Do You Need a New Knowledge Platform? How to Decide Before You Buy Software
When documentation is a mess, the instinctive fix is a tool. A wiki, a knowledge base, a document management system, an AI search layer: the market offers many, and every demo makes the problem look simple. Yet many organisations have bought a platform, populated it with enthusiasm and watched it decay within a year.
Lakhlani International's Knowledge Maturity Management page puts the principle plainly: a new software platform is not necessarily required, and the problem, current tools and user behaviours should be assessed first, with the platform decision following from requirements. This article turns that principle into a decision process. It names no vendors and makes no claims about any product.
Why platform-first decisions disappoint
Tools fail for reasons that have little to do with features.
- →No owner. Content nobody is responsible for goes stale.
- →No habit. If people do not look there first, they will not contribute either.
- →Wrong problem. A tool cannot capture judgement that lives in someone's head, and it cannot resolve a dispute about who owns a process.
- →Hidden cost. Migration, tagging, permissions, training and maintenance often exceed the licence fee.
The capability page makes a related point: effectiveness depends on participation, ownership and ongoing maintenance. A platform provides storage and search. It does not provide those.
Step 1: Define the problem in business terms
Write the problem as a sentence that a manager would recognise. Examples:
- →New joiners take too long to become productive because they cannot find how things are done.
- →Customer questions get inconsistent answers because the latest policy is not accessible.
- →When our senior planner is away, nobody knows how to handle exceptions.
- →Teams redo analysis because earlier findings are scattered.
A good problem statement says who is affected, what goes wrong and how you would notice an improvement. If you cannot write it, you are not ready to choose a tool.
Step 2: Audit what you already have
Most organisations already own several places where knowledge lives: shared drives, email, chat, a wiki, a ticketing system, project tools. List them and note, for each:
- →What kind of knowledge it holds.
- →Who uses it and how often.
- →Who owns it and keeps it current.
- →How easy it is to find things.
- →What people do instead when it fails.
This maps to the capability's step of mapping current repositories, ownership and maintenance practices. Often the audit shows that the issue is conventions (naming, structure, ownership), not the absence of a platform.
Step 3: Observe real behaviour
Watch how people actually look for answers. Ask five people to find three specific pieces of information and note where they go, how long it takes and where they give up. The capability page speaks of assessing user behaviours first. What people do tells you more than what a survey says they would do.
Step 4: Write requirements
Only now should you list what a platform must do. Cover:
Requirements checklist
- Content. What types (procedures, decisions, FAQs, templates, recordings)? How many items?
- Findability. How will people search or browse? Do you need structure, tags or both?
- Access. Who can view and edit? Are there confidentiality needs or client restrictions?
- Ownership and review. Can each item have an owner and a review date, with reminders?
- Integration. Does it need to work with the tools people already use daily?
- Migration. How much existing content must move, and who will clean it?
- Adoption. What will make using it the easiest option?
- Cost. Licences, setup, migration, training and maintenance over a few years.
- Exit. Can you export your content if you leave?
If you handle personal data of customers or staff, include data protection requirements and take advice. In India, that includes how the Digital Personal Data Protection regime applies to what you store.
Step 5: Compare options, including no new platform
Lay out realistic options:
Options and their risks
- →Better conventions on existing tools. Suits small content volume and a clear problem of structure and ownership. Main risk: relies on discipline, with weak search.
- →A shared wiki or knowledge base. Suits procedures and FAQs that many people need. Main risk: stale content without owners.
- →Document management with permissions. Suits controlled documents and version history. Main risk: heavy for informal knowledge.
- →Process or workflow tools. Suits knowledge tied to specific steps in a process. Main risk: narrow scope.
- →AI-assisted search over existing content. Suits finding information across scattered sources. Main risk: depends on the quality, currency and permissions of the underlying content.
The last option deserves a note. AI-assisted search is increasingly offered, but its usefulness depends on the content it searches. That assessment reflects this guide's analysis; there is little independent evidence on how well it performs in SMB settings. If your content is outdated or poorly permissioned, an AI layer can surface outdated or inappropriate answers faster.
The comparison should include the option of changing nothing in tooling and fixing ownership and conventions first.
Step 6: Pilot before committing
Choose the platform or approach for one priority domain, such as onboarding for one team or the top three at-risk processes. Run it for a limited period and measure use: do people look there first, does new content get created, are owners reviewing it? The capability's last step, measuring adoption and revising the system based on actual use, is the test.
Signals that you probably do need a platform
- →The volume of content has outgrown shared folders and no one can find anything.
- →You need permissions and version control that existing tools cannot provide.
- →Several teams need the same answers and keep copying files.
- →You have audited current tools and the gap is a missing capability, not missing discipline.
Signals to fix the basics first
- →Nobody owns the existing documentation.
- →There is no agreed place to look.
- →Staff say they do not trust what is written down.
- →Critical know-how is mostly tacit and has not been discussed.
An illustrative scenario (hypothetical)
Consider a hypothetical 40-person services company where the operations head wants to buy a knowledge base to fix inconsistent customer answers. The audit shows that three versions of the pricing policy exist across a shared drive, email and a team chat, and nobody owns it. Observing five staff finding the current policy takes several attempts and mixed results.
The company decides that its first need is one owned, current policy document in a clearly named location, with a review rule. A simple wiki page is created for the top ten customer questions. After a trial, usage data suggests whether a fuller platform is justified. This scenario is invented for illustration and does not describe any client.
Next step
Write your problem statement and list your current repositories before you take another vendor demo. The Knowledge Maturity Management capability sets out how an assessment runs. For technology and process change more broadly, see the Digital Transformation capability.
Knowledge Maturity Management
An assessment that starts from the problem, the current tools and how people actually behave.
Explore the capabilityReducing key-person dependency
How to rank knowledge risk and capture what matters before choosing any tool.
Read the key-person guideDigital Transformation
For the wider process, platform and customer-experience change around a knowledge programme.
Explore Digital TransformationFrequently asked questions
Do we need knowledge management software?
Not necessarily. Start by defining the problem, auditing your current tools and observing how people find information. If the gap is a missing capability, such as permissions or version control, a platform may help. If it is ownership and habit, software alone will not fix it.
What is the difference between a wiki and a knowledge base?
Usage varies. A wiki is typically collaboratively edited and loosely structured. A knowledge base is often more curated, with defined article types and review processes. Check what each product actually offers.
How do we choose a knowledge management tool?
Write requirements first covering content, findability, access, ownership, integration, migration, adoption, cost and exit. Compare options against them, including improving what you already have, and pilot before committing.
Why do knowledge bases go stale?
Usually because nobody owns the content, there is no review rule and people have no habit of using or updating it. Assigning owners and review dates addresses this more than any feature.
Should we use AI search on our documents?
It may help find information across scattered sources, but its usefulness depends on the quality, currency and permissions of the content. Consider data protection and take advice before pointing any AI tool at personal or confidential data.
How long should a pilot run?
Long enough to see whether people use it unprompted and whether owners keep content current. Agree measures in advance, such as how often people search, what they find and what content gets updated.
References
- →Valiukečienė, L. Analysis of Knowledge Management Maturity Models. Vilnius University. https://www.journals.vu.lt/open-series/article/view/43556
- →ISO 30401:2018, Knowledge management systems — Requirements (catalogue record). https://genorma.com/en/standards/iso-30401-2018
- →Lakhlani International. Knowledge Maturity Management capability page. https://lakhlaniinternational.com/capabilities/knowledge-maturity-management