We believe the AI agents can really unlock the upside which is li really unlimited in terms of revenue generating and enable your workforce to do more than they currently doing. Another thing is I will also go back to the deep research AI analyst agent is that you can really now for the first time can really um unlock the full potential of the try to retro data that's customer uh interactions then the level of insights that traditionally you cannot get to and you can now use that to improve operations that ...

At a Greylock Change Agents event, Glean CEO Arvind Jain discusses how enterprises need searchable agent libraries and orchestrator systems to manage agent discovery and deployment, along with careful curation to keep agents organized and effective. Clip from The Enterprise Brain for AI Agents with Glean and Cresta: https://youtu.be/1ZMDA6atOhc #greylock #Glean #artificialintelligence #AI #Agents #AgenticAI #software #AgentLibrary #Agentorchestration ...

Bringing the right context to the models is actually a very hard problem. But luckily for us, like you know, that's the thing that we we've been working on for the last six to seven years. Like our whole like the the model for Glean has always been go deep, you know, connect with all the enterprise systems and now start to build an understanding of how this business actually works.Yeah. >> Understand like you know the data, the knowledge, understand business processes work and like you know given any task l ...

As you uh develop more and more agents, as you take u these human-driven processes and agentify them, you have to think about like how do you bring that full comprehensive enterprise context to all all of these different agents. And ideally, we feel like AI should be like electricity. It just disappear.We don't, you know, be using AI without even knowing using AI. And it's almost like augumented reality before work. Arvin Ping, thank you so much for joining us for Greylock Change Agents.As you know, change ...

One thing that you've said time and time again is, you know, we've been building software for humans >> and uh and now we're really building software. You can imagine a world where we're building software for agents. >> I'm very uh convinced that users are still going to want to see a lot.>> Okay. And and so that as long as the user wants to see a lot and touch buttons and move things around, um then I think you have a a very strong case for for sort of dedicated, let's just call it SAS, you know, plays per ...

If you went to Coca-Cola and Pepsi and you gave one the world's worst CRM system and one the world's best CRM system but the same talent profile at Coke and Pepsi, I don't know that the the revenue would be com so like in a massively compounding way that different, right. If you gave them the world's best ERP system and the third best ERP system, you know, I don't think that like the supply chain is going to be so totally different. But if you give a team right now the state-of-the-art in coding, >> yeah, > ...

Over time, as you think about the box business model or even abstracted away into this any system of record business model, where does the value in that business model acrew to. Over time, >> systems of record were built for a TAM of human employees. >> Yeah.>> And that was sort of where the value kind of maxed out at. And now I at least imagine I think anybody, you know, doing AI agents imagines that maybe we'll have like a hundred times more agents than people or a thousand times more agents than people. ...

AI & Agent Adoption in Enterprises - Box initially focused on cash flow positivity but now prioritizes delivering functional AI agents, believing margins are currently less important [1][74] - The company is adopting an "AI-first" strategy, aiming to resemble a new startup in its approach to AI implementation [8] - Box is excited about the potential of AI agents to tap into unstructured data, which constitutes approximately 90% of corporate information, including contracts, marketing assets, and research files [14][15] - The company envisions a future where agents can answer customer questions and automate workflows related to unstructured data [15][16] Agent Differentiation & User Experience - Differentiating agents involves two main categories: seamless integration where users are unaware of the agent's presence and scenarios where users directly interact with agents to delegate tasks [19][20] - A key challenge is designing a delightful and transparent user experience for human-agent interaction, balancing transparency with simplicity [22][23][24] - The company emphasizes the importance of data advantage, noting that a large percentage (potentially over 95%) of enterprise data is underutilized [30] - The company believes applied solutions are crucial because understanding the nuances of each use case is essential for user comprehension and effective implementation [39] Business Model & Enterprise Transformation - The company is moving towards a business model that incorporates both user seats and consumption-based pricing for agents, utilizing "AI units" to normalize different use cases [129] - The company anticipates a shift towards selling outcomes to customers, requiring a deeper understanding of their businesses and potentially leading to new service level agreements (SLAs) focused on accuracy [130][131][132] - The company expects the transformation of enterprises with AI to be a gradual process, allowing for retraining and repurposing of the workforce [85][86] - The company predicts a significant increase in the use of tokens (AI processing units) for tasks not currently performed, indicating new opportunities and workflows [87][88]

What does an agentic uh workflow look like for legal and for Harvey. >> I would think of there's kind of two ends of this spectrum that we think about when we think of building this in the product. One is how do you do the very open-ended agentic tasks.So if you give Harvey any question you would give an associate, can we have some system that kind of looks like deep research that can take that task, create some plan, figure out which tools to use and solve this. And then the other end of the extreme is a l ...

Is data a big moat for RV. >> Yeah, I think there's two types of data. And so when we started the company, I think everyone thought of legal data as case law. >> And we've now partnered with Lexus and I think having access to case law and these other data sets is super valuable.But I don't think that is when we think of like the data mode that we think we're going to have as a company. The data that I think is really valuable that I think now people are realizing is valuable is the RLHF data or the reasonin ...