How Paddle consolidated its AI strategy and achieved 91% company-wide adoption

Hendrik Brackmann has built data functions at five UK and European fintechs. MarketFinance, Finiata, Tide, Stenn, and now Paddle, where he serves as VP of Data & AI. When you've seen that many companies from the inside, you start noticing patterns. The one that matters most for this story: every fast-growing B2B company eventually hits the same wall with AI tooling.

At Paddle, the Merchant of Record platform helping 10,000+ digital product companies operate and grow globally, that wall looked familiar: every department was already moving fast on AI adoption but was independently using their own tools. Sales had one. Customer Success had another. Engineering was experimenting with something else and so on. 

The catalyst at Paddle wasn’t a top-down executive mandate announced at an all-hands. Instead, it was a natural evolution of the company’s operating values, which actively encourages autonomous ownership, innovation and experimentation. Because of this, grassroots AI adoption exploded across the organization. Recognizing the need to move from a fragmented tool ecosystem to a centralized strategy, leadership came together to develop a company-wide AI strategy. 

Paddle built an internal AI strategy around three pillars: a champions network to drive adoption from within, a unified agent platform (they chose Dust), and governance to keep things coordinated. The framing was never "we need AI." It was "we need to consolidate, govern and accelerate what is already happening."

If you're a B2B SaaS company with 200, 300, 500 people, there's a good chance the same thing is quietly happening inside your org right now. Teams are solving their own problems with their own tools but there isn’t yet a centralized approach. 

Paddle didn't just roll out a platform and hope for the best. They went from no centralized AI tooling to 342 active users in under six months. 

But the numbers that really stand out are about depth, not breadth. By week eight, the vast majority of active users were doing high-value work: querying company data, building custom agents, and connecting external tools. Not just chatting with an LLM.

To put that in context: the typical enterprise customer at a similar stage has about 32 people actively building agents and around 170 agents in use. Paddle had 114 builders and over 470 agents. More than three times the average on both counts.

That kind of adoption doesn't happen by accident. Hendrik and the leadership team drove it deliberately. They organized a hands-on hackathon where leadership built workflows themselves, rather than delegating it to a task force. They ran internal events. They created a champions network where early adopters helped their teammates get started. And they pushed adoption across the majority of the company’s departments: 

The key insight is that leadership treated it as a fundamental operating shift, rather than an IT rollout. When leadership builds agents alongside its teams, adoption follows.

What surprised everyone was which capability drove the most organic adoption. It wasn't the chatbot. It wasn't search. It wasn't automation.

But when Hendrik talks about what really got people excited, he keeps coming back to one thing: data visualization.

Paddle is a data-heavy company. Financial services companies tend to be. The ability to pull numbers from a data warehouse, combine them with context from Slack or call recordings, and produce a visual output that a non-technical person can actually use? That spread faster than any chatbot or automation workflow.

Hendrik sees this as a pattern specific to the kind of company Paddle is: "Financial companies tend to be a lot more data-driven. Dust's ability to take in data and visualize it might be more relevant in fintech than other industries."

Here's the insight that surprised even Hendrik. When he looks at where Dust gets the most traction inside Paddle, it's not in payments-specific workflows. It's in the problems every B2B company shares.

The highest-impact use cases at Paddle are ones like merging quantitative data from their data warehouse with qualitative context from Slack and Gong. Giving account managers a richer picture of their customers. Helping sales teams prep faster and with better context. These aren't fintech workflows. They're B2B workflows.

Paddle did build a fintech-specific agent for payment acceptance root cause analysis, which Hendrik acknowledges is "a very financial topic." But that's a deep use case, not the one that drove broad adoption. The horizontal B2B value came first. The vertical, industry-specific stuff followed naturally.

Paddle has strong engineering. Go microservices, API-first architecture, they even shipped their own MCP server with 80+ tools. You might expect a company like that to say "we'll build it ourselves."

They didn't. As Hendrik says: 

“In a fast-growing fintech, the team’s time and focus is really critical to how we operate. By partnering with a world-class platform like Dust, we gain an enterprise AI platform that helps everyone work smarter and allows us to focus on our customers’ needs and building the best product.” 

There was no meaningful engineering resistance at Paddle. The reasoning was simple: building and maintaining an internal AI platform pulls engineers away from the product that actually generates revenue. And the IP concern that sometimes blocks "buy" decisions didn't apply here. As Hendrik put it: "I don't think Dust would typically fall into this category of sensitive IP we'd be outsourcing."

If you're running a B2B company and debating whether to build an internal AI layer or adopt a platform, Hendrik's advice from five fintech data teams is clear: partner, and redirect your teams’ time to what makes your product different. 

If you're wondering whether your company is heading toward the same moment Paddle hit, Hendrik's assessment is practical. The trigger isn't a CEO posting about AI strategy on LinkedIn. It's not a "Head of AI" job opening. Those are lagging indicators.

The leading signal is when there is clear use and momentum but departments are each using a different set of tools. That cost-consolidation moment is when the CTO or VP of Data gets the mandate to unify. At Paddle, it happened organically. At your company, it might be happening right now.

Paddle’s story isn’t unique to payments; it’s the blueprint for any fast-growing B2B organization reaching AI maturity. It’s the story of a company that recognized organic enthusiasm for AI across its departments and chose to build on that momentum by creating a company-wide strategy. 

The playbook that worked for them:

Channel momentum, don't restrict it: Instead of pushing an abstract corporate AI mandate, look at where your teams are already experimenting and give them a single, unified environment to do it securely.

Integrate with your existing ecosystem: Adopt a platform like Dust that seamlessly connects to the tools your teams already live in. (Slack, Notion, Salesforce, and your data warehouse.)

Lead from the front: Treat the transition as an operational evolution, not an IT rollout. When leadership actively builds alongside their teams, high-value adoption follows.

Don’t distract engineering resources: Partner with a specialized platform for your AI infrastructure so your internal team can stay focused on what you do best. 

By shifting from fragmented tool sprawl to a centralized AI strategy, Paddle achieved 342 monthly active users, 114 active agent builders, and over 470 custom agents running across the business, all within six months. The result is an organization where 91% of employees don't just use AI as a novelty, but rely on it as core operational infrastructure. 

Often the most powerful AI strategy doesn’t come from the top-down, it’s seeing where your employees already have momentum, and giving them the platform to run with it.