To make Wealthtender free for readers, we earn money from advertisers, including financial professionals and firms that pay to be featured. This creates a conflict of interest when we favor their promotion over others. Read our editorial policy and terms of service to learn more. Wealthtender is not a client of these financial services providers.
➡️ Find a Local Advisor | 🎯 Find a Specialist Advisor
[The accelerating rate of AdvisorTech innovation is truly breathtaking. No sooner than a new tech tool appears, it starts an inexorable climb of new enhancements and iterations (2.0, 3.0, 4.0…) with creative technologists developing an ever-increasing and widening array of capabilities and applications. Nowhere is this more apparent than in the critical areas of data management and AI technology, as we are still in their early stages of development.
Once data is properly collected and organized, the realization sets in on how companies are sitting on a goldmine of opportunities. What can you do with all that data? How can technology create more value from that data?
To better understand the dynamic behind this AI tech evolution and the opportunities it can provide for advisors and wealth managers, we reached out to Institute members Era Jain, Co-Founder & CEO, and Divam Jain, Co-Founder & CTO of Zeplyn – a purpose-built AI assistant platform that streamlines nuanced wealth management workflows and improves client experience and engagement.
As former Google engineers with decades of expertise in building AI products, we asked them questions to better understand how they developed their initial flagship product, Zeplyn Meeting Assistant, into an AI-powered, workflow intelligence and practice management platform with new productivity features that were specifically designed for advisors, wealth managers, and Broker-Dealers.]
Hortz: Can you briefly explain to us what an AI Agent is? What is the technology behind AI Agents?
Divam: Unlike traditional reactive AI tools, such as chatbots that respond to single prompts, AI agents are autonomous digital workers. They are designed to understand context, plan, and execute multi-step workflows, and adapt along the way. Once given a defined goal, they operate with minimal human oversight.
At Zeplyn, we believe the next evolution in wealth management will not be unlocked by another dashboard. It will come from agentic AI: intelligent assistants that can proactively execute workflows end-to-end, making decisions, and taking action just as a teammate would. Imagine an assistant that does not just summarize meeting notes, but also drafts personalized follow-ups, updates your CRM, analyzes portfolio impact, and schedules your next check-in, automatically. That is the difference with an agentic system.
The technology behind AI agents typically includes two core components:
Orchestrators: These are the strategic planners. They determine which tasks need to be executed, in what order, and by which agents or tools. Think of the orchestrator as the brain of the operation that is responsible for planning tasks, adapting to new information, and coordinating workflows across systems.
Workers: These are the executors. They execute specific tasks assigned by the orchestrator, such as pulling data from client records, generating reports, or interfacing with CRMs. A worker could be a generative AI model like GPT, a tool for accessing databases or APIs, or even another smaller agent optimized for a particular function.
It is this combination of intelligent coordination and execution that makes AI agents so powerful.
Hortz: How do you then train an AI agent to perform specific functions for advisors, like note-taking?
Divam: Training an AI agent to perform a specific function, like note-taking for advisors, requires more than handing it a script. A typical Zeplyn agent is purpose-built with four components:
- A clear task instruction – defining exactly what it needs to accomplish.
- A large language model (LLM) – chosen for that specific task (different LLMs work better for different kinds of tasks – for example, summarization, reasoning, creativity, classification).
- Context – including industry norms, firm-specific workflows, and advisor preferences.
- Inputs – from other agents or humans.
Zeplyn Meeting Assistant captures actionable meeting notes from client conversations across any channel: video conferencing, phone calls, or in-person meetings. At face value, note-taking seems simple. In wealth management, however, nuance really matters. A misattributed task or missed detail can break client trust or lead to operational risk. That is why we decompose this “note-taking” workflow into multiple specialized agents:
Speaker recognition agent: This agent uses a diarized transcript and meeting context to accurately assign who said what. This is essential for attributing action items correctly. Did the advisor commit to something, or was it the client?
Meeting type classification agent: Next, we use an agent trained to classify the meeting purpose: annual review, planning, service team follow-up, etc. This step is crucial because it helps the system prioritize relevant content and filter out noise.
Action item detection agent: Using inputs from the prior agents, this agent then extracts structured, actionable tasks – who’s responsible for doing what and by when. For example, “Send updated estate documents” becomes a tagged task assigned to the advisor or the client service team.
Intelligence extraction agent: Beyond tasks, we have a collection of customizable agents that collect data relevant for different purposes. For example, we use dedicated agents for generating meeting summaries, structured compliance records, relationship signals, and anything else an advisor could want to document explicitly.
Advisor-in-the-loop execution: Once identified, tasks can be passed downstream to execution agents, such as a CRM updater. For compliance reasons, this step in Zeplyn is advisor-assisted, meaning we require final advisor review and approval before any data is pushed.
Ultimately, we are building solutions that leverage both fully autonomous and human-assisted agents to meet the compliance and regulatory standards of the wealth management industry, while minimizing the administrative burden on the advisory firms.
Hortz: Starting with an AI agent for advisor note-taking, how did you leverage that technology into a workflow intelligence platform, adding new tools and capabilities for advisors and wealth managers?
Era: When we started out, we noticed a significant data gap in the industry: 60% of client data gathering was happening during client meetings and less than 25% of these meetings were properly documented in the CRM. The manual nature of the task was too time-consuming. Our first solution, Zeplyn Meeting Assistant, was built to close this gap.
And what started as a smart notetaker was really a launchpad into workflow intelligence. By applying our agentic AI technology to highly accurate and structured meeting notes captured by Zeplyn Meeting Assistant, we are now able to help advisors streamline the full meeting lifecycle from pre-meeting prep to post-meeting follow-ups and execution.
This includes, but is not limited to, composing personalized client follow-up emails, equipping client service teams with prioritized next steps, updating CRMs, preparing agendas for next meetings, and surfacing client trends and insights to empower firms to optimize practice management. Through this, we reduce advisor administrative burden by more than 90% and give firms actionable insights to grow their business.
We are building toward a future where agentic AI supports and empowers advisors to work in an entirely new way, one that allows them to focus on strategy and nurturing client relationships.
Hortz: Can you share with us what your thought process and vision was in reimagining wealth management from your experiences as a Google engineer and AI-native perspective? How did that inform your decisions on how to build the overall functionality of the new platform?
Era: At Google, we built AI systems that powered products like Search, Assistant, and Speech-to-Text. We lived at the intersection of unstructured data and intelligent automation, engineering systems that could understand language, extract meaning, and act on it at scale. And while we were automating billions of tasks a day on the web, professionals in high-trust, high-stakes industries like wealth management were still buried in manual work.
Zeplyn was born from that contrast. Divam and I saw an opportunity to draw heavily from the principles we used to build accurate and scalable AI at Google to build specialized AI assistants to understand advisor workflows and handle advisor busy work.
We have a big vision: reimagine the financial advisor experience from an AI-native perspective. To do that, we focused on five core tenets:
- Ground every automation in context: Generalized AI is not enough in wealth management. We built specialized AI agents to understand the industry’s nuances, so advisors can get industry-leading accuracy in capturing relevant insights from every client interaction. With a highly contextualized solution, our output is reliable and free from hallucinations.
- Design with interoperability: Advisors do not need another siloed tool. They need intelligent systems that work well with the investments they have made—and with how they naturally work. Zeplyn integrates seamlessly with advisors’ core tech stack (meeting platforms, email, calendars, CRMs, etc.). With an onboarding time of less than five minutes, advisors can immediately start seeing value.
- Build for enterprise-level trust: Trust is non-negotiable. From day one, we designed Zeplyn to meet the industry’s rigorous compliance and security standards. This includes recording-free notetaking, off-the-record mode, PII masking, customized retention, end-to-end encryption, and no client data ever used for general model training.
- Build scalable, customizable systems: Every firm has its special way of operating. We honor that. Our approach is to give financial advisory firms enough flexibility to bring their own data, process knowledge, and collaboration patterns to the experience. At the same time, our platform does the heavy lifting of integrating that context into the agents. The result is a system flexible enough for any financial advisory firm but customized to fit how a specific firm works.
- Keep the human at the center: We did not build Zeplyn to replace the advisor. We built it to amplify their productivity. At the end of the day, our vision is for advisors to spend less time on admin and more time on advice. Our platform is designed to embrace the advisor-in-the-loop approach, putting them in control of every agentic workflow while removing the operational friction that slows them down.
At Google, we saw firsthand how powerful AI can be when applied responsibly, thoughtfully, and with purpose. At Zeplyn, we are bringing that power to the people who need it most – humans guiding other humans through life’s biggest decisions.
Hortz: What are the key benefits you designed into your platform for advisors and wealth managers?
Divam: We designed Zeplyn to solve real pain points that advisors and firms face every day, starting with the manual and administrative tasks that keep data from being properly recorded and advisors from advising.
Our flagship product, Zeplyn Meeting Assistant, transforms unstructured client conversations into structured, actionable insights. It saves advisors 12+ hours a week on note-taking, follow-ups, task creation, CRM updates, and more – and that alone frees them up to focus on higher impact work.
When we expanded, just recently, and introduced a new suite of practice management capabilities, we took those meeting insights and extended them into new workflows that give firms visibility into actionable trends and streamlines the client engagement experience even more. Whether identifying life event opportunities or tracking best practice adoption, Zeplyn gives advisors and firms the ability to act on what matters most.
Now, all of this is powered by agentic AI – which means that these insights can extend and automate multi-step workflows across the advisor tech stack, not just within the meeting experience context. Using the latest advancements in LLMs and agentic AI frameworks, Zeplyn will soon make workflow automation across the disparate advisor stack a reality, turning meeting insights into real action and outcomes. This new technological foundation will transform how firms operate and advisors engage clients.
Hortz: What kind of responses and feedback did you get from your presentations and discussions at the recent 2025 T3 Conference?
Era: AI was at the center of nearly every presentation and conversation at T3 this year. That is how quickly AI is becoming embedded into every area of a firm’s operations. And while there was a lot of curiosity and excitement about it, there was still some understandable skepticism. What value does it actually deliver? Can I trust it? Is it here to replace me?
At T3, we set out to give advisors some answers to those questions while expanding their perspective of the value AI can create for their firms. Once advisors saw how agentic AI can help them identify and act timely on open client opportunities, track how often they are hitting compliance benchmarks, or even systematize something as human as asking for referrals, that is when the opportunity became very clear for advisors. They realized how AI could augment their capacity.
Hortz: Any advice you can offer advisors and wealth managers to help them with their technology purchasing decisions and developing their overall AI and tech stack strategy?
Era: The most important advice I can give advisors and firms is not to treat AI as a one-off tool. It’s not. Treat it as a foundational layer of your business. AI is going to be embedded into every business process, so be thoughtful about your AI strategy, but do not wait too long to take action. Your competitors who did not hesitate, will outpace you.
Here are a few tips to get you started:
Identify the high-friction areas in your business: manual data entry, fragmented meeting workflows, inconsistent follow-ups, or delayed CRM updates. These are areas where agentic AI can immediately reduce workload and improve outcomes.
Think about your data strategy. AI is only as powerful as the data it has access to. Accurate, relevant, and structured data from client interactions powers more intelligent workflows, and domain-specific AI that understands this industry can turn even unstructured inputs into meaningful insights.
Partner with vendors who understand the wealth management space, not just generic AI providers. AI that understands the nuances of this field can make or break your ability to service clients and unlock opportunities to scale.
Looking beyond the short-term fixes is critical. Even if you are solving for one specific pain point today, like meeting notes or CRM updates, make sure the vendors you choose have roadmaps that align with your long-term AI strategy. You want a platform that can scale with you, integrate across your tech stack, and evolve as your needs change.
This article was originally published here and is republished on Wealthtender with permission.“
About the Author
Bill Hortz
Founder Institute for Innovation Development
To make Wealthtender free for readers, we earn money from advertisers, including financial professionals and firms that pay to be featured. This creates a conflict of interest when we favor their promotion over others. Read our editorial policy and terms of service to learn more. Wealthtender is not a client of these financial services providers.
➡️ Find a Local Advisor | 🎯 Find a Specialist Advisor