While most enterprises are talking about how they approach AI transformation, increasingly knowledge management and arranging the required context are important factors in whether or not AI can work effectively. This shift is creating demand for a new role called the Knowledge Architect.
We have Systems Architects to design the overall structure of software and hardware systems. Knowledge workers execute on work across different functions like Sales, Marketing and Engineering to achieve business goals. But the concept of a Knowledge Architect is new, and perhaps what’s needed to untangle the complexity of an organization’s documentation and enable better collaboration across functions.
A Knowledge Architect is the bridge between the history of what an organization has documented, how it makes decisions today, and how it should operate in an AI-augmented world. Put simply, it’s the role that architects the conditions for AI to do useful work. It could be the difference between a successful or unsuccessful AI transformation. Measuring success in your organization should not be about how much AI is deployed, but instead by how effectively the workforce is transformed.
At Atlassian, we’ve been thinking about where this role can add most value. Here’s how it could look.
What is a Knowledge Architect?
The purpose of this role is really two-fold:
- Architect the systems and frameworks that enable AI tools to draw on organizational knowledge effectively.
- Upskill team members to leverage the tools and systems effectively across their functional areas.
The concept draws on the structural rigor of a Systems Architect and leverages the documentation of knowledge workers within the business, while introducing a new dimension to adapt for AI transformation: context engineering.
Context engineering is the process of understanding how to structure, curate, and deliver the right information so that both humans and AI agents can make high-quality decisions. It requires thinking deeply about what context is lost in handoffs, what context AI needs to perform well, and what context humans need to trust AI outputs.
A key distinction to help distinguish the role is that Systems Architects design how technology scales, while Knowledge Architects design how context scales across the business. For example, when a marketing team’s AI agent produces inaccurate campaign briefs because it can’t find the brand guidelines buried in a shared drive, a Knowledge Architect designs the system that solves that problem permanently.
The five core responsibilities of a Knowledge Architect
1. Design the flow of context across teams
Ultimately, the role is tasked with designing how knowledge moves between people, teams, AI agents, and systems. They should map the organization as a knowledge system that is constantly moving and adapting, and identify where knowledge is siloed, fragile, or invisible. The goal is to structure, curate, and deliver the right information to the right teams, so both humans and AI agents can make good decisions.
2. Define the rules of engagement between humans and AI agents
The role helps develop guidelines and assess which tasks suit human judgment, which suit AI execution, and which require a collaborative loop. This may also include surfacing opportunities for role imagination by identifying where knowledge workers can move up the value chain as AI absorbs routine cognitive tasks.
3. Upskill knowledge workers on context management
A crucial part of the role is to build AI fluency at scale, but also to enable knowledge workers to embed effective documentation into their ways of working to help architect AI-accessible context. The role also plays a role in educating the team to address any fears, build safe spaces for experimentation, and emphasize that AI tools augment rather than replace ways of working.
4. Build knowledge infrastructure
Developing the operational systems that power knowledge management, such as taxonomies, ontologies, metadata standards, and content governance frameworks. Building a working understanding across the organization of AI ecosystems, such as LLMs, agents, and knowledge graphs to architect how they integrate with human workflows.
5. Steward responsible AI use and build trust
Build the feedback loops and validation mechanisms that ensure work meets organizational standards, accuracy and accountability. Ensure trust is at the centre of all AI use across the organization and create a Responsible AI framework to provide clear guidelines. Some examples of this include the following principles:
- Transparency: Can you explain why AI made a decision? If not, it shouldn’t make that decision autonomously. This might include showing inputs, confidence, and allowing access to auditable logs tied to the model version
- Guardrails: What hard limits exist? (eg. AI should never auto-closes a high-risk task; AI should never send external communications without human review)
- Accountability: When AI is wrong or impacts production workloads, who is responsible for it? Clear ownership prevents scapegoating the algorithm.
Knowledge Architect should also establish mechanisms and incentives to keep knowledge and context updated and relevant to reinforce trust for data interpreted by AI.
The qualities that make a good Knowledge Architect
As AI technology is rapidly evolving and documentation needs across your business will likely change, being adaptable is an important quality for this role. It is best suited to systems thinkers who have a knack for operationalizing strategies. The key skills to look for include:
- Cross-functional translation: They should be excellent at cross-functional stakeholder management and translating the needs of engineers, executives and end-users into working systems.
- Future-thinking: Designing for what’s next, not just what’s now. An organization’s knowledge needs to be maintained and updated for its proper functioning.
- AI literacy: They possess a working understanding of AI architectures including LLMs, retrieval-augmented generation, agent frameworks, knowledge graphs, embeddings, and vector systems. This is to architect how they integrate with human workflows and organizational knowledge.
- Context engineering: Understanding how to structure, curate, and deliver the right information so that both humans and AI agents can make high-quality decisions.
How this role scales across an organization
This is a horizontal role by design – intended to connect lots of different verticals across the business. Across smaller organizations, just one Knowledge Architect could serve as a connective tissue across the business. In larger enterprises, the function scales into a small, distributed team: a central Knowledge Architect who owns the organization-wide standards, governance frameworks, and context layer, and domain-embedded counterparts sitting within major functions like Engineering, Sales, Marketing, and Customer Support, who translate those standards into the realities of their domain. I picture them as mid-senior IC roles, with room for growth.
The value this team drives shows up where executives care most: increased ROI. When embedded effectively, the work of Knowledge Architects should result in better, more reliable AI outputs, faster adoption from other employees due to increased trust, resilient institutional knowledge systems as your business grows, and better performing AI investment.
We know that enterprises are spending millions on AI infrastructure without thinking about their knowledge management solutions – and therein lies the risk. The companies that move fastest to address this challenge will produce better AI outputs, build trust faster and successfully embed AI tools across their business.
The way forward
AI models are no longer the moat for your business – context is.
Knowledge Architects are a role-in-the-making which will likely evolve over time. It’s all about starting the conversation about role design as organizations adopt AI tech.
What I would encourage you to think about, is if you already have someone in your organization doing this work, acknowledge them and resource them. And if you don’t, consider creating the opportunity. These are the people your AI strategy depends on.
