From Chatbots to Balance Sheets: What the New AI Agent Wave Really Is
Enterprise AI agents are autonomous software systems that connect to a company’s data and tools so they can plan, decide, and execute repetitive or complex workflows end-to-end with minimal human oversight, rather than stopping at static recommendations or chat-style answers. The funding spike around these agents is not about smarter chatbots; it is about machines that can approve invoices, assemble deal packages, and ship production-ready designs inside real business systems. The key shift is from AI as a helper that “suggests” to AI as an operator that takes responsibility for back-office, financial, and transaction-heavy tasks that previously demanded large, often outsourced teams. That change is now pulling in serious AI startup investment as boards ask for cost savings, faster cycle times, and systems that retain institutional knowledge instead of letting it walk out the door.
Freehand and the Supply Chain AI Land Grab
The clearest sign of this new direction in AI agent funding is Freehand, which raised USD 75 million (approx. RM345 million) in a Series B to scale its autonomous AI agents for complex supply chain spend and back-office operations at large enterprises. Freehand’s agents sit inside existing systems, read contracts and emails, check supplier bills against real-world milestones, and even take “business calls” that once required tribal knowledge, to the point of “no human intervention” on many decisions. One quotable summary is that “Freehand has raised USD 75 million (approx. RM345 million) in a Series B funding round to scale its autonomous AI agents, which manage complex supply chain spend and back-office operations for enterprise companies”. This is supply chain AI aimed directly at Fortune 500 governance, not another expense-card tool. Investors are betting that tariffs, taxes, and immigration policy straining the old outsourcing model will push enterprises toward autonomous AI platforms that can replace sprawling offshore back offices with in-house, software-driven control.
Henry AI: Automating the Memory of Commercial Real Estate Deals
Commercial real estate is one of the most spreadsheet-heavy, email-bound sectors left in modern finance, which is exactly why Henry AI’s USD 16.5 million (approx. RM76 million) Series A matters. The startup, used by more than 150 firms, claims its platform has produced over 20,000 client-ready deliverables tied to more than USD 150 billion (approx. RM690 billion) in underlying deal value, while cutting analyst production time by about 90%. That is not a chat interface; it is an operating system for deals. Henry’s new Henry Deal product tries to capture every action across a transaction, pull a firm’s scattered files and comps into an ordered “context engine,” and make the next deal faster because the last one taught the system something. In other words, enterprise automation here is about institutional memory. The firms that adopt it first will have agents that remember every rent roll and comp; those that delay will keep paying analysts to reinvent the same deck over and over.
Paper vs. Figma: When AI Agents Design in Code, Not Pictures
On the design side, Paper’s USD 34 million (approx. RM156 million) Series A is a shot at the dominant design tool precisely because it is built for AI agents, not humans alone. Paper argues that design tools built before the AI coding era force an unnecessary translation step from proprietary formats to production code, while it renders everything directly in HTML and CSS. Its MCP server ships 24 bidirectional tools, letting AI agents both read from and write to the canvas, compared with three read-only tools in the incumbent platform. After Paper Desktop launched, the company says ARR grew 25x in a single month, and benchmarking cited by the company claims its agent workflows are faster, more accurate, and use fewer tokens than agents in other design tools. The implication is blunt: if AI coding agents become central to how software is designed and shipped, then design surfaces that double as working code—and as shared workspaces for agents and humans—will displace static mockup tools.
Why Investors Love Domain-Specific Autonomy—and What Comes Next
Across Freehand, Henry AI, and Paper, the pattern is clear: funding is rotating away from general-purpose LLM wrappers toward domain-specific autonomous platforms that attack concrete operational pain. Supply chain investors highlight Freehand’s focus on the largest shippers, not intermediaries, and on “clear, measurable business outcomes” from automating financial governance. In commercial real estate, Henry is positioning itself as a context engine embedded in the transaction, with claims that nine of the top 10 brokerages already use it. Paper is betting that design tools wired for AI coding agents—down to rendering model-friendly HTML and CSS—will define how teams ship software. Enterprise AI agents are no longer demo toys; they are executing back-office, finance, and deal-making workflows, and even spreading beyond design teams to engineers, sales, and growth as a visual interface for their agents. The next wave of AI startup investment will likely reward teams that pick a messy, high-value workflow and give an autonomous agent the keys, rather than those that ship another generic chat window.
