A Lightweight Knowledge Management Approach for Policy Teams

A brain alight with ideas. Image generated by Midjourney

Resource-constrained policy teams are often tempted to deprioritize Knowledge Management (KM) efforts. “There’s too much work and never enough time” is a lament that every policy team can relate to. I would argue, however, that the more resource-constrained a team is, the more thoughtful and targeted it should be about managing the team’s knowledge.

How do we reap the benefits of Knowledge Management without being unduly burdened by excessive documentation?

How are these four types of artifacts used?

How does a team create and maintain these artifacts?

What information do we track about each artifact?

How do we keep the knowledge base organized?

What kind of metrics should we track?

How can my team use Generative AI to support our KM efforts?

What can a team’s leadership do to encourage effective knowledge management practices?

Consider what happens when policy team members are left to figure out knowledge management on their own. The team spends an inordinate amount of time searching fruitlessly for reference material, comparing versions, or recreating them. Each person maintains their own set of bookmarks as they stumble across useful references. Pockets of KM efforts may emerge organically, but each exists in isolation, resulting in duplication of effort and the proliferation of inconsistent practices. Knowledge transfer between team members is laborious, and staff turnover is highly disruptive and chaotic. In extreme situations, even a basic inventory of existing and planned policies and their current status can take weeks to produce.

In contrast, investing in knowledge management is akin to pausing regularly to sharpen your saw when you need to cut a lot of wood. A lightweight and targeted KM approach1 delivers multiple benefits—it saves time, reduces errors or omissions that lead to rework, improves transparency, encourages a culture of continuous learning, and improves the consistency and quality of the team’s deliverables—without wasting the team’s time and energy.

More importantly, KM efforts give the team’s leaders visibility into the state of the team’s documented collective competencies. This visibility allows leaders to identify gaps and make thoughtful decisions about where to allocate resources.

How do we reap the benefits of Knowledge Management without being unduly burdened by excessive documentation?

An effective way to organize knowledge on a policy team without unnecessary overhead is to focus the team’s KM efforts on distilling and capturing knowledge into four types of artifacts: Templates, Checklists, Playbooks, and Acceptance Criteria.

An artifact is a tangible or digital work product that embodies knowledge. These artifacts are created, updated, or used while performing work—typically as inputs to guide how work should be done or as outputs that serve as evidence of the work that was performed.

While most people have an intuitive understanding of the four types of artifacts, here are quick definitions:

Templates. Pre-made, pre-formatted artifacts that serve as the starting point for creating documents, decks, spreadsheets, diagrams, and so on. Templates provide a standard layout and structure, a consistent look and feel, and placeholders for filling in specific information.

Acceptance Criteria. Conditions that must be met or satisfied by a work product or deliverable before it is deemed acceptable. These acceptance criteria may be internal (e.g., what a manager or policy lead uses to review the work of a direct report) or external (e.g., what a customer uses to assess the quality of a product or service).

Checklists. Written guides that enumerate the key steps in any complex procedure. A procedure can be considered complex if it involves multiple steps, has interdependencies, or requires specialized knowledge that makes it challenging for an individual to consistently perform just from memory alone. Checklists minimize errors, omissions, and inconsistencies. Where applicable, a Checklist will prescribe the use of specific Templates to produce work products and the Acceptance Criteria for reviewing them.

Playbooks. Comprehensive guides (sometimes depicted as process maps) that outline the organization’s processes, practices, and strategies. A Playbook helps its readers make decisions by specifying the Checklists that are appropriate for different needs and contexts.

Note: Your team or organization may prefer to use a different set of terms for these types of artifacts. The exact term used doesn’t matter as much as having a shared and well-documented definition for each type.

How are these four types of artifacts used?

Let’s say you want to standardize how your team responds to crisis events. In this scenario, we would use a combination of these four types of artifacts in this manner (see also Table 1 below):

A Crisis Management Playbook will direct the team to an initial set of Checklists to (a) assess if the situation meets the company’s definition of a crisis, and (b) determine the nature of the current incident.

Once a crisis has been declared and the nature of the crisis has been determined, the Playbook will direct the team to additional Checklists that specify what the team must do to respond to the crisis. For example:

A Crisis Stakeholders Checklist may specify which stakeholders should be notified and how they should be informed once the nature of the crisis has been determined.
A Rapid Response Team Checklist may specify the exact combination of crisis management roles that should be activated as well as how they should be activated.
A Rapid Response Action Checklist may detail the step-by-step procedures for each role to follow.

One or more of these Checklists may require the use of specific Templates. For example:

A Stakeholder Update Email template may prescribe the structure and format of emails the crisis team sends to senior management to keep them updated on the company’s crisis response.
A Crisis Response Log template may outline how and where response actions should be logged for future review.
An Impact Assessment Report template may outline how the impact of the crisis on stakeholders and systems should be documented.

One or more work products may have Acceptance Criteria to evaluate their quality. The criteria are conditional statements that the work product must satisfy before it can be considered acceptable. For example, the Impact Assessment Report may have the following as acceptance criteria:

All types of stakeholders must be explicitly identified, including customers, employees, partners, investors, and regulators.
The impact assessment for each stakeholder must cover the following areas: financial, reputational, legal, and operational.
The report must offer senior management a set of recommended next steps (e.g., stakeholder engagement actions and a communication plan) designed to mitigate the impact of the crisis.

When properly maintained and linked, these four types of artifacts improve the team’s ability to avoid mistakes and perform work consistently, produce work products of predictable quality, and help the team achieve desired business outcomes.

Table 1: Examples of Crisis Management Artifacts

Crisis Management Artifacts

How does a team create and maintain these artifacts?

The team can use the approach shown in Figure 1, which illustrates how these four types of artifacts fit into the team’s workflow.

The diagram is meant to be read from left to right. On the left, we see the four types of artifacts and how they are related. Most teams start by creating templates and checklists first (as denoted by their purple color in the diagram) since these artifacts directly and immediately benefit ongoing work. Acceptance Criteria typically follow later, and Playbooks only become necessary when multiple checklists exist for different scenarios.

Figure 1: A Lightweight Knowledge Management Approach

The following points are worth noting:

The use of distinct artifact types encourages planned, piecemeal additions. We avoid monolithic and unwieldy documents by dividing the documentation across Playbooks, Templates, Checklists, and Acceptance Criteria.

Knowledge management is not a one-and-done task. Expect to regularly update the artifacts to reflect changes to tools, workflows, and roles and to incorporate lessons learned through experience. While a flurry of rapid updates is expected when you first start out, the frequency of updates will eventually slow down, and the team will likely settle into a quarterly update cadence—frequent enough to be useful and seldom enough to not be disruptive.

These artifacts can be applied at different levels of granularity. KM artifacts are useful for a variety of activities, ranging from a single task (e.g., writing a policy), to a project that involves multiple tasks (e.g., performing a policy audit), to incidents with no pre-planned scope or length (e.g., a crisis event), and even broader still to a cross-functional initiative (e.g., launching a new safety feature).

Not all activities require KM artifacts. What are signs that a KM artifact will help?

Consider creating a template when different people produce the same type of work product but have outputs that differ in content or structure for no valid reason.
Create a checklist when you have to repeatedly explain the steps for an activity to multiple people at different times or when people forget mandatory steps or perform steps in the wrong sequence.
Acceptance Criteria are useful when (a) you have people applying different rubrics to assess similar deliverables and wish to standardize them, or (b) when setting expectations with a direct report when you first assign them to work on a deliverable.
Playbooks are useful when the team needs guidance on what combination of checklists, templates, or acceptance criteria they should use for a specific business situation or scenario.

These artifacts serve both the current team and future hires. By creating and maintaining these artifacts, you promote more consistent work performance and deliverables from your current team members and enhance your team’s onboarding materials for new hires.

Onboarding is incomplete if a new hire does not contribute to at least one of these artifacts. The onboarding process for each new hire must include a task where the new hire is expected to contribute an idea or suggestion to improve at least one of these artifacts, preferably within their first 60 days on the job. This assignment serves multiple purposes: (a) it confirms that the new hire has access to the knowledge base, (b) the new hire learns how to search the knowledge base for existing artifacts, and (c) it reinforces the message that everyone on the team is expected to contribute to artifact improvements.

What information do we track about each artifact?

Expect to track the following metadata for each artifact.

Title. A short title that describes the contents of the artifact.

Purpose. Specifies who the artifact is for and when it should be used.

Version History / Change Log. Provides data on when the artifact was created and when it was subsequently updated, with a description of what has changed with each update. Ideally, you will store copies or snapshots of each version.

Owner. Identifies the individual or team responsible for maintaining the artifact. They review suggestions for improvement, decide what changes are worth implementing, and ensure that the necessary approvals are obtained before changes are made.

Categories and Tags/Keywords. Used to organize the knowledge base and make the artifacts searchable.

Related Artifacts. Links to other artifacts referenced by or that make reference to this one. For example, a Policy Development checklist (see example in the gray box below) would be expected to link to a Policy Template and a set of Policy Acceptance Criteria.

Ideas for Future Updates. A mechanism to collect suggestions to improve the artifact. Create a way for team members to add notes to a “Future Updates” section of the artifact (if supported by the tool) or include a discreet link to a feedback form where people can submit ideas for the artifact owner to consider when they work on a new update.

The ease of tracking this information depends on the knowledge management tools used by the team. Some platforms offer built-in version control and commenting features (e.g., Confluence, Notion), which simplify tracking metadata. Take the time to understand the capabilities of your current tools/systems and adopt the necessary practices to use them effectively.

Example: A Policy Development Checklist

The images below show portions of this Evidence-Based Policy Development Checklist and are used with the permission of Del Harvey.

Policy writers can use the checklist to ensure all mandatory steps in the policy development process are completed. Specific tasks can be assigned to different people working together. The %Completed figure in the header automatically updates as each task is checked off. Checklist users are expected to replace any [placeholder text] that appears in square brackets.

Metadata about the checklist is tracked in a separate ReadMe tab.

Partial screenshot of Del Harvey's Evidence-based Policy Development Checklist

How do we keep the knowledge base organized?

We’ve all seen intranets or wikis that are a mishmash of outdated pages. When different people own different artifacts in the knowledge base, there is a real risk that the knowledge repository will devolve into a disorganized mess of conflicting and outdated information.

Thus, a Knowledge Base Manager must consistently guide the process of documenting, organizing, and maintaining the artifacts as a cohesive whole, even if they don’t own the individual pieces.

A Knowledge Base Manager must:

Define and document the categories, tags, and metadata fields that all artifacts will use. They work with artifact owners to ensure that these are consistently updated as the contents of the knowledge base evolve.

Maintain the conceptual integrity of the knowledge base by monitoring its state (e.g., what’s been recently updated and what hasn’t been updated for quite some time) and working with artifact owners to resolve any inconsistencies, overlaps, duplications, and gaps.

Provide guidance and answer questions about adding to, searching through, and updating artifacts in the knowledge base. Where needed, they help team members structure early drafts of new artifacts and give feedback to improve their quality and usability.

Provide regular updates about knowledge base changes, including changes to categories and tags, as well as an easy-to-reference list of recent artifact additions and updates.

Track and report knowledge base metrics to assess the team’s overall KM efforts (see the Metrics section below for more.)

Ideally, the person serving as the Knowledge Base Manager has the Scribe persona (as seen in Policy Personas). We define the Scribe persona as someone who is “adept at recognizing, documenting, and organizing tacit knowledge—information about how we work or how decisions are made here—that is typically acquired through years of practice and hands-on experience.” Most policy teams will have at least one person who derives satisfaction from this work and enjoys bringing order to chaos. They are likely to become your Scribe by default.

What kind of metrics should we track?

Depending on the system or tool used as the knowledge repository, some metrics may be easier to track than others.

Consider starting with a few metrics and evaluating if the effort to collect the data is justified by the types of decisions they enable about the knowledge base and the team’s efforts to maintain it.

Some types of metrics to consider:

Artifact Volume	Number of artifacts in the knowledge base as of this periodNumber of artifacts identified (existing + planned) as of this period, with a %complete figure
Knowledge base Updates	Number of artifacts added, updated, and retired this periodMost actively updated category or tag this period
Knowledge base Contributors	Number of unique knowledge base contributors this periodMost active knowledge base contributors this period4New knowledge base contributors this periodKnowledge base contributors who became inactive this period (where inactivity means no updates have been made over a pre-defined timeframe)
Knowledge base Access/Views	Number of artifact views or downloads per artifact this periodTrends in knowledge base views over timeMost frequently accessed artifacts this period
Artifact Feedback	Number and type of entries added to “Future Updates” this periodArtifacts with the most entries added to “Future Updates”
Search Metrics	Number of abandoned searches (i.e., searches where the user did not click on any of the links in the search results or subsequently refined their search criteria)Search keywords that yielded no results

Table 2: Candidate KM Metrics

Note: While it’s tempting to collect as many metrics as possible, it’s worth remembering that metrics are only useful if they help you decide where to invest resources in the team’s knowledge management efforts.

How can my team use Generative AI to support our KM efforts?

Because the artifacts are distinct but related pieces of knowledge, we can use Generative Artificial Intelligence (GenAI) to improve them individually and use one artifact to review another.

For example, GenAI can help you:

Enhance a Template: Upload the latest version of a template and its metadata to an AI service, then ask the AI to suggest improvements so the template better achieves its purpose.

Improve a Checklist: Upload a checklist, then ask the AI to (1) flag jargon or terms that may be confusing, (2) identify ambiguously worded instructions, (3) suggest edits to improve the checklist, and (4) propose additional checklists and templates that may be useful.

Suggest an initial set of Acceptance Criteria: Upload a well-written work product and ask the AI to derive an initial set of acceptance criteria for this type of deliverable based on the example provided.

Identify more Playbook scenarios. Upload a Playbook and ask the AI to identify gaps, i.e., scenarios that are likely to occur but are not yet covered by the Playbook.

Perform a first-pass review of a work product: Upload the work product and its related acceptance criteria and ask the AI to highlight any areas where the work product fails to meet the Acceptance Criteria. Ask the AI for suggestions to improve the work product. Note: this step does not replace actual human review by a manager or a policy lead, but it can supplement the review process.

Enhance categorization and tagging. Upload a set of related artifacts and ask the AI to recommend categories and tags for each to make them easier to find.

Perform consistency checks: Upload all artifacts related to a specific process or deliverable, then ask the AI to flag any inconsistencies in terms, steps, roles, or guidance that appear across artifacts.

Summarize suggested updates. Upload all submitted ideas for future updates for a related set of artifacts and ask the AI to group the feedback into themes.

Look for ways Generative AI can reduce manual effort, identify patterns, spot inconsistencies, and enhance the completeness and usability of the artifacts. The goal is to make the artifacts as useful as possible without spending more time than necessary to maintain them.

What can a team’s leadership do to encourage effective knowledge management practices?

To encourage effective knowledge management practices, consider the following actions:

Designate a central KM repository. Mandate the use of a system or tool (be it a shared drive, an internal wiki, a third-party KM solution, or something built in-house) as the team’s knowledge management system, where all artifacts must be stored.

Name your Knowledge Base Manager. Ensure that a Knowledge Base Manager is designated and that they specify naming conventions, a standardized set of categories and tags, and metadata fields for each artifact type.

Lead by example. Make a point of contributing regularly to the knowledge base. Since managers must set clear expectations and most work products are reviewed by managers, the team’s managers and policy leads are the ideal owners of Acceptance Criteria artifacts.

Add knowledge management duties to job descriptions across all levels. Make knowledge management an explicit job expectation by including KM work in job descriptions.

Reward knowledge-sharing behavior. Recognize and reward team members who actively contribute to the knowledge base. Incentivize KM contributions by explicitly recognizing them as work goals and a prerequisite to promotion.

Foster a learning culture. Consider formally adopting a Blameless Retrospectives approach to emphasize that learning comes from successes and failures. The team will not successfully extract lessons from painful experiences and codify them in artifacts if they are afraid to share openly without fear of criticism or penalty.

Ask team members to create new artifacts in the course of normal work. If you’re explaining a task or deliverable to a team member and there’s no existing artifact to share as a reference, consider asking the team member to take notes and turn those notes into a new checklist.

Allocate resources and time for KM work. Build time into the team’s schedule for knowledge base updates and retrospectives lest these tasks fall by the wayside.

Team members will only take knowledge management work seriously when it’s evident that the team’s leaders consider it important.

In Closing

Knowledge Management is an essential competency for any policy team.

Without well-defined KM practices, the team spends an inordinate amount of time searching fruitlessly for reference material, comparing versions, or recreating them. Each person maintains their own set of bookmarks as they stumble across useful references. Pockets of KM efforts may emerge organically, but each exists in isolation, resulting in duplication of effort and the proliferation of inconsistent practices.

When a team is resource-constrained, a lightweight knowledge management approach becomes vital. By focusing on four distinct types of artifacts—Templates, Checklists, Playbooks, and Acceptance Criteria—a policy team can direct its KM efforts toward artifacts that have a meaningful impact. These KM artifacts save time, reduce errors that lead to rework, improve transparency, encourage a culture of continuous learning, and improve the consistency and quality of the team’s deliverables.

While the approach described in this post is lightweight, it still requires sustained investment and effort. Policy team leaders must actively participate, support, and recognize KM initiatives if they wish to see their teams reap the benefits of this work.

Author’s Note: Sincere thanks to Del Harvey for letting me use screenshots of her checklist as an example in this post. My thanks also go to IR, LM, DH, and YR. Their feedback helped make this post more useful.

Recommended Reads Elsewhere