AI agents can already write code, analyse data and execute workflows. Turning them into customers requires something more: they need to find a service, authenticate, understand its pricing and complete a transaction within defined limits.
For SaaS companies, API providers and data businesses, that creates a three-part readiness test: Can an agent access your product, can you charge for how it uses the product and can it pay you safely?
In a recent EUVC conversation hosted by Andreas Munk Holm, Viggo Stenseth, Co-Founder and CEO of SolvaPay, joined Redstone General Partners Samuli Sirén and Mickaël Bellaïche to discuss what payment infrastructure needs to change as AI agents become economic actors.
One signal suggests that shift may be arriving faster than expected. Viggo had assumed fully autonomous agent-to-agent transactions were still a few months away, until prospective customers began asking for them now.
Among those early use cases is a platform where creators build specialised datasets, dashboards and analytics tools. The company wants agents to spend autonomously across those services within its existing revenue-share model. It needs the transaction layer now.
That turns agentic payments from a distant infrastructure thesis into an immediate commercial question: What revenue are digital businesses missing because their products are accessible to software but still difficult for software to buy?
Machine access is only the first step
Agents increasingly consume APIs, data and specialist tools while completing larger tasks. Most digital sales infrastructure still assumes a person will choose a plan, enter payment details and approve each purchase.
Viggo argues that businesses need a parallel sales layer designed around how software buys. An agent should be able to discover a service, authenticate, test it and purchase additional access based on usage.
SolvaPay currently offers recurring, usage-based, one-time and hybrid pricing for APIs, web applications and Model Context Protocol (MCP) servers, alongside identity, tool-level access, agent credits and spending limits.
The bigger question is whether a business’s pricing, permissions and product architecture can support a customer that buys programmatically.
Redstone invested before traction could answer the question
SolvaPay recently raised €2.4 million in pre-seed funding, led by Redstone and MS&AD Ventures, with other investors joining.
For Mickaël, the investment case did not come from traction. SolvaPay was only a few months into its journey when Redstone invested. The conviction came from the missing transactional layer, the team and the number of business-to-business use cases that would eventually require agents to spend.
“There are no payment rails for the agentic world.”
Samuli places the opportunity within a familiar technology cycle. AI carries plenty of hype and overreaction, but he expects a substantial underlying shift to remain.
“It feels like a hype, but it’s quite a bit more than that.”
For Redstone, payments are particularly interesting because they add infrastructure and regulatory complexity that generic AI process automation may lack.
One integration needs to reach several agent ecosystems
Payments are only one source of fragmentation. AI providers, financial companies and infrastructure businesses are developing different protocols and marketplaces, often within their own ecosystems.
Viggo calls this the “battle of protocols”.
A business selling digital services will not want to rebuild its commercial infrastructure for every agent platform. SolvaPay’s bet is to become a connective layer across them, closer to a universal standard such as USB than a closed marketplace.
Rather than predicting which protocol will win, the goal is for businesses to connect once and sell across the ecosystems where agents operate.
Existing financial rails make the product usable now
SolvaPay is deliberately building on existing financial infrastructure. Viggo is not dismissive of crypto or stablecoins, but businesses still need payments to connect with the banking and financial systems they use today.
“It has to work today.”
That principle shapes the product. Stripe provides underlying payment infrastructure while SolvaPay adds the identity, billing, permissions and agent-facing transaction layer.
The aim is to let agents buy in new ways without requiring businesses to abandon existing financial rails.
Regulation could become part of the moat
That approach introduces a difficult operating model. SolvaPay must develop the product, work with customers and pursue regulatory approvals in parallel.
Viggo says relevant licensing processes can take 12 to 18 months. For a company founded by three technical co-founders, Viggo, Tommy Berglind and Ingemar Svensson, bringing in compliance expertise early allows that work to progress alongside product development.
Mickaël sees this complexity as part of the potential defensibility. Many AI-enabled workflows can be replicated by well-funded competitors or incumbents, while payment infrastructure and regulation create harder barriers to entry.
The agent-readiness test
For SaaS companies, API providers and data businesses, three questions matter:
Access: Can an agent discover your service, authenticate and use the right tools?
Pricing: Can you charge in a way that reflects how software consumes the product?
Payment: Can the agent transact within clear budgets, permissions and controls across the ecosystems where it operates?
A product that passes only the access test may be visible to agents without converting them into customers.
Turning agents into customers requires all three: discovery and access, pricing built for machine usage and payments within clear guardrails.
Listen to the full conversation below to pressure-test your agentic revenue layer.
