How Payabli Unified Their Developer Experience with AI-Powered Documentation

TL;DR

Payabli enables SaaS platforms to embed, manage, and monetize payments via a single, unified API.

1. Payabli needed to scale the process of surfacing their documentation content to many places as more users onboarded on to the platform.
2. To simplify how users found new technical details of their product, the Payabli documentation team looked to Inkeep's AI assistant to surface answers from various sources while embedding it in multiple website locations.
3. Payabli also used Inkeep's "Keep" copilot directly inside their ticketing system, cutting down on repetitive questions and speeding up responses.
4. In less than two days post-rollout, new help content created from support feedback was already resolving real customer issues, proving fast time-to-value.

Problem: When "AI-Powered" Isn't Intelligent

In late 2023, Payabli faced a challenge familiar to many developer-focused companies. Their prior documentation platform needed to scale as more users onboarded to use their platform.

"There's gotta be something better out here" recalled Casey Smith, Payabli's Head of Documentation. That's because Casey knew that the conversation nature of AI chat can help enhance their customer's search experience through a chat assistant.

But the problems ran deeper than search:

• Repetitive Customer Questions: Support reps were answering the same questions repeatedly without linking to documentation
• Manual code generation: Teams were hand-crafting JSON configuration examples for demos and integrations
• Siloed knowledge: Support agents had to leave their ticketing system to search documentation, adding friction to every interaction
• Language barriers: Payabli's partners had nearshore and offshore engineers who needed documentation in Spanish and Ukrainian, but their docs weren't localized

Solution: One AI Experience Across Every Touchpoint

Casey and the Payabli team wanted to build an experience that gave their users instant access to documentation information, while enabling their support team to handle more complex questions.

When Casey discovered Inkeep, the difference provided in value was immediate. That's because Inkeep's suite of products enables Payabli to:

1. Generate Instant Code from Configuration Docs: Users simply ask Inkeep chat to "spin me up a config example for these parameters?" and instantly get working code.
2. Integrate AI in Support Workflow: The support team now uses Inkeep's copilot (called "Keep") directly within their ticketing system. "They don't even have to leave where they work on tickets, which is phenomenal, 'cause those seconds do add up."
3. Provide a Unified Experience Everywhere: From Slack channels to support tickets to documentation sites, users get the same AI experience. This consistency proved crucial – Casey compares it to the jarring experience of switching between Claude and ChatGPT: "I instantly notice that it's a different vibe... I have to prompt it differently."
4. Discover Content Gaps: Payabli's documentation team can now see details on "about top queries and what people clicked on". This helps Payabli's "catch doc gaps" & thereby prioritize content that users actually need based on their query behaviour.
5. Not worry about translating docs: Engineers can ask questions in Spanish, Ukrainian, Vietnamese and French and receive answers that "don't read like Google Translate." Inkeep understands context and technical terminology across languages.

Internally, Payabli implemented Inkeep and branded it internally as "Docabli" across their entire ecosystem.

Results: From Implementation to Impact in Days

Payabli's adoption of Inkeep yielded an 80% deflection rate, outperforming industry benchmarks at 50-60%.

The rollout followed a strategic path:

1. Started with documentation: Integrated Docabli into their existing docs
2. Expanded to Slack: Added to internal channels first, then gradually to partner channels
3. Integrated support copilot: Brought AI directly into the support team's workflow
4. Created feedback loop: Support team tags tickets where FAQs would be useful

The impact was nearly instantaneous.

Within 2 days of creating a new knowledge article based on support feedback, Casey received a Slack message: "Hey, look how great Keep Copilot… a customer asked this question," and the newly created content had already helped resolve a customer issue.

Today, a quarter of Payabli's customer support conversations now use Keep to accelerate the support ticketing process, thereby getting quick service to their customer experience.

Expanded Adoption: Inkeep MCP to Safely Integrate Payabli Docs

As Payabli's partnership with Inkeep evolved, they adopted Inkeep's Model Context Protocol (MCP) server. Inkeep's MCP enables Payabli's users to access Payabli's documentation directly within development environments like Cursor, Windsurf, Claude Desktop, and Claude Code.

The immediate benefit of this is that it enables developers to query Payabli's documentation without leaving their IDE. Thereby reducing context switching and saving time.

But the true benefit of this lies in security.

"We love Inkeep because it knows everything about Payabli but is completely airgapped from our actual applications," explains Casey. "Through Inkeep's MCP, we are able to safely release AI integration assistance without exposing ourselves to security vulnerabilities."

Payabli's MCP server works exclusively with their public-facing documentation, creating a secure boundary between AI capabilities and production systems. This approach eliminates the need to build custom APIs or risk exposing sensitive application data.

The result? Payabli's partners get an intelligent & context-aware assistance for integration work, while Payabli maintains complete security isolation.

As Casey notes, "Inkeep makes us feel comfortable doing AI because we're building using only our public sources."

Key Takeaways for Developer-First Teams

1. Start Where Your Users Already Are: "We added it to a couple internal channels, and then we started slowly rolling it out to external channels." Don't force new workflows – integrate AI into existing ones.
2. Code Generation Is a Game-Changer: With well-defined config specs for configuration docs for embedded components, your AI Assistant can also become a code assistant. Payabli's teams use Docably to generate configuration examples on-demand, eliminating manual JSON manipulation.
3. Consistency Matters More Than Features: "Our users, both internal and external, get a consistent search and chat experience across the board... I don't think a lot of people realize that consistency in user experience is gonna be so important."
4. Human-in-the-Loop Is Non-Negotiable: "I will never use a tool that just allows people to introduce garbage into my system... AI is amazing. However, even Inkeep I want a human in the loop."
5. Think Beyond Deflection: While ticket deflection is the "big goal," the immediate wins come from:
   1. Empowering support teams with instant answers
   2. Enabling sales teams with demo materials
   3. Breaking down barriers (e.g. language)
   4. Creating a feedback loop between support & documentation teams

Looking Forward: The Future of AI-Augmented Documentation

Casey's vision extends beyond current implementations. The next milestones include:

• Automated ticket deflection with intelligent routing
• Support self-service where agents can contribute to the knowledge base directly

For technical writers and documentation teams, Casey offers this perspective: "You need to have a good command of these tools. You need to understand how to write for AI or you're gonna be unemployable, because like it or not, even the enterprise companies are moving in this direction."

The lesson from Payabli's journey is clear: When AI actually works, adoption happens naturally. Teams find unexpected use cases, workflows improve organically, and the technology becomes indispensable.

As Casey puts it: "I've got Inkeep stuck in everything now."