The classic knowledge base
Have you ever built a knowledge base for your company or “a company brain”? W ith AI that idea is getting new life and taking a new form.
The old idea of a knowledge base is a potential solution to some of the challenges all businesses face:
- people have their knowledge and intuition stored in their heads.
- when a person leaves, most often their knowledge leaves with them.
- we’d like people to share their learnings and skills across the organisation to help everyone become more effective.
The classic solution has been to build an internal knowledge base and encourage our teams to search through the database when they need guidance on how to solve a task.
But this is changing.
Why the role of the knowledge base changes with AI
Now, AI enables our teams to benefit from internal knowledge in a much more active and high value way.
What does that mean?
Take the example of a sales team responding to an RFP (request for proposal). Simply put, an RFP is where a buyer is looking for a solution and asks a handful of potential providers to respond with structured information to make the buying decision easier.
In practice, the sales team receives a template with 100 questions about their company and product they need to fill out. This can take ages to complete but is essentially a task to collect relevant information about the business and communicate it well in a structured output.
Traditionally, many sales teams operate from their Sales Playbook, stored somewhere within the internal company knowledge base. The sales person would review the playbook for how we handle RFPs and get started.
However, today this process can be done much differently. In a way that’s more effective and where the company’s internal knowledge is used more productively.
How so?
Well, first you’d give Claude the request template and ask it to fill it out based on the previous RFPs you have responded to, stored somewhere in your files. You can add additional context about your business e.g. from your codebase or website. You could use the “How to respond to RFPs” guide from your knowledge base as a pre-defined skill for how Claude should draft your response. Same with your branding guidelines and tone of voice to make it look and sound right. You could add connectors e.g. to your CRM that pulls in data about the prospect.
You will have a first draft in 10 minutes.
And importantly, if this is set up right, the draft would largely reflect your internal best practices for handling RFPs, and your company’s branding and tone of voice.
However, that requires setting up the infrastructure for how your AI system can access your tools, context, “how to guides”, etc.
And that’s the opportunity here: Structure your internal knowledge and context, build and store specific “skills” (or, “how to do x”), make them shareable, integrate your third party tools, and make it all available in the AI your company uses to make it accessible in your teams daily work.
Finally, and importantly, this new form of knowledge base is much more accessible because rather than relying on the user stepping out of their way to search the database, it’s now embedded directly in the now familiar chat interface that many people rely on in their day to day. It’s already where they are.
What it means for business
So this sounds quite wild. But what’s the benefit?
In short, it means the old idea of the knowledge base is actually coming to life. If executed well this could be a major productivity unlock for the organisation. And it compounds. Perhaps can even become a competitive advantage.
A bunch of companies are experimenting with this. Some public examples you can look into are Ramp. Coinbase. Block. Shopify.
But this is not something that’s only relevant to strict tech companies. We believe most knowledge work to some extent can be supported this way.
We will learn how valuable the knowledge in the company really is.