A $45M Bet on AI That Acts, Not Just Analyzes
San Francisco-based Zocks closed a $45M Series B round co-led by Lightspeed Venture Partners and QED Investors, bringing total funding to $65M.[1] The round follows a $13.8M Series A in March 2025.[2] More than 5,000 financial firms now use the platform, including Ameritas Life Insurance, Cambridge Investment Research, and Carson Group.[3]
What makes Zocks notable is not just the size of the raise. It is the category the company operates in: agentic AI. Unlike traditional AI that surfaces insights and leaves humans to act on them, agentic AI takes the next step. It fills out forms, drafts follow-up emails, opens accounts, and routes tasks through compliance workflows. Zocks claims its platform saves financial advisors more than 10 hours per week on administrative work by turning client conversations into structured data and automated actions.[1]
The new capital will expand those capabilities beyond administrative automation. Zocks plans to use agentic AI to identify planning opportunities across an advisor's entire book of business, essentially telling advisors what to do next based on contextual insights drawn from every client interaction.[3]
What Zocks Does: From Conversation to Action
Zocks is a privacy-first AI platform built for financial advisors.[3] At its core, the technology listens to client conversations and converts them into structured, actionable data. But the "agentic" part is what separates it from standard transcription or note-taking tools.
Here is what the platform does after a client meeting:
- Generates structured records from unstructured conversation, extracting key data points like account details, life changes, and financial goals.
- Triggers follow-up workflows automatically, including drafting email replies, initiating account openings, and populating compliance forms.[1]
- Integrates bidirectionally with existing financial planning and CRM platforms, pushing data into the systems advisors already use.
- Identifies planning opportunities by analyzing patterns across the advisor's full client base, surfacing recommendations proactively.[3]
The pattern here is clear: capture data from natural interactions, structure it, and then act on it without waiting for a human to do the manual work.
Why Agentic AI Matters for Lending
The same pattern Zocks applies to financial advisory (conversation to structured data to automated action) maps directly to challenges in lending operations.
Alternative lenders process thousands of applications monthly. Each application triggers a chain of verification steps, document collection, borrower communication, and compliance checks. Most of these steps still involve manual handoffs, email follow-ups, and human data entry.
Agentic AI changes that equation. Instead of generating a report that an underwriter reads and acts on, an agentic system would:
- Pull entity verification data and route the application based on the result, automatically flagging dissolved or suspended businesses for review.
- Draft borrower communications requesting missing documentation, personalized to the specific gaps in each file.
- Monitor portfolio health continuously, alerting risk teams to entity status changes or compliance triggers without waiting for a periodic review cycle.
- Fill compliance forms and generate audit-ready documentation as a byproduct of the underwriting process, not as a separate step.
The operational savings Zocks reports (10 or more hours per week per advisor) hint at the scale of impact when similar automation reaches lending workflows where the volume of repetitive tasks is even higher.[1]
The VC Signal: QED Investors Co-Leading Says Something
The investor composition of this round deserves attention. QED Investors is not a generalist venture firm. Founded by Capital One co-founder Nigel Morris, QED focuses exclusively on fintech.[4] Their portfolio includes Nubank, SoFi, Kavak, and CreditKarma.[5]
When QED co-leads a $45M round in agentic AI for financial services, it is a directional signal. The firm has deep conviction, built over decades of operating in financial services, that AI-driven automation is entering a new phase. The shift from "AI as analysis tool" to "AI as execution layer" aligns with what QED sees across their portfolio companies: the biggest ROI comes not from better dashboards but from fewer manual steps between data and decision.
Lightspeed Venture Partners, the other co-lead, brings enterprise SaaS expertise with investments in Snap, Mulesoft, and Nutanix.[6] Their participation reinforces the enterprise readiness angle. Zocks is also expanding security, compliance, and integration capabilities with this round, all signals that the platform is moving upmarket into regulated financial institutions.[3]
What Alternative Lenders Should Be Watching
Agentic AI is not a theoretical concept for lending. Several concrete applications are already emerging:
Underwriting workflow automation. Instead of underwriters manually checking entity status, pulling documents, and making routing decisions, agentic systems can execute that entire chain. The underwriter reviews the output and makes the final call, but the research, data gathering, and initial routing happen automatically.
Borrower communication at scale. Agentic AI can draft and send personalized follow-ups to borrowers with incomplete applications, request specific missing documents, and handle routine inquiries. This reclaims hours currently spent on email and phone follow-up.
Portfolio monitoring. Rather than running periodic batch reviews of borrower entity status, agentic systems can monitor continuously and flag changes in real time. A business that moves from "Active" to "Suspended" in a state database triggers an automatic review workflow, not an email to an analyst asking them to check.
Compliance documentation. Every verification, communication, and decision point generates an audit trail automatically. Instead of compliance teams reconstructing what happened after the fact, the documentation is created as a byproduct of the workflow itself.
Building the AI-Ready Data Infrastructure
There is a prerequisite that makes all of this work: clean, real-time, structured data. Agentic AI is only as good as the inputs it receives. If the entity data feeding an automated underwriting workflow is stale, incomplete, or inconsistent, the AI will make confident decisions on bad information. That is worse than no automation at all.
This is where data infrastructure becomes a strategic priority. AI systems making automated decisions about business entities need authoritative, real-time verification data from primary sources. Automated SOS verification, for example, provides the kind of structured, timestamped entity data that agentic AI systems require to make accurate decisions without human intervention.
The Zocks raise is one data point in a larger trend. As agentic AI matures across financial services, the lenders who have already invested in clean data pipelines, API-first verification, and structured compliance workflows will be positioned to adopt these tools first and benefit most. The ones still running manual processes will find the gap widening.
The $45M question is not whether agentic AI will reach lending. It is whether your data infrastructure is ready when it does.
Cobalt Intelligence provides real-time Secretary of State data and business verification APIs for lenders, fintechs, and compliance teams. Learn more or schedule a demo.
.jpg){kind=small}
.png)