108: OneStream Splash 2026 Recap - Part 2

Welcome to the CPM Customer Success Podcast, where we help office of finance leaders enable an exact solution to enhance their corporate performance. Without spending countless nights and weekends in the office. Hi, I'm your host, Andy Smetana from Nova Advisory. Nova Advisory is a 100% OneStream professional services diamond implementation partner. We help OneStream customers and prospective OneStream customers get the most out of their OneStream investments. For CPM customer Success listeners, we now have a weekly email newsletter to keep you informed on the latest CPM customer success news by subscribing today. We'll keep you informed on episode releases, including customer success stories. OneStream interviews and capabilities to keep you up to date about one Stream's unified platform for finance. You can check that out now at Nova Advisory's website, Nova advisory.com/podcast and subscribe to the CPM Customer Success podcast newsletter today. Today we dive into part two of our Splash 2026 recap. In part one, we focus on the platform foundation behind OneStream's Splash 2026 announcements. We cover the new high-performance cube engine, OneStream's developer studio, expanded sensible AI capabilities, and Sensible AI Forecast Express. The theme was speed, scale, extensibility, and practical AI forecasting. In this episode, we're moving into the next layer of the conversation: governed AI, native agents, cluster analysis, and what all these announcements mean for finance leaders building out their OneStream roadmap. Splash 2026 made it clear that OneStream is not only investing in faster performance and better forecasting, they're also building out a future where finance users can interact with data, documents, analysis, and workflows in a more conversational and intelligent way. The key point for finance leaders is governance. AI in finance has to be trusted. It has to respect security, understand business definitions. It has to be grounded in the financial model, and it has to support decision-making without creating a new risk. This is the lens for today's episode, so let's get into it. part one: the finance agentic layer, governed AI for finance. So one of the more forward-looking announcements at Splash 2026 was the finance agentic layer. This is OneStream's framework for governed AI agents in finance. The overview describes the finance agentic layer as having three major pillars: an agentic gateway to route and govern AI agent requests with controls around security, identity, and cost. Finance specific tools, including a semantic layer to keep AI grounded in the customer's OneStream data model. And third, packaged skills and plugins for standardized finance tasks. The word that's key here is governed. Finance leaders are interested in AI, but they also have legitimate concerns around control, accuracy, security, auditability, and data access. The agentic layer is designed to create that controlled framework. OneStream also described a semantic layer that allows agents to use role-based business language and provide more accurate responses. It points to openness, allowing external authenticated AI solutions such as Copilot, ChatGPT, or Claude tap into a governed OneStream model for analysis and answers. And that's an important direction. The future of AI and finance needs to be built around trust and control, and that is exactly what OneStream emphasized at Splash. Splash twenty twenty-six also covered native AI agents embedded directly in OneStream. These agents are designed to help finance teams retrieve data, analyze results, and execute tasks using natural language while operating inside OneStream's data model and security framework. OneStream announced four pre-built agents, the finance analyst, the forecast agent, the search agent, and the deep analysis agent. The finance analyst agent is designed for natural language analysis of financial data. Users can ask for analysis in plain English and receive real-time results, things like dashboards and visualizations from OneStream's financially intelligent engine. Use cases include variance analysis, P&L deep dives, and performance analysis during planning, forecasting, or reporting cycles. This can reduce friction for FP&A teams and finance managers who need quick insight without building a new report every time. The forecast agent allows finance teams to interact with forecasts using natural language, asking questions, exploring scenarios, and adjusting assumptions directly within OneStream's governed forecasting model. In FP&A , That interaction is critical. Forecasts are shaped by assumptions, drivers, scenarios, and business judgment. The forecast agent helps teams explore those elements more quickly with greater transparency. A conversational forecast experience can help finance teams test scenarios faster, understand forecast drivers more clearly, and spend more time on decision-making instead of manual analysis. The search agent is focused on conversational search across documents and knowledge sources. It can help users search unstructured data such as platform guides, policy documents, accounting procedures, historical reports, and knowledge bases. It returns answers with source references, which make it useful for controllers, auditors, finance staff, and administrators who need to retrieve supporting information quickly. The deep analysis agent is Designed for More Complex cross-data analysis. It's capable of processing large volumes of unstructured content to answer more complex questions across contracts, financial statements, and other sources. Potential use cases include strategy support, risk assessment, due diligence, scenario analysis, and executive decision support. Taken together, these agents point towards a more conversational user experience inside the OneStream platform. Finance users will still need strong processes, clean data, and thoughtful analysis, but these agents are designed to reduce manual effort, speed up access to information, and help users move more quickly from data to insights. Another sensible AI capability highlighted at Splash was cluster analysis. Cluster analysis automatically groups business entities based on operational similarities rather than only relying on static attributes, things like region, business unit, or predefined categories. Allowing finance teams to compare stores, customers, products, or entities that behave similarly. it runs natively in OneStream using the same data model and governance and does not require data science expertise. This can help finance teams improve analysis. Traditional hierarchies are useful, but they can also hide patterns. Two stores may sit in different regions, but behave similarly based on sales mix, customer traffic, labor profile, or seasonality. Two plants may report to different divisions but have similar margin characteristics. Two customers may sit in different channels but show similar purchasing behavior. Cluster analysis helps finance teams identify more meaningful peer groups. They can improve variance analysis, anomaly detection, benchmarking, and performance management. A manufacturing company can compare plants with similar production profiles. A customer profitability model could group customers based on behaviors rather than static customer type. This is a practical use of AI because it helps finance sharpen the comparison set. Better comparisons lead to better analysis, and better analysis leads to better decisions Let's back up for a minute. Because one of the most important messages OneStream wanted finance leaders to hear is that to prepare for AI, teams should start by strengthening the basics. Things like data quality, metadata, workflow, security, integrations, dimensionality, and ownership. AI readiness starts with CPM maturity. AI can only be as useful when the underlying model is clean, trusted, and well-governed. If metadata is inconsistent, AI can amplify confusion. If hierarchies are unclear, AI might struggle to interpret the results. If data quality is weak, AI can generate responses quickly without improving decision quality. If business definitions are inconsistent, AI-generated answers can create more debate and more clarity. And this is where the role of an implementation partner becomes so important. A strong partner can help finance teams identify the right use cases, assess their readiness, prepare data foundations, design governance, and build adoption plans that match the organization's maturity. Customers need more than awareness of the product roadmap. They need a practical path for turning those capabilities into business value. And that's what we'll be covering next week in the third and final episode of our Splash series. We'll talk about where to start, how to prioritize, and how to think about applying both the core platform capabilities and the more advanced AI-enabled features in a way that supports measurable progress. So we'll see you then. Thanks for joining us for this episode of CPM Customer Success Reminder to check out the CPM Customer Success Weekly email newsletter to stay informed of the latest CPM customer success episodes and OneStream News. You can check that out now at the Nova Advisory website, Nova advisory.com/podcast and subscribe to the CPM Customer Success podcast email newsletter today. Thanks again for listening, and I'll catch you next week for another episode of CPM, customer Success.