.avif)
.svg)
Forum's Venture Studio brings together ambitious people, brilliant ideas, and capital to build the best B2B SaaS businesses in the world, from 0 to 1. In addition to capital and an idea, we support founders with a full design and development team, go-to-market support, strategy, and back-office (accounting, legal, HR) support. We also offer full fundraising support, including narrative and pitch deck creation, investor outreach plans, and investor introductions. Our model is designed to help you move faster, develop better insights, and build companies that have a higher success rate than startups built in any other way.
At Forum, we believe AI will fundamentally change how business is done across every industry. It’s now been 2.5 years since ChatGPT launched. At this point, no one can seriously argue that AI isn’t here or that it hasn’t gone mainstream.
There’s no single playbook for building an AI company today, and that’s what makes this moment so exciting. We’re seeing companies being built across different models, and we’re backing the best of each.
Vertical SaaS: AI-powered tools that automate painful workflows in specific industries.
Dev Tooling: Infrastructure that makes it easier for teams to build, integrate, and scale AI across the stack.
Full-Stack AI Companies: End-to-end platforms that reinvent how entire industries operate by rebuilding their internal tools and workflows with AI at the core.
Our venture studio is launching five new ideas we’re looking to invest in. We’re on the hunt for founders to turn them into real companies.
The Ideas:
Automating NPMS Compliance for Pipeline Operators
Context
The pipeline industry faces an annually recurring, federally mandated headache, NPMS (National Pipeline Mapping System) submissions. These filings require meticulous geospatial accuracy, precise metadata formatting, and compliance with complex regulatory standards. Mistakes or late submissions trigger steep penalties of up to $200,000 per violation per day and heightened regulatory scrutiny.
Currently, most operators rely on manual processes or expensive consultants to navigate the Operator Submission and Validation Environment (OSAVE), a system that, while free to access, is notoriously difficult to use. The task is time-consuming, high-stakes, and ripe for error, especially given the highly technical GIS requirements. As regulations tighten and the scale of pipeline infrastructure grows (including the addition of hydrogen and carbon pipelines), the burden of compliance will only increase.
The Market
This is a high-pain, no-choice entry point into the much larger energy compliance market. All 3000+ hazardous liquid and gas transmission operators in the U.S. are federally required to submit NPMS data annually. These companies already allocate between $4K–$15K per submission internally (often more when outsourced), and failure to comply carries penalties of up to $200,000 per violation per day.
While the immediate addressable market for NPMS is narrow, it's incredibly acute: every operator must comply, every year, and there are no viable end-to-end software solutions serving this need today.
More importantly, NPMS is just the beginning. The same customers face dozens of similar, complex, and recurring reporting and certification requirements, covering everything from pipeline abandonment to environmental impact, emissions reporting, and PHMSA audits. Building trust through NPMS creates a direct path to expand into adjacent workflows across geospatial data management, environmental compliance, and integrity validation.
The Opportunity
Build a submission automation and compliance platform for NPMS reporting. This platform will handle everything from geospatial data validation and attribute packaging to metadata formatting, abandonment certification preparation, and OSAVE-compatible file generation. It will also support ongoing submission management, ensuring operators remain in full compliance year over year.
Beyond NPMS, the platform can serve as a launchpad into other underserved, compliance-heavy workflows in the oil and gas sector, such as PHSMA integrity certifications, state-level pipeline filings, or emissions reporting.
Why We’re Excited
This is the type of problem AI was made to solve: high-volume, high-stakes, and rules-based, with enormous consequences for error. With mandatory use, annual recurrence, and a fragmented services-based status quo, we believe this is a rare wedge into a highly defensible vertical. The customers are giant, multinational companies with large, dedicated compliance budgets and this is the type of market where a first-mover AI platform can become indispensable infrastructure.
Securing the Next Layer of AI Infrastructure: MCP Safety Auditing
This specific idea focuses on MCP security, but we see MCP as a foundational layer in the AI stack. We're actively looking to invest across the MCP ecosystem, including infrastructure, developer tooling, monitoring, and safety. If you're building anything within the MCP ecosystem, we want to chat.
Context
As AI agents become more autonomous and embedded into enterprise workflows, Model Context Protocol (MCP) is rapidly emerging as the standard that connects LLMs to tools and data. Unlike traditional APIs, MCP was purpose-built for AI, enabling agents to dynamically discover, query, and execute tools without human intervention. But this flexibility comes at a cost: security vulnerabilities.
Recent research has shown that even the most aligned models can be manipulated into executing harmful actions, exposing organizations to risks like remote access control, credential leakage, and retrieval-based deception attacks. While companies like Postman, Composio, and Microsoft are racing to expand their MCP capabilities, they’ve left gaps in the security layer.
The Market
MCP is no longer experimental, it’s being deployed at scale. Microsoft has embedded MCP capabilities directly into Windows 11, OpenAI has integrated it into its Agents SDK, and Postman now supports MCP requests as part of its core offering. These moves signal a shift: AI-native protocols are becoming foundational to modern software.
While the MCP ecosystem is still in its early days, we believe it will become a core part of the AI infrastructure stack. We expect this market to explode over the next 3 to 5 years.
The Opportunity
Build a security auditing and monitoring platform purpose-built for MCP servers and agents. This platform will enable companies to proactively identify vulnerabilities, validate the safety of their agent workflows, and ensure compliance with internal and external security standards. Rather than a general-purpose observability tool, this would be a deeply integrated product that speaks the language of AI agents and understands the unique threats introduced by LLM-powered automation. By embedding security into the MCP lifecycle, across development, deployment, and iteration, we can help organizations prevent exploits before they occur.
Why We’re Excited
As the MCP protocol becomes foundational to how AI agents interact with the world, enterprises will demand guardrails that ensure these systems don’t compromise safety or security. Today, those guardrails don’t exist.
The AI-Native Freight Brokerage
Context
Freight brokerage in the U.S. is still astonishingly manual, despite operating in a $51.7B+ market with 24/7 demands, heavy compliance needs, and razor-thin margins. Traditional brokers are overrun with fragmented processes, rate negotiation, carrier vetting, document handling, and real-time updates, all managed through phone calls, emails, and outdated TMS platforms. Digital freight marketplaces like Uber Freight and Flexport made early waves, but their marketplace-centric models often fall short in a relationship-heavy industry that still depends on human intuition and operational nuance. Meanwhile, point solutions abound, but adoption is shallow, especially among mid-market brokers who continue to rely on spreadsheets and phone calls. The industry is overdue for a radical rethink, not more tools, but a new kind of brokerage.
The Market
Freight brokerage is a massive, fragmented industry with nearly 80,000 players and no dominant winner. The freight brokerage market is expected to grow at a 6% CAGR through 2032, and the broader U.S. trucking industry generates nearly $1T annually.
But most startups are focused on selling software to existing brokerages limiting the ability for their to be real innovation in the industry. Given the operational complexity, owning the entire stack is the more compelling long-term play.
The Opportunity
Build a vertically integrated, AI-native freight brokerage from the ground up. This brokerage will be powered by autonomous agents capable of executing every key operational function, from carrier matching and rate negotiation to compliance vetting, document generation, shipment tracking, and customer communication.
By controlling the entire experience, we can deliver a level of speed, precision, and 24/7 reliability that traditional and digital brokerages simply can’t match. Rather than layering automation on top of outdated workflows, we’re rethinking the brokerage itself as a fully AI-orchestrated operation, purpose-built for scale, trust, and resilience in one of the most complex logistics environments in the world.
Why We’re Excited
As the freight industry moves toward consolidation and automation, a greenfield opportunity exists to leapfrog incumbents by designing a brokerage purpose-built for AI.
Agentic automation allows us to compete not by marginally improving broker workflows, but by reimagining them entirely. With mounting compliance pressure, rising fraud, and growing complexity across the supply chain, carriers and shippers need a partner that’s fast, reliable, and always-on. If done right, this company will completely reshape a trillion-dollar industry.
AI-Powered Inspection Readiness for Foreign Manufacturers
Context
The FDA’s May 2025 policy shift to expand unannounced inspections at foreign drug, device, and food manufacturing plants fundamentally changes compliance expectations. No longer can companies rely on last-minute scrambles, facilities must maintain constant audit readiness. Yet, most still operate with fragmented systems, hybrid paper records, and face high QA staff turnover. This makes proactive compliance extremely difficult and puts billions in U.S. market access at risk.
A single Form 483 observation can cost $1–1.5M in remediation and halt revenue for up to nine months. While existing QMS players focus on digitization and defect detection, none are built to anticipate regulatory findings or dynamically orchestrate corrective action in real time.
The Market
With ~3,000 foreign sites audited annually and over 30% receiving Form 483s, the opportunity to prevent costly compliance failures is massive. The global pharmaceutical quality management software market is projected to exceed $2B by 2025, growing at a 13% CAGR. Moreover, 60%+ of prescription drug and API manufacturing is now offshore, yet these facilities face language barriers, limited cloud access, and frequent QA turnover. Despite funding in adjacent categories (Dot Compliance, TraceLink, Qualio), there’s a clear whitespace for an AI-first, FDA-specific inspection solution that addresses these structural barriers head-on.
The Opportunity
Build an AI-powered inspection copilot that predicts likely FDA deficiencies and orchestrates real-time CAPA (Corrective and Preventive Action) workflows. This platform will facilitate historical Form 483 pattern analysis, PDF/SOP ingestion, multi-lingual NLP for deviation logs, and auto-prioritized remediation tasking. We can transform FDA compliance from reactive to predictive, functioning as a regulatory radar system for foreign QA teams.
Why We’re Excited
We love building software in response to new regulatory shifts—they create urgency, keep problems top of mind for customers, and act as a powerful forcing function in the buying process.
As the FDA ramps up surprise audits abroad, manufacturers must embed always-on readiness into their operations. The compliance stakes are high, the pain is constant, and the whitespace is unmistakable.
Data Observability for CAT Models in Insurance-Linked Securities
Context
As climate volatility increases and extreme weather events become more frequent, the reliability of catastrophe (CAT) models used to underwrite and price risk is under growing scrutiny. These models are foundational to Insurance-Linked Securities (ILS), which now represent a $100B+ asset class. Yet, the data powering CAT models is often incomplete, outdated, or miscalibrated especially as events outpace historical precedents. Regulators like the NAIC (US) and Solvency II (EU) are intensifying requirements for data quality audits and model governance, but most modeling entities still lack the tooling to ensure transparent, trustworthy data flows. Current solutions are generic and not designed for the complex, fragmented, and context-heavy datasets used in CAT modeling.
The Market
The CAT modeling and ILS markets are rapidly growth, with over $36B in outstanding catastrophe bonds and an expanding ecosystem of reinsurers, MGAs, brokers, and insurtechs. At the same time, AI is reshaping expectations around explainability, traceability, and data integrity, creating demand for domain-specific data observability tools. Unlike broader observability platforms (e.g., Monte Carlo, Cribl), which aren't tailored for insurance, a focused solution for CAT data validation and monitoring has few direct competitors.
The Opportunity
Our current thesis is to build a data observability platform purpose-built for catastrophe modeling in the ILS space. The platform will act as a real-time audit layer for CAT data, offering end-to-end visibility into the health, structure, and reliability of data flowing into models. It will surface when something breaks, what broke, and how to fix it, enabling underwriters and fund managers to make decisions with confidence.
The product will serve as a “check engine light” for model inputs, delivering intelligent monitoring, anomaly detection, and compliance readiness for Solvency II and NAIC frameworks. With increasingly complex datasets and rising demand for transparency in capital markets, this kind of domain-specific observability layer is both overdue and mission-critical.
Why We’re Excited
Climate-driven losses are accelerating, pushing capital markets to play a larger role in absorbing and pricing climate risk. This has propelled ILS into a $100B+ mainstream asset class. But with that growth comes mounting pressure, from regulators, investors, and rating agencies, to ensure the underlying models are transparent and resilient.
As scrutiny increases, the data pipelines powering CAT models are emerging as the weakest link. We believe this mirrors past shifts in sectors like fintech and healthcare, where data integrity evolved from an operational issue to a strategic imperative.
Build with Forum
We believe the next wave of iconic SaaS companies will be AI-native from day one. The ideas we’re backing are just the starting point, what matters most is the founder behind them. If you’re a builder ready to reshape how industries operate, we want to hear from you.
If you’re interested in building with Forum please apply through the corresponding founder-in-residence description. If none of these areas fit what you’re working on, feel free to apply through the Forum Studio Founder posting. Please highlight relevant experience and the area you’re most interested in.
Founder-In-Residence Opportunities:
-min.png)
.jpg)
.avif)
