The AI conversation in investment management technology has moved fast in 2026. What was speculative 18 months ago, AI that operates within a CRM rather than alongside it, is now in production. What was promised as a roadmap item a year ago, agentic workflows that act rather than just summarize, is arriving in platforms that have built the underlying architecture to support it. And the technology protocol that makes most of this possible, the Model Context Protocol, has matured from a research standard into what one technical analysis described as an infrastructure layer comparable to what REST was for web APIs.
For investment management firms evaluating what AI will actually mean for their IR operations, the practical question is not whether AI will transform CRM but which specific changes will arrive in the next 12 months and what firms need to have in place to benefit from them.
What MCP Actually Is and Why It Matters for CRM
The Model Context Protocol, or MCP, is an open standard originally developed by Anthropic and donated to the Linux Foundation in 2026, that defines how AI systems connect to external data sources and tools. Before MCP, connecting an AI model to a CRM required a custom integration for each combination of AI system and CRM platform. Every firm that wanted AI to operate on their CRM data had to build and maintain that connection independently.
MCP changes this by providing a standardized protocol layer. An AI system that speaks MCP can connect to any data source that also implements MCP without requiring a custom integration for each pairing. The AI receives live, structured data from the connected system rather than a snapshot or a static export, which means the AI’s responses reflect the current state of the data rather than last week’s version.
For investment management CRM, this matters because the most valuable AI applications in IR require live access to relationship data, not historical snapshots. A meeting brief generated from a CRM export that is two days old may reference a contact who has since changed roles or miss an activity log that was just entered this morning. A meeting brief generated through an MCP connection to live CRM data reflects the relationship record as it actually stands at the moment of generation.
MCP enables dynamic context exchange across CRM, ERP, analytics, and workflow systems so responses stay grounded in real-time data. For IR teams whose most time-sensitive AI use cases, meeting preparation, relationship health monitoring, and pipeline status review, depend on current rather than historical data, the architectural difference between MCP-connected AI and AI operating on exports is significant. CData Software
Where Satuit Stands in This Transition
Satuit’s MCP/AI layer and the Satuit Agent introduced in v18.4 represent the firm’s current production implementation of this architecture. The MCP layer provides the protocol infrastructure through which AI models interact with SatuitCRM’s investor relationship data within the platform’s security and access control framework. The Satuit Agent is the first user-facing implementation built on top of that infrastructure.
Further down the horizon, Satuit plans to integrate self-service reporting, investor onboarding automation, multilingual support, and next-best-action intelligence powered by AI. Each of these roadmap items depends on the same MCP infrastructure that powers Satuit Agent today. The roadmap is not a collection of independent features. It is a series of capabilities that build incrementally on the data access architecture that MCP provides. Satuit Technologies
The current Satuit Agent capability set covers meeting preparation briefs assembled from live investor records, activity log drafting assistance, relationship health summaries, and follow-up task generation. These are the high-value near-term applications of embedded AI for IR teams. The roadmap moves toward more autonomous operation: AI that initiates actions rather than responding to prompts, AI that monitors relationship data continuously and surfaces alerts without waiting to be asked.
What the Next 12 Months Will Deliver
Based on where the MCP protocol standard and the investment management AI market are heading, the next 12 months will bring several specific changes that investment management firms should anticipate.
Agentic monitoring becomes standard. The shift from AI copilot to AI agent, software that acts within defined parameters without waiting for a prompt, is arriving in CRM platforms in 2026. KPMG reports that, as of early 2026, 24% of asset management and private equity firms already had AI agents deployed in production, and 68% were actively piloting such solutions. In CRM terms, this means agents that monitor relationship engagement patterns continuously and flag at-risk investor relationships before a human relationship manager would notice the pattern, agents that detect pipeline staleness and prompt appropriate follow-up, and agents that surface compliance documentation gaps as they develop rather than when a quarterly audit catches them. FinTech Global
Meeting preparation becomes automatic, not prompted. The current Satuit Agent requires a relationship manager to request a meeting brief. The next generation, enabled by agentic infrastructure and calendar integration, generates the brief automatically before each scheduled investor meeting and delivers it to the relationship manager without a request. The brief is current because it draws from live CRM data through the MCP layer at the moment of generation.
Investor onboarding automation reduces manual process. The investor onboarding process after a fund close involves a large number of structured, repeatable steps: portal access activation, document permission configuration, capital account record setup, compliance documentation review, and communication preference recording. Automating this workflow through AI agents operating on the CRM’s live data eliminates the manual coordination overhead that currently makes onboarding the most resource-intensive post-close activity.
Natural language CRM queries replace structured filters. Relationship managers who currently navigate filter menus to find specific investor segments or pull pipeline reports will be able to ask the CRM in plain language: which investors have not been contacted in 45 days, which pipeline records are in the due diligence stage for more than 30 days, which investors have side letter co-investment rights that have not been exercised. The MCP layer that connects the AI to live CRM data makes these natural language queries possible in a way that static exports cannot support.
Self-service reporting reduces operational dependency. Leadership and relationship managers who currently depend on a CRM administrator or operations team member to produce specific reports will be able to generate those reports themselves through AI-assisted query interfaces. The self-service reporting capability on Satuit’s roadmap is the practical implementation of this direction.
What Firms Need to Have in Place to Benefit
The AI capabilities arriving in the next 12 months will deliver meaningful operational improvement for firms that have two things in place. Firms that lack either will get less from the technology regardless of how capable the platform becomes.
The first is complete, consistent CRM data. AI is only as good as the data it touches. If your data is messy, AI will scale the mess. A meeting brief generated from an incomplete investor record is an incomplete meeting brief. An agent that monitors relationship engagement and surfaces at-risk investor relationships based on activity log data can only surface what has been logged. The data quality program that makes AI useful is the same program that makes the CRM useful without AI. The AI amplifies what is there. CData Software
The second is the MCP-ready platform architecture. Not all investment management CRM platforms have implemented MCP. Platforms that are not MCP-enabled cannot offer the live data connectivity that makes embedded AI operationally meaningful. Firms evaluating CRM platforms in 2026 should confirm that the platform’s AI capabilities operate on live CRM data through a structured protocol layer rather than on periodic exports, because the distinction determines whether the AI is genuinely useful for time-sensitive IR workflows.
Schedule a demo with Satuit to see how the MCP/AI layer and Satuit Agent are working in practice today and what the roadmap delivers in the next 12 months.
