Riseplan from Roast & Rise
Company Knowledge System
Create one trusted version of how the company works
Turn scattered documents, decisions, examples, and team know-how into a maintained company memory people can trust and use in daily work.
Course thesis
A company knowledge system is not a prettier folder or a bigger wiki. It is an operating system for keeping decisions, context, workflows, examples, standards, and customer knowledge trustworthy enough for people and AI-assisted workflows to use.
What you leave with
A practical company memory with a knowledge priority list, navigation map, ownership model, review rhythm, retrieval tests, and rules for what belongs in the system.
Learner
Teams that keep answering the same questions, losing context between projects, onboarding slowly, or struggling to make AI useful because knowledge is scattered.
Workflow
Knowledge prioritization, memory mapping, source and ownership design, review routines, retrieval testing, onboarding, and AI-assisted decision support.
Behavior
Move from scattered knowledge and repeated questions to a maintained company memory with owners, review dates, source clarity, retrieval quality, and explicit rules for stale or uncertain information.
Outcomes
What the learner should be able to do after finishing this public Riseplan.
Decide which knowledge is durable, high-value, and repeated enough to belong in the company memory.
Design a simple memory map that reflects how the company actually sells, delivers, decides, and improves.
Assign ownership, source rules, review triggers, and confidence labels so trust does not decay quietly.
Create the first high-value entries using examples, decisions, templates, and operating routines from real work.
Test whether people and AI-assisted workflows can retrieve answers with enough accuracy, source clarity, and usefulness.
Chapters
01
Define What Must Be Remembered
Separate durable company knowledge from temporary notes, chat history, and low-value archive material.
The first mistake is saving everything. A strong company memory starts by deciding what must remain true, findable, and reviewable over time.
Durable knowledge includes offers, customer segments, positioning, process, decisions, examples, standards, operating routines, pricing logic, delivery rules, and lessons learned. Temporary discussion belongs somewhere else.
The best starting point is repeated questions. If people keep asking the same thing, the company has already revealed a knowledge gap that costs time and consistency.
Prioritization matters because knowledge systems fail when teams try to migrate the whole archive before proving usefulness. Start with the questions that block work today.
Exercise
List the ten questions people repeatedly ask.
Collect repeated questions from chat, meetings, onboarding, sales, delivery, operations, and leadership. For each question, write who asks it, why it matters, what a good answer must include, where the answer currently lives, and what breaks when the answer is missing or wrong.
Artifact
Knowledge priority list.
Use this at work tomorrow
Ask three teammates which question they are tired of answering and where they currently look for the answer.
02
Design the Memory Map
Create a structure for offers, customers, process, decisions, examples, culture, and operating routines.
A company memory needs a map people can understand without training. If the structure is too abstract, the system becomes another place to search and abandon.
Use categories that match the business: what we sell, who we serve, how we work, what we decided, what good looks like, what changed, and what needs review.
The map should support action. Every section should help someone decide, write, onboard, sell, deliver, support, manage risk, or improve quality.
Design for retrieval, not storage. A useful map includes names people would search for, example questions each section answers, and links between related decisions and workflows.
Exercise
Draft the first navigation model for company knowledge.
Create five to seven top-level sections. Under each, add three example entries, the real question each entry should answer, likely owner, source of truth, and review trigger. Test whether a teammate can place five messy documents into the map.
Artifact
Company memory map.
Use this at work tomorrow
Create a one-page draft of the memory map and test whether a teammate can find where three real entries belong.
03
Assign Trust and Ownership
Decide who owns each knowledge area, how sources are shown, and how stale information gets corrected.
Knowledge quality is an ownership problem. If nobody owns an entry, everyone eventually distrusts it, and AI-assisted workflows amplify the confusion.
Each important entry needs an owner, a backup owner, source links, a review rhythm, and a visible last-reviewed date. This makes uncertainty manageable instead of hidden.
Ownership should follow expertise and workflow, not hierarchy. The person closest to the knowledge should usually maintain it, while leadership owns decisions that set direction.
Add confidence labels for entries that are draft, inferred, outdated, or waiting for approval. A useful system can say 'we do not know yet' without pretending.
Exercise
Assign owners, sources, and review dates to the first ten entries.
For each priority entry, assign an owner, backup owner, source link, review date, confidence status, and trigger that means the entry must be updated sooner. Mark which entries are safe for AI-assisted retrieval and which need human confirmation.
Artifact
Knowledge ownership table.
Use this at work tomorrow
Choose one high-value entry and add an owner, source, confidence status, and next review date.
04
Create Useful Entries
Turn raw knowledge into pages, templates, examples, and decision records that people can actually use.
A useful knowledge entry is not just a note. It answers a real question, explains when it applies, shows examples, names the owner, and tells people what to do next.
The strongest entries combine a short answer, context, examples, source links, related decisions, and update rules. This makes the knowledge usable for humans and safer for AI-assisted workflows.
Templates and examples matter because they reveal what good looks like. Without examples, teams keep asking for clarification or recreate work in different styles.
Start with a small set of high-value entries and improve them through use. A living memory is built by answering real work, not by perfecting taxonomy in isolation.
Exercise
Write three high-value memory entries from the priority list.
For each entry, include the question it answers, short answer, context, examples, owner, source, last-reviewed date, related entries, and what someone should do if the answer seems wrong. Ask one teammate to use each entry on real work.
Artifact
First company memory entries.
Use this at work tomorrow
Turn one repeated question into a complete memory entry with source, owner, and example.
05
Make Retrieval Trustworthy
Test whether stored knowledge produces answers, templates, decisions, and onboarding support people can trust.
A knowledge system is only useful if retrieval produces answers people trust. Test it with real questions, not with happy-path demos.
Good retrieval gives answer, source, confidence, context, and next action. It should distinguish confirmed fact from interpretation and flag when a human owner should verify.
Every failed search is product feedback. Use failures to improve naming, structure, examples, owner assignment, and stale content rules.
For AI-assisted retrieval, require source visibility. The system should make it easy to inspect where an answer came from before it becomes a customer message, decision, or internal standard.
Exercise
Test five real questions against the knowledge system.
Ask five common questions and score the answers for accuracy, completeness, source clarity, confidence, and usefulness. For every weak answer, decide whether the fix is content, structure, ownership, naming, review rhythm, or tooling.
Artifact
Retrieval quality checklist.
Use this at work tomorrow
Test one real onboarding or customer question and note exactly where the answer breaks down.
30-day path
Week 1: identify repeated questions, critical decisions, and the first ten high-value knowledge entries.
Week 2: build the memory map, owner model, source rules, and confidence labels.
Week 3: create the first entries and test them against live sales, delivery, onboarding, or operations questions.
Week 4: run retrieval tests, fix weak answers, and install the review rhythm.
After 30 days: expand only after the team trusts the first entries and owners keep them current.
Success signals
Repeated high-value questions are answered from one trusted place.
Every critical entry has an owner, source, confidence status, and review date.
New team members can find core context without interrupting the same people repeatedly.
AI-assisted answers cite maintained entries rather than scattered or stale material.
Retrieval tests produce specific fixes instead of vague complaints that the wiki is messy.
Reflection prompts
Which company knowledge is most painful when it is missing, stale, or wrong?
Where do people currently go first when they need an answer, and why?
What would make the team trust an answer from the knowledge system?
Which knowledge should not be used by AI without human confirmation?
Manager checklist
Choose the first ten high-value entries, not the whole archive.
Assign owners, sources, confidence labels, and review dates before scaling.
Test retrieval with real team questions every week for the first month.
Remove, update, or mark stale content instead of letting trust erode.
Require source visibility for AI-assisted knowledge use.
Want this shaped around your company?
Risey can research your company foundation first, then build a version of this path around your real workflows, customers, and culture.
Start with your company