI does everything, it does nothing, because it really describes how we've kind of, and people in my profession specifically, AI implementers, have failed a lot of our organizations. Kind of to that famous MIT study from this year, 94% or 98% of AI implementations fail. It's because we haven't made it personalized enough. We've kind of just given an organization, ChatGPT or Gemini, and said, use it. We haven't really made the case for our employees about how it's going to make their lives better and how we're going to make their jobs more effective and how we'll get them back to the things that they loved about the industry that they're in. In social, that's like the most important. Like we really are looking for ways that we can make AI as effective as possible.
What we've been waiting for are moments like today, where we're looking for ways where AI can be integrated into our existing workflows, into tasks that we need to accomplish, and making it so that it's kind of task specific. When we get away from AI does everything, we can actually start getting away from the it does nothing and finding those moments where it's useful. I think that, you know, things like looking at the MCP integration with ChatGPT, that's going to be a huge win for professionals who are looking for better ways to integrate it into their workspace.
Thank you so much, Jeff. You know, I love that phrase. AI can do everything, but can also do nothing. This next one is for Tazi. How do we get access to Trellis? Do we have to specifically request it?
Great question.
As Meg announced earlier, which we're so excited about, all of our customers who attended today who have social listening will be moved to the front of the line for early access. We are in beta right now. Early access is going to be rolling out over the next month or so. You will get an email from Sprout with your CSM CCed with all of the information once it is activated on your account. There will be a lot of really fun resources to make sure you're getting the most out of Trellis.
Thanks, Tazi. This next one is for Meg. What's the difference between calling a generative AI like ChatGPT that can format and synthesize all these things and a true agentic AI?
Yeah, that's an excellent question.
It really, you know, as we think agentic AI, or maybe first we'll talk about generative AI, the kind that we see in tools like ChatGPT, is designed to really be responsive. It takes a prompt and uses it to write and summarize and synthesize the information that you provide. It's a fantastic tool, but it doesn't really independently take action on its own. In agentic AI understands intent and is really built to take steps on your behalf. This means it can select from a set of tools that it's provided and retrieve the necessary data and coordinate and sequence tasks to help you achieve outcomes, like some of the things we were showing earlier. It really moves beyond just creating content to actually helping you manage and execute workflows.
Thank you so much, Meg, for that breakdown.
This next question is for Meg as well. Is there a minimum amount of data you need to get accurate insights from Trellis?
Yeah, for sure. That's a super practical question about getting started. The great news is that there's no hard minimum requirement for data. Trellis is designed to be super flexible and can generate valuable insights from both small as well as very large data sets. That said, it's really important to know that the quality and depth of insights will really improve as you have more data available over time. The richer the data set, the more granular and actionable the results will be that you get to experience within Trellis. This is exactly why, as Paige mentioned earlier, we are continuously adding more data sources into our listening product.
Our goal is to ensure you have the richest possible data foundation for Trellis to generate insights from, really to help your team. Yeah, hopefully that answers the question. Yes, I definitely think so.
Thanks so much, Meg. This next question is also for Meg. The chat is loving you today. Is Trellis only available if you have the social listening tool?
Yeah, I love all these questions. This is great. I love that you're eager to get started with Trellis. We are starting with a very focused rollout. Trellis will initially be available for our listening customers on the professional plan and above. As we shared earlier, all of you attending today's event will be receiving exclusive early access to Trellis over the next month. That is super exciting.
Beyond that, our plan is to quickly expand Trellis and this access to everyone as Trellis becomes embedded across more of the Sprout platform, both in publishing as well as the Smart Inbox and covering more workflows over the next few months. Critically, we are building Trellis as a core capability. We know how important it is for all of you to stay ahead of your competition. We will be making this available to all customers on the professional plan and above at no additional cost.
Thank you so much, Meg. I appreciate that. Wow, those were so many knowledge gems. I loved all of those insights. It does look like we are close to the hour mark. We have covered some great questions in the chat. Thank you again to our guests, Jeff and Bianca, and to everyone in the audience for such thoughtful questions.
This Enter the Chat has been so insightful. I'm going to pass it back to Tazi, who's bringing us home with final words.
Amazing. Thank you so much, Jamia. And again, everyone, for participating and asking such great questions. Thank you again also to our panelists in Enter the Chat. We are out of time. We'll be wrapping things up for this edition of Breaking Ground. Thank you so much for joining us. Have a wonderful rest of your day. Bye, everyone.