Oops! Something went wrong while submitting the form.
This quarter, we deepened our impact across finance and insurance by delivering product updates, completing two global accelerators, and sharing our perspective on AI adoption at major industry events.
From launching conversational interfaces to rethinking how teams scale expertise, every initiative this quarter focused on the same goal: turning institutional knowledge into production-grade AI workflows.
Key Takeaways
We launched Chat 1.0, a purpose-built conversational interface that delivers contextual, auditable AI interactions for business units
Instant Setup 1.0 introduced streamlined agent deployment, enabling users to configure agentic AI workflows in minutes
AgentFlow gained traction through two major accelerators, showcasing how it preserves institutional knowledge and integrates into existing systems
Multimodal was named to the AI Hot 100, recognized for building production-grade AI that addresses operational pain points in regulated industries
The team actively shaped the industry conversation, with leadership speaking at InsurTech Insights NY and sharing insights on startup–enterprise collaboration
Feature Release: Chat 1.0
We introduced Chat 1.0, a purpose-built conversational interface designed to support enterprise-grade interactions with AI agents.
Unlike generic chatbots, Chat 1.0 operates within structured workflows and persistent knowledge Spaces, ensuring every response is contextual, traceable, and production-ready. Users can:
Use selected Spaces as knowledge bases
Navigate across sessions via a left-hand pane
Interact in a central conversation window with clear AI responses
Verify each answer through a dedicated Trust Center showing the agent's reasoning and source trail
This structure allows underwriters, claims processors, and analysts to engage in fast, auditable dialogue with the agents managing their work.
It’s also built for iterative improvement: users can rate each answer with a thumbs up/down and copy any message. Behind the scenes, each session is tied to a selected set of Spaces, meaning every chat is grounded in relevant, reviewable context.
We are already working on Chat 2.0, which will allow users to see sources and in-text citations for each response.
New Feature: Instant Setup 1.0
We also introduced the first version of Instant, a streamlined way to build and run an AI workflow with minimal upfront configuration.
Users can also select a desired model and schema for each agent:
For Document AI, users can set classification and extraction instructions and import the desired schema from a JSON file
For Decision AI, users can set up decisions, logic, and outcomes
For Report AI, users can define a report structure
Users can select a desired model for all agents individually, except for Unstructured AI
While Instant 1.0 is functional and already powering quick pilots, we’ve already scoped a 2.0 version. Development starts in a week and will include extended agent configuration options and even faster onboarding. Instant 2.0 will also allow users to leverage Instant Spaces for reusable context.
Google Cloud AI Accelerator: Demo Day and What’s Next
We also wrapped up our time in the Google for Startups Cloud AI Accelerator, where Multimodal was one of just 15 startups selected globally. The program concluded with Demo Day, where our VP of Engineering, Andrew McKishnie, pitched our platform on stage to investors, enterprise leaders, and product experts.
Over the course of the accelerator, we worked closely with the Google Cloud team and a global cohort of founders to confront some of the most persistent challenges in enterprise AI. For us, that meant pressure-testing how our agentic AI platform, AgentFlow, tackles operational backlogs and processing delays without forcing teams to rip and replace core systems.
AgentFlow is designed to integrate invisibly, move securely, and scale intelligently inside banks and insurers. That mission resonated strongly throughout the accelerator, and we’re walking away with sharper product focus, new strategic relationships, and real traction heading into Q3.
This quarter, we completed our 12-week journey with the FinTech Innovation Lab New York, a program run by Accenture and the Partnership Fund for New York City.
The experience culminated at Demo Day, where our founder and CEO, Ankur Patel, presented how Multimodal is addressing a critical but often overlooked challenge: the quiet loss of institutional knowledge.
When M&As, reorgs, or retirements hit, companies often lose the tacit expertise that keeps operations moving. AgentFlow tackles that head-on by turning subject-matter expertise into AI agents, making expert workflows scalable, traceable, and auditable.
Throughout the Lab, we partnered closely with financial institutions to validate new use cases, refine our product roadmap, and strengthen go-to-market alignment.
Company Recognition: Multimodal Joins the AI Hot 100
The AI Hot 100 Summit brought together some of the most promising startups in generative AI, alongside enterprise leaders, investors, and researchers, to explore what meaningful innovation looks like in this next wave of AI adoption.
We were featured on the AI Hot 100 list, joining a cohort of companies shaping production-ready applications of AI. Not just building flashier interfaces, but addressing the deep operational friction inside enterprises.
In our case, that means solving a problem most teams feel but few can name: losing institutional knowledge when employees leave. Whether it’s an M&A, a reorg, or staff turnover, critical expertise disappears, and with it, speed, consistency, and compliance.
AgentFlow turns that expertise into auditable, repeatable AI workflows across underwriting, servicing, claims, and more. Built specifically for regulated environments and integrated directly into existing systems, it’s helping banks and insurers do more with the teams they already have.
InsurTech Insights NY: Accelerators as a Testing Ground
This quarter, our founder and CEO took the stage at InsurTech Insights NY, joining a panel on the Blue Stage alongside Dr. Thomas Rodewis, Tim Rollender, Caitlin Herling, Ashlyn Lackey, and Alexander Paruschke.
The topic: what makes startup–corporate partnerships actually work and how accelerators can serve as a real-world proving ground for that collaboration.
For Multimodal, accelerators are about pressure-testing how fast we can help insurers move from experimentation to deployment. That means surfacing business-critical pain points and showing how AI, when designed for operators, can actually solve them.