How to Build Powerful Internal AI Productivity Tools
🧠Internal tooling is the new AI battleground. Here’s some ideas on how to build your own - using Ramp's impressive new product and the latest technologies as inspiration.
🔒 The Knowledge Series breaks down emerging AI technologies with practical playbooks designed specifically for product teams. Get 100+ guides and practical tutorials covering everything from Claude Code and MCP to agentic workflows, vibe coding, and more.
Historically, working on internal tooling wasn’t particularly glamorous - and the product teams who worked on these tools would often itch to get back to work on customer-facing features.
But during the AI boom this attitude seems to be changing.
Whether or not you believe in the so-called SaaSpocalypse, there is a new, emerging role amongst some of the most AI-progressive companies: the ‘Internal AI’ specialist:
Companies like OpenAI, Figma, Replit and others have either hired or are actively hiring internal AI tooling specialists and this week, fintech startup Ramp’s own internal tooling specialist Seb Goddjin explained how they built their own internal productivity suite.
Based on the current trends, it looks as though as it becomes easier for companies to build their own tools, they’ll hire people to do it - and then use these tools as a flex for their AI excellence. Rightly or wrongly, relying on third party SaaS products across your entire internal operations could be seen as a sign of weakness and AI immaturity vs building your own.
In this Knowledge Series, we’ll take a closer look at what exactly Ramp’s internal tool does, the technologies it uses, and how you can take some of these principles to build out your own set of powerful internal tools designed to make your product development process faster and more efficient.
Coming up
What Ramp built and how it works
Key principles and takeaways for product teams to think about
Ideas on how you might be able to apply these new internal AI development principles in your own product:
Examples from verticals including SaaS, fintech and consumer products you can use as inspiration
What Ramp built
Here’s a snapshot of the new tool that Ramp’s internal AI teams built:
The product is called Glass and it is designed to make AI adoption and collaboration easier between employees easier to boost their productivity.
They found that while initial adoption of traditional AI tooling was high, many people were confused by the different workflows and technologies involved to get up and running properly. Glass was built to address that problem.
The core pieces unpacked
The internal tool includes 5 core capabilities:
Auto-configured single sign on and integrations
A Skills marketplace called Dojo to re-use custom built Skills
A Skill guide called Sensei which which looks at which tools you've connected, your role, and what you've been working on, then recommends the skills most likely to be useful
Persistent memory that lets users pick up from where they left off
Scheduled automations that run locally from a users’s laptop
How these pieces work together to boost productivity
Glass eliminates the usual setup friction by coming pre-configured on install. Employees sign in once via Okta SSO and every tool they need, from Gong and Salesforce to internal Ramp products, is immediately available. From there, a persistent memory system builds automatically based on those authenticated connections, giving every session full context on colleagues, active projects, Slack channels, Notion docs and Linear tickets.
On top of this foundation sits Dojo, a Skill marketplace where employees package their best workflows into reusable markdown files that teach the agent how to perform specific tasks.
Over 350 Skills have been shared company-wide and to prevent the catalog from becoming overwhelming, Glass includes an AI guide called Sensei, which looks at your role, connected tools and recent activity to surface the handful of skills most relevant to you. A new product manager or designer never has to browse 350 options; instead the right skills find them. Finally, Glass lets scheduled automations run on daily, weekly or custom schedules, posting results directly to Slack, so work that previously required an engineer can now be set up by anyone in an afternoon.
The strategic impact of building and owning these types of AI tools
The functionality of Glass is impressive but Ramp says that owning this infrastructure has a strategic impact, too.
They argue that Glass is a competitive advantage because internal productivity and speed is a moat and, perhaps most importantly for product teams, building these types of internal products also informs the customer-facing part of their business:
How to use the core principles from this to build your own internal tools that make your product development processes easier
Reading a case study is one thing, but now let’s dig a little deeper into some of the core principles and takeaways from this case study and explore how you might be able to use this to build your own internal tool set.
For this, we’ll analyze some of the technologies that can be used by product teams on internal tooling and then consider ideas for how you can build internal tools across 3 different product types:
B2B SaaS
Fintech
Consumer