Suggestions
Maria Palma
General Partner at Kindred Capital VC and Nonprofit Co-founder
Maria (Brewer) Palma is a General Partner at Kindred Capital VC, a venture capital firm based in London that focuses on equitable investment strategies. She has held this position since January 2021. Palma has a strong background in venture capital and business development, having previously worked as a Principal at RRE Ventures from 2018 to 2020, where she focused on fintech and consumer investments, and played a significant role in sourcing and diligencing numerous deals.
In addition to her role at Kindred Capital, Palma has been actively involved in various organizations. She served as a Board Member for the African Entrepreneur Collective from 2016 to 2022, supporting entrepreneurs in East Africa. She also contributed to the Industrial & Systems Engineering Advisory Board at the University of Wisconsin-Madison from 2019 to 2022, guiding curriculum and student experiences.
Palma co-founded NYC BLEND, a nonprofit aimed at connecting underrepresented founders to the venture ecosystem in New York, and she has held various leadership positions at Eyeview, where she focused on business development and strategic initiatives.
Her educational background includes a Bachelor of Science in Industrial Engineering and an MBA from Harvard University, further solidifying her expertise in the field of venture capital and entrepreneurship.123
Highlights
10 years in VC taught me something I wish I’d known from day one - the best thing you can do for a founder is often nothing at all. https://t.co/EvjoDXf3Df

When Machines Pay Machines: The Era of Agentic Payments Has Arrived
Agents aren't just coming - they're already here. Yet, their full potential won't materialize unless they can autonomously transact and handle money seamlessly. This isn't entirely new - think about the frictionless payment experience of Uber, where transactions invisibly settle behind the scenes. However, extending this model to fully autonomous AI agents introduces unprecedented complexity and scale.
Why Current Infrastructure Isn't Ready
Two major blockers stand in the way of scaling agentic payments: identity authentication and the inherent probabilistic nature of AI decision-making.
From Bad Bots to Good Bots
Historically, 99.9% of bot traffic has been malicious, designed explicitly for fraud and exploitation. Existing financial infrastructure is consequently optimized to combat bad bots, not embrace good ones. But AI agents flip this narrative. As headless browsers and automated agents proliferate, traditional authentication methods break down. Every time an AI developer describes their "headless browser," bank fraud teams cringe. How do we distinguish legitimate autonomous agents from malicious bots, especially when the agents themselves multiply exponentially and operate at superhuman speeds?
Traditional identity verification, KYC, and fraud detection processes were designed for human-led transactions, not agents operating independently. Who exactly are we authenticating - the developer, the AI model itself, or the ultimate beneficiary? Today's frameworks offer few answers.
Probabilistic vs. Deterministic Payments
AI operates on probabilistic reasoning to get outputs, where "close enough" often suffices. Yet payments demand absolute precision. A model hallucinating text might be harmless; hallucinating transaction details - a quantity, a decimal, a recipient - is catastrophic. Imagine your agent mistakenly orders one million units instead of one. Probabilistic errors in financial contexts simply aren't permissible, yet current payment rails lack robust verification mechanisms capable of interpreting and correcting probabilistic decision-making.
Unprecedented Scale & Frequency, and Variable Pricing
The human brain struggles to grasp exponential growth. Autonomous agents don't sleep - they operate in parallel, instantly spinning up infinite workflows, transactions, and interactions. Consider startups running thousands of autonomous MVP experiments simultaneously to optimize product-market fit, or personal agents autonomously negotiating purchases, services, and contracts around the clock.
This isn't incremental change - it's an order-of-magnitude leap. The traditional financial rails - FedACH, Visa, Mastercard - weren't built for instantaneous, 24/7, infinite scalability. Yet AI-driven commerce demands precisely this.
New infrastructure will likely be required, as it's important to move information alongside money - especially because the cost of completing a task can vary. For example, if you ask an agent to book a flight, the price is known in real time. But if a chain of agents is executing a workflow, costs can fluctuate based on which LLM is used, which freelance agent is hired, and other factors. Take the simple task of generating a logo - you'd want to confirm that the work has been delivered and meets expectations before releasing payment from one agent to the next.
What Should Agentic Payments Infrastructure Look Like?
The future payments system for agents will definitely run on traditional fiat rails for regulatory compliance, and may also include crypto elements such as programmable smart contracts and stablecoins. Ideally, these transactions should:
• Validate and verify completion of tasks autonomously before releasing payment, much like smart contracts today. • Enable instant, secure, compliant microtransactions at unprecedented scale. • Seamlessly handle variability in operational costs - such as dynamically priced compute resources or API calls from various AI models. • Have a way to handle liability that is baked into each step (as it is today in the current financial system) so it's clear who is responsible for a charge-back, return etc
Who's Already Building?
This market barely existed 12 months ago but is now bustling with innovation. Here's how the early movers are shaping it:
• @AtumLabs : Building a blockchain-based Visa-like network on Solana and Base for open agentic commerce. • @crossmint : A crypto platform for people to build onchain, including agents • @fewsats (Lightning Network): Leveraging Bitcoin's Lightning Network and the internet's original HTTP 402 payment protocol to facilitate microtransactions. • @Nevermined_io : Creating decentralized AI-native payment protocols tailored explicitly for autonomous workflows. • @PaymanAI *: AI compliant bank accounts and infrastructure - recently executed the first real, compliant autonomous bank transaction initiated by an AI agent. • @trySkyfire : Former Ripple team reimagining the financial stack for an AI-first economy.
People are also leveraging old financial infrastructure in new ways. I spoke to a stealth company that was leveraging traditional financial proxies for agent payments.
Why Won't @stripe Just Own This?
Yes, Stripe is a payments powerhouse with deep pockets. I mean - Stripe is worth $90+ Billion dollars and is capable of going after what they want, but agentic payments pose significant risks. The cost of prematurely enabling autonomous agents without solving authentication, fraud, and liability concerns is immense. Stripe's cautious moves into agentic payments - mainly through partnerships and virtual cards - highlight the risks inherent in rushing forward, and they have much more to lose if things go wrong on the regulatory front. Plus most of Stripe's true advantage is on the merchant side which doesn't necessarily translate here. Early startups can innovate faster and take more targeted bets without existential risk to legacy systems.
A Glimpse Into the Near Future
The coming era of agentic payments remains somewhat hazy, so here is blurry ball of predictions versus a crystal one:
•Hybrid & New Infrastructure: Whether it's existing FIAT or crypto infrastructure or a patchwork quilt of both, the rails and at least the concepts exist today to build this the way it should be built.
•New Merchant Interfaces: Websites may have dedicated checkout gateways designed explicitly for agent interaction, separate from human-facing systems.
•Fraud & Liability Reimagined: Teams ignoring fraud prevention, consumer protection, and clear liability definitions will fail spectacularly. Those proactively embedding these considerations will thrive.
•Humans to Automation: Short-term solutions will keep humans in the approval loop, but complete autonomy will rapidly become the norm. Agents will autonomously transact, settle, and reconcile transactions at a scale never before seen.
One thing is crystal clear - what wins will be the team that can get compliant, fast payments into the hands of as many developers as quickly as possible as they build out agents (ultimately getting it into the hands of businesses and consumer).
The Opportunity is Now
We're in the opening innings of an enormous transformation. AI, fintech, and crypto founders who move fast to solve these challenges will unlock vast new categories of commerce and reshape global financial infrastructure. VCs who back the right teams today will fund the financial backbone of tomorrow's autonomous economy.
If you're building, innovating, or investing in this explosive new frontier, we'd love to connect, exchange ideas, and chart the future together.
The agentic payments revolution has begun. The only question now: who's going to win the race?
- Disclosure - I'm an investor in Payman