How Armira runs agent-led origination to spend more time with founders

Results at a glance

• Armira tracks a target universe of roughly 20,000 companies, built systematically over several years and visible to the whole firm in one place, replacing personal spreadsheets, disconnected notes and fragmented inboxes.
• Around 50 people across the firm work in the system, extending well beyond the investment team to include Capital Formation and Executive Assistants managing investor relationships, event coordination and scheduling.
• Manual CRM logging is virtually eliminated. Activity is captured automatically, while structured meeting notes are pushed to records by AI agents, removing the manual data entry that causes many CRM implementations to fail.
• Meeting preparation has shifted from hours of research to a simple review. Teams open a briefing with a relationship history, notes, and pipeline context already assembled rather than piecing it together from emails and calendars.
• List hygiene now takes seconds instead of hours. Armira deduplicates every new target list against the CRM before outreach.
• Armira is now among the most active users globally of Affinity's MCP server worldwide.

Where it all began

There was no single dramatic trigger. The challenge was visibility.

Before Affinity, Armira's institutional knowledge lived in inboxes, personal files, and people's heads. For a firm increasingly running proactive, structured outbound sourcing, that fragmentation was both inefficient and a competitive disadvantage.

The firm needed more than just another CRM system. It needed a system that captured activity automatically and provided a single place where anyone could see who knew a company, when Armira last spoke to them, and what had been discussed.

Why Armira chose Affinity

Armira Growth launched in 2021 with no legacy system and chose Affinity from day one. Of the platforms the team evaluated, Affinity was the one with a working, live Outlook integration, and that proved to be the deciding factor. Once Affinity demonstrated its value within the Growth team, the confidence in the platform grew quickly and the wider firm gradually standardized on it.

The team evaluated the usual mix, from private-equity-specific systems to modern CRMs, against four criteria:

• proximity to private-markets workflows
• flexibility without relying on external consultants
• low maintenance and overhead
• and, above all, the ability to capture relationship data automatically from email and calendar activity.

Reliable data is what drives adoption

The Outlook integration may have opened the door, but what made Affinity the long-term platform was the quality and completeness of the underlying relationship data. The one genuine point of internal skepticism was email and calendar synchronization. Some senior partners were initially hesitant. 

Rather than forcing adoption, the team focused on transparency. Clear rules were established from the outset around what information would sync and who could access it. The turning point came during pipeline meetings. The partners saw their own relationship history accurately assembled without having logged a single email or calendar event themselves. Skepticism quickly turned into confidence. Adoption spread not because it was mandated from the top-down, but because the system consistently produced better information than anyone could assemble manually.

"Affinity removes the classic CRM failure mode, where the system slowly dies because nobody logs anything." — Niklas Krusche

How Armira built its relationship data layer

Because Affinity captures every email and calendar interaction across the firm automatically, Armira’s relationship graph continually builds itself. There is no manual logging, and no dependence on individual discipline. That automatic capture became the foundation for a single pipeline the entire firm could trust.

Over the past year, Armira brought its sourcing into one shared pipeline. Every target has an owner, a status, interaction history captured automatically, and structured meeting notes pushed to the record. Armira Growth built and expanded the outreach effort first, before the Mittelstand migrated onto the same platform. Capital Formation now manages investor relationships in exactly the same system.

The practical test comes during pipeline meetings. Someone mentions a company, and the answer is already there. Who has spoken to them? When? What was discussed? What previously required asking around the office and searching through inboxes is now available with a single lookup, and increasingly, a single question to Claude.

Armira continuously enriches and protects its pipeline through two automated checks.

Warm-intro mapping

Before the firm goes in cold, it checks whether anyone across Armira or its wider network of advisors and partners already has a path into the target. Warm-intro mapping is the practice of surfacing the shortest relationship route to a company through people the firm already knows. When a relationship exists, Armira replaces cold outreach with a warm introduction through the right colleague, allowing conversations to start with existing context and trust rather than from zero.

Collision avoidance

Armira cross-checks every new target list against the CRM before outreach goes out. This check runs weekly, sometimes daily, catching companies where a colleague already has a live conversation running. Today those collisions are prevented before the first email is sent, protecting the firm’s reputation and the founder experience.

Relationship scoring changed how the team prioritizes.

"The team no longer starts from gut feel. It starts from a ranked list." — Niklas Krusche

Rather than asking “Which company should we call next?”, the team begins with a continuously prioritized view of the market. Affinity engagement data is combined with public and licensed datasets, ownership signals, and news to surface the companies where time is most likely to generate a meaningful relationship.

The edge that gives Armira is often invisible from the outside. "The most valuable non-public layer is our own history," Niklas notes: knowing what a colleague discussed with a founder two years earlier fundamentally changes how the next conversation begins. External signals matter too, but it is the firm's own record that most often decides when and how to go in.

How Armira rebuilt origination around agents

This is where Affinity stops being a CRM and becomes one of the foundations of Armira’s AI operating model.

"Being at the frontier of technology and AI ourselves is what lets us be the best possible partner to the companies we back, and it’s what lets us scale further. But the real reason we do this is time: the more of it our agents take off our plate, the more we have for the personal relationships that actually drive this business." — Niklas Krusche

Armira’s operating thesis is simple: if the relationship data is trusted, AI agents can take over the repetitive work, allowing investment professionals to spend more time exercising judgment and building relationships.

"Technology replaces none of that. It makes sure we remember all of it." — Niklas Krusche

The firm built its agent layer on that foundation. Today, AI is woven into the firm’s daily workflow. Claude interacts with Affinity through the Model Context Protocol (MCP), allowing agents to read and write relationship data securely under defined permissions. Instead of navigating menus or exporting spreadsheets, team members simply ask questions in natural language. Preparing for meetings, updating CRM records, checking pipelines, deduplicating research lists, and recording meeting notes all happen conversationally, with Affinity remaining the single source of truth.

On top of that, Armira runs scheduled agent workflows it designed itself, reading from and writing to Affinity under scoped access rights, defined guardrails, and human review of material changes. Importantly, Armira has been deliberate about governance. AI agents handle repetitive operational tasks, while anything material remains subject to human review and approval. The objective is not to automate investment judgment, but to automate everything surrounding it.

The clearest example is the firm's market-intelligence sync, an agent Armira built and runs. The process runs almost entirely in the background. Every week the agent reviews market-intelligence inputs, updates pipeline records within predefined guardrails and surfaces new opportunities for human review. What was previously a manual workflow has become a continuous intelligence process.

Affinity's MCP server also feeds the rest of Armira's stack. Alongside Claude, the firm connects Affinity to its market-data provider and its AI note-taking tool, and the agent layer is the heaviest consumer, with email triage, meeting-notes processing, and research workflows all reading pipeline context from Affinity through MCP. In Armira's stack, Affinity moved from one destination among many to the structured layer underneath the firm's entire external-engagement effort.

"If it is not in Affinity, our tools do not know it, which is exactly the discipline we want." — Niklas Krusche

Results: a firmwide relationship system

What began as an origination platform has evolved into the firm's relationship operating system. Capital Formation manages the full investor lifecycle through Affinity. M&A advisors, industry experts and broader relationship networks all sit within the same ecosystem, alongside event planning and coordination.

The Capital Formation team is extending the data structures Armira built on the origination side to the investor side, which is straightforward because the process shapes are similar and AI-assisted data ingestion has made the work far lighter.

"It has evolved from a sourcing tool into the firm-wide relationship system." — Niklas Krusche

For a growing share of Armira's investments, the firm can trace the relationship back to a proactive first contact made years before a formal process began. That rests on something that did not exist before: full-firm pipeline visibility.

For firms weighing where agentic AI fits in their own origination, Armira's experience comes down to one thing: the agent layer is only as good as the data underneath it, and most of the work is getting that data right first. For Armira, Affinity is where that data lives.

Key takeaways

• Clean relationship data is a prerequisite for successful AI. Without trusted data, AI simply scales inconsistency.
• Automated capture solves adoption at the root. Because logging is automatic and notes are pushed by agents, Armira never had to mandate use. Partners adopted the system once they saw their own relationship history assembled without having logged anything.
• Agents extend a lean team's reach. An 80-person firm runs origination at a scope well beyond its size by handing routine work to scheduled agents under scoped access and human review, so people spend their time on judgment instead of assembly.