Agentic Customer Intelligence with Gainsight

Julia Nimchinski:
And next, Brady Bloom. Welcome back! Long time I’ll see. How have you been?

Brady Bluhm:
Nice to see you. I’ve been great. Busy. Busy, busy.

Julia Nimchinski:
What’s the latest and greatest with Gainsight? What’s in your agenda class?

Brady Bluhm:
My personal one? Man, I am playing with Claude all the time. I’ve been a Claude fanboy for almost 2 years, we’re gonna get to 2 years pretty soon here, so, all the way through before projects, and building in projects and playing there. But, I’ve been in Claude Code for the last 6 months, and now, with the new scheduling feature that they have, I feel like Alright, I’ve been wanting to build my own open claw, but now I feel like I’m able to kind of do the same thing with Claude Code locally, so yeah, there’s just all sorts of things you can do right now. It’s really exciting. You just need to think about what you want to do, and then you can do it most of the time.

Julia Nimchinski:
It’s exciting and it’s crazy. It’s the only consistent answer we’re receiving today across all sessions. So, Anthropic, it’s dominating.

Brady Bluhm:
Yeah, yeah, and hey, they’re a Skilljar customer, which Skilljar is a Gainsight customer, so if you’ve taken their courses, their courses are great, by the way, and they’re all through Skilljar as well, so, just a little plug there. I’m the Staircase product manager, though, so I’m gonna talk a little bit about there. Do you have any… anywhere you’d like to start, or a question for me to kick off, or want me to take?

Julia Nimchinski:
Let’s just kick it off.

Brady Bluhm:
Okay, great. For those that have not seen Staircase AI, Staircase is a product that Gainsight purchased a little over a year and a half ago, and I’ve been the senior product manager of the platform throughout that time, and have seen how it’s growing. Ultimately, what Staircase does is it looks at all your unstructured conversational data with customers. And so you’re looking at meetings and recordings, you’re looking at emails with customers, support tickets, chats, those are our, like, core data sources, along with CRM datasets that can be… that can be brought in in usage data. And we use all of that to take all the unstructured data and make it structured. And at Gainsight, we are really the creators of the customer success space, ultimately, and so we bring our expertise. And we really have an opinion with Staircase about the data models and the way that we shape the data and the workflows that people use for this system. So, I’m gonna share my screen. I’m going straight into the product, no slides today, because that’s my preferred way to go anyway. The, like, core essence of Staircase, again, we’re scanning all that conversational data, everything is coming through this system. One of the big benefits right out the gates is that you just get, like, full context across all of your accounts. A couple years ago, it was really hard to get, like, up to speed on an account. You’d have to ping someone, ask them to create, like, an account brief, and drop that into Gainsight Timeline, and then… and then you can share that across with different stakeholders. etc. Now, you can get all of it right here, just with a query, or by landing on a page. Some of the, like, core signals that we’re looking for are churn risks, negative messages, whether an account’s dark and they’re communicating with you. We also do positive messages. Have there been renewal discussions? So, it’s really all the post-sales conversational context. To understand the health of a customer, and where you need to intervene or lean in. And then today, I’m going to focus on some of our newer features that we have, which are our risk analysts. So essentially, by purchasing Staircase, you’re hiring us to build agents that help to take all of the context from conversations, and give you exactly what you need to do, and keep it fresh and up-to-date, both here, but then also to push it into all of your other systems agentically, with MCP connections and APIs. So, on this page, you can see this customer, you can see all of their lifecycle events, any events that it triggered over the last year, kind of get up to speed. You can see their engagement and your response time with the customer. There’s a lot of rich data, along with, like, stakeholder connections. So you get a whole stakeholder heat map of everyone that’s engaged and their health of the engagement around that. And now, 6 months ago, we released our first true agent inside of Staircase, and this is, it has multiple sub-agents at work. Whenever we detect a conversation about an opportunity that’s happening within an account. We fire off a flow that… it has multiple agents that come in and analyze the recent conversations with this customer, and build out all the details of each of the different opportunities that are happening, and… It’s incredibly rich. You’ll see each opportunity that’s at play, so this one’s a seat growth opportunity. Let me actually just zoom in so it doesn’t kill your eyes. This seat growth opportunity has a mini summary here. It gives you the timeline, what’s the expectation on it, the decision maker, if there’s any competitors mentioned. We create an action plan, and this action plan is built off of an experience library that our agents build from your team’s actions. For similar opportunities, what did your teammates do that drove it to success? That’s what this action plan is based off of. So it gives you the next best action based off of what works in your organization and personalizes it in that way. And then also, from the data model we have, what’s missing? What should you ask the customer in the next call that you have with them? So that we can gather that and curate that information for you as well. Then it’s all grounded and rooted in the communications that it’s referencing. So, this one is 3 different emails that happened, but it could be meetings, etc, that it would be pulling from for this rich data set that builds out all the current opportunities across all accounts. Then you can query that also. So, you can ask Staircase about these opportunities. At the cross-account level, you can ask about all the opportunities at play, and which ones have the highest readiness in my strategic tier. And then you can get a brief written for you and an output to show you all of that research. And that’s either in Staircase or with our MCP, you can do it from Cloud, or Cloud Code, or ChatGPT, etc, wherever you’re working. Two months ago, we took this same model, and we applied it to risk, and so let me jump to a different account. Where I believe I have a risk analyst beta. So, this is live for all of our customers as well, and so whenever we detect a churn risk from conversations, from either extremely negative or situations where it’s a little bit dicey, we automatically fire, again, multiple sub-agents in an agentic flow to do deep analysis, tell you which products are at risk, what are the root causes of that risk, so it goes back over the last six 60 days of every engagement that’s happened, and scans for any potential reasons that are leading into this risk and causing it. So you can see it summarizes those, gives you the severity, and any upcoming timeline pressures, again, that are happening here. And now, we’ve added also, on this side area, you’ll see two panes. One, it analyzes all the stakeholders involved. and tells you whether they’re neutral, if they’re a detractor in this risk, or if they’re an advocate in the risk as well. And it tells you, based on how they communicate, what’s going to be your best approach that you should take with this stakeholder to get them on track and help get their buy-in. And then we also create a playbook. So similar to that action plans, we’ve enriched it even further to say, hey, this should be the owner of that role. when should it happen, so we’re looking out over the next period of time, which actions should be taken, and now we’re iterating, and this action feed is going to be feeding into Gainsight CTAs, which can also be synced to Salesforce objects as well, and assign tasks out, and get your teams aligned in the work that they’re doing in that sense. So this is our most recent release on an analyst, and I do have a prototype that I can show of a data model that we have running… we dog food all of our analyst agents, so we have this running internally right now, and I’m going to show you… I was actually at… an off-site that we did internally. I’ve been in customer success for 13, 14 years now, and I’ve been a part of so many lifecycle builds in, like, Miros, or whiteboarding, or, like, alright, what are all the stops along the customer lifecycle? And we were doing that here at Gainsight across all of our products, and it just kept coming up that, like, there’s this, like, handoff period from here to hear that is such a challenge, and we need to capture it, and we always end up having to ask customers multiple things, and so that inspired me to talk with my engineers, and now we have… the first iteration on it is the sales to post-sales handoff. And so whenever a new sale happens, and we get that initial sale date coming through into Staircase, we run analysis across all of the pre-sales conversations to look for their goals, their pain points and drivers, any commitments that were made, and who made those. So, for your salespeople that might overcommit, this allows you to do coaching or other things like that. If they’re making promises that might not. I’ve experienced receiving some of those, and I’ve experienced being the sales guy making some of those promises, too, in the past. What are any concerns or open risks that they have? Were they resolved? Are they open right now? What’s their buying rationale and the decision maker and that head honcho that you need to be aware of? And then their onboarding readiness, too. So, where do we have gaps in the data right now? Do they have any blockers internally? What’s adoption gonna look like? Are there any integrations that we need to be clear on and understand? And then here’s some initial questions, again, that your team can start with to enrich the data model. model. And, again, actions and stakeholders on this, too. So, this is our next iteration on an analyst model, but ultimately, with Staircase, you’re… you’re getting an intelligence engine of all customer understanding. And one of my core goals, Julia, is for CS people, post-sales and salespeople, to not have to manually enter data ever again. I think People can improve the way that they ask questions in meetings, and gather the right context from customers, and then allow the AI models to be the connective tissue, to do that logging and tracking and everything we’ve seen here in these data models, and also the meeting prep. All of the, hey, what don’t we understand? What do we need to ask? Where should we dig deeper to gather more context? And drive more action forward to ultimately drive True customer success. Yeah, I was gonna model a little bit of the MCP, too, like. This is just an output. The richness of the context in Staircase blows me away every time that I use it. I already ran this. Luckily, I was gonna do it live, but for time, I won’t. But ultimately, this ran 4 queries across Staircase, and you can see it’s asking, like, about the stakeholders, and about their health, and about, do they have any open items or concerns? It’s gathering all that context, and then Claude takes takes that and puts it into an incredibly rich meeting prep doc that is… I’m using this every day, and when I’m brought into calls, and every time, I just feel over-prepared beyond what I was ever able to do before.

Julia Nimchinski:
This is amazing, Freddie. We have a couple of questions from the audience. Let me just address those. How do you translate signals into actions without creating false positives?

Brady Bluhm:
It’s a good question. I’m always really amazed at the high quality of data that exists in conversational data sets. But one of the things that we are looking at as we align the staircase playbooks that are AI-generated with Gainsight CS is the actual merging of Gainsight has a very rich playbooks that are, like, you have to build them beforehand as, like, these are kind of rules-based playbooks. I think rules are shifting with AI right now, because we can get such rich intelligence and personalization, so we’re finding a middle ground of allowing you to set structures of, hey, these are some standard plays that we have, but then let AI fill in the gaps and give the Details and the personalization along that.

Julia Nimchinski:
Makes sense, and one more question here, quite futuristic, but what does a fully, agent-driven CS work do differently from today’s model?

Brady Bluhm:
I… one of my real predictions on customer success, and especially the CSM role themselves, one, I think CSMs are going to be able to handle so many more accounts without the chaotic craziness of handling so many accounts, and so I think scaling of CSMs is going to be… we’re already seeing that. It’s by force of the market, but it’s going to get easier for them to do it well. So that’s one thing, but again, like, I think the way CSMs will work is they’ll be able to be on back-to-back-to-back calls through their day. Like, they are the human interface. And so my recommendation always to CS teams is to start practicing how you’re asking questions. If you detect a product feature request, you as the CSM should be the person that digs in and asks those follow-up questions of, hey, tell me more about that. How would that benefit you? What might that look like? Do that on your call so you don’t have to bring a PM on the call. That call then is surfaced to the PM because you already did the work on it, and you asked those follow-up questions. So I think CS is going to shift in that way as that human interface of a business with its customers.

Julia Nimchinski:
Makes sense. And last question, Eddie. I know I asked months ago, but still, what’s in your roadmap? What are you allowed to share, and what excites you the most?

Brady Bluhm:
Yeah, what excites me the most right now is the potential with MCP, both, like, for individual use, and being able to build out your own skills and get creative and just ping us for the context you need when you need it, right? And so, by us shaping that context right, it feeds the AI so much more effectively, and so the outputs are… like, context is king when it comes to AI. And so, we’re really focused on curating that. We do have more analyst agents planned. We’re getting close to, like, where I think we’ll be with those, but we want to iterate on the ones we have and make them more recursively learning, and improving our experience library as well, so that it’s always learning and adapting to your organization when it’s plugged in, and then pushing those also into systems. So, we are, building our API connectors, but then also just enriching our MCP with more tool calls and tool options, and allowing our MCP to be used in system-to-system MCP calls, I think will allow our customer intelligence to be piped into everything your AI agents are doing.

Julia Nimchinski:
Thank you so much, Betty. Huge fans on Gainside, and yeah, where should our community go? What’s the best next step?

Brady Bluhm:
Gainsight.com is a great space to stay up to date on that. We do have our Pulse conference coming up in May, where we’ll be launching some new features. I guess one other thing I could say that is near-term roadmap is product-level differentiation of communication signals coming in. From my research, nobody is doing that right now, because it’s honestly a really hard problem to solve, and we’re We’re hard at work on it right now. And so for larger enterprise organizations, that product-level understanding is there. And then you can find me on LinkedIn, Brady Bloom, B-L-U-H-M.

Julia Nimchinski:
Awesome. Thank you again.

Brady Bluhm:
Thank you, Julia.