It is extraordinarily rare for us to write about ourselves; one of our core values is “Keep client interests paramount.”

But I wanted to make an exception and discuss Callan’s artificial intelligence (AI) strategy in order to help educate our clients about the ways in which we use AI, the ways we do not, and how these tools benefit our work for you.

Callan’s AI philosophy is simple: encourage broad experimentation, let people use AI wherever it appears useful, see what actually improves our work, and then take the best recurring use cases and build them into enterprise tools, workflows, or agents. The current model is intentionally bottom-up. We are learning from actual usage rather than trying to impose a single firmwide system before the practical use cases are clear.

I surveyed my colleagues about their use of AI, and their responses suggest that this approach is already producing beneficial results across the firm. In most cases, people are not delegating judgment to the tools. They are using them to shorten the first pass through work, organize information, speed up repetitive tasks, and free up more time for review, interpretation, and client-facing decisions.

Three points stand out:

1. AI is already embedded in ordinary work at Callan. It is not limited to one group or one kind of task. It is being used by investment professionals, operators, developers, legal/compliance staff, and reporting teams.
2. The strongest current use cases are reducing repetitive workflows.
3. There are a few patterns that look mature enough to consider building agents.

Callan and AI: Highlights from the Survey

Below is a summary of the general themes of the responses.

1. The dominant use case is assistance in the early phases of a task or project.
   The most common pattern is not “AI does the work.” It is “AI gets me to a usable first pass faster.” People describe it as an assistant, intern, writing partner, coding partner, or second set of eyes. The final view still belongs to the professional using it. Callan is using AI to support professional work, not to substitute for professional judgment.
2. Summarization and information compression is quite common in the ways we use AI.
   This was the most consistent theme in the survey. People are using AI to process DDQs, quarterly investor letters, manager commentary, meeting notes, legal summaries, research papers, diligence materials, and operating instructions. That is not surprising. A large part of Callan’s work begins with too much text, too many sources, or too much repetition. AI is proving useful at reducing that mass into something short, organized, and easier to convey to clients.
3. Writing support is broad, but it is mostly for editing and refinement.
   The responses show repeated use for email editing, flow improvement, clarity, brevity, memo outlines, presentation scripts, and plain-English explanations of technical or legal points. That is a very practical use case. It is low-friction, widely applicable, and immediately helpful. It also fits the way people seem to be using the tools: not to invent views for them, but to help express those views more clearly and more consistently.
4. The technical groups are getting substantial leverage out of AI.
   In particular, our developers are using AI to reduce the time-cost of smaller code improvements that used to be easy to defer. That is important because once these small improvements become cheap, more of them get done. Over time, that should translate into better tools and cleaner internal processes.
5. AI helps with research and data gathering, albeit with limitations.
   Several respondents described some version of the same experience: AI can get a user to 70% or 80% of an answer very quickly. That is often enough to cut hours out of source-gathering, public-document review, or third-party data collection. In most of these cases, it shortened the setup work and left more time for analysis.
6. A meaningful share of usage is quality control.
   Not all of the examples involve content generation. Several involve checking, ranking, validating, critiquing, or pressure-testing. Users are asking AI to identify missed points, compare peer data, spot classification errors, test themes, and review drafts for inconsistencies. This suggests people are learning to use the tools not only to draft faster, but also to reduce omissions and catch issues earlier.
7. A few use cases are starting to look like potential enterprise level agents.
   Several responses point beyond individual experimentation: automated manager collateral summaries; pre-meeting manager and strategy dossiers; CallanDNA Chatbot trained on internal notes and database information; prospecting, RFP, and market-intelligence tools; recurring report-update workflows; and internal prompt libraries. Those are the places where the current bottom-up philosophy starts to feed into building enterprise agents for our CallanDNA platform.
8. The responses are realistic about where AI still falls short.
   The responses are notable for their lack of hype. People are generally matter-of-fact about the strengths and the limitations. That sober tone helps. It makes the overall picture more credible.

What AI Is Not Doing at Callan

This is worth stating plainly:

• AI is not replacing investment judgment, legal judgment, or client judgment.
• AI is not eliminating the need to verify facts, numbers, or sources.
• AI is not equally useful across all tasks; quality varies meaningfully by use case.
• AI is not yet a fully built-out enterprise platform at Callan; much of the current value still comes from individual experimentation.
• AI is not reducing the importance of subject-matter expertise; it is helping that expertise operate more efficiently.
• AI is not a tool to reduce headcount but is used to make our associates even more productive for clients.

As we continue to refine our work with AI and encourage even more robust experimentation by a wider array of people at the firm, and as we learn from those who are already using it, I expect our work for clients to become better, more efficient, and more focused on addressing the questions they have. I will continue to keep you posted on our efforts, through my annual letter to clients and other venues, but do not hesitate to reach out directly to me with questions or concerns.