Will AI save Knowledge Management?

-

Since its emergence 30 years ago, knowledge management has sought to capture, share and enhance the collective knowledge of organizations. But despite technological and methodological advances, one central challenge persists: how to make explicit the tacit knowledge that resides in the minds of employees? As AI makes its thunderous arrival on the business scene, we can’t help but wonder whether it might be able to perform the long-awaited miracle in this area, and become the trigger for truly widespread and effective KM.

Older readers of this blog will remember that knowledge management was, in its early days, a very present theme here, and an article published at the beginning of the year which I rediscovered this weekend (Knowledge management: back to the future FR) led me to wonder about the progress made in this field since then.

It’s useful to put things in context to understand why I’m suddenly asking myself this question.

AI in the Web 2.0 era: an unfulfilled promise

We were at the point of what the author describes as a major evolution in KM tools and practices, particularly in the operation of communities of practice.

CoPs have proven over time to be highly effective in increasing knowledge transfer and social learning. Technology was still needed to support KM activities, but was now seen as a facilitator rather than the main driver. Around the mid-2000s, social media tools began to appear and be used internally by organizations as Web 2.0 collaborative tools.

The idea was simple: for people to express the knowledge in their heads in a tangible, reusable way, they needed two things: a medium and a stimulus.

At the time, the medium was the blogs and wikis proliferating on the web, giving everyone a practical and simple way of sharing information, and the stimulus was precisely the ability to easily react to what was published. Back in the day, Loïc Le Meur, nicknamed the “Pope of Blogs” in France, used to say that “blogs start conversations”.

Businesses immediately saw enormous potential in terms of KM.

For those of you in the know, BlueKiwi Software, an adventure I was lucky enough to join in its early days, was born out of a request from a major French business to offer a KM tool inspired by blogs and Web 2.0, and its slogan was “for fruitful conversations”.

Then Gartner invented the “enterprise social software” category, and BlueKiwi and other Yammer products joined it, leaving the KM category behind . This was consistent with their protean nature: some saw them as KM tools, others as collaboration tools, others as engagement tools or even skill identification tools. In fact, it was a bit of everything.

But we have to admit that the promise was not kept. I won’t go into the details, as I must have written a hundred articles on the subject at the time, but social tools in business follow more or less the same rule as on the web: the 1-9-90 rule.

1% contributors, 9% participants, 90% passive readers. On the scale of the web, we’re talking about hundreds of millions of contributors, but on the scale of a business, that’s infinitely less, and that’s a real problem. At the time, companies like IBM were often cited as pioneers, but with 500,000 employees and a genuine culture of communities of practice, it was “easy”.

So, in the absence of critical mass, something else was needed. 

The first was culture. Danone was a great French success story when it came to corporate social networks, but at the time Danone had a very advanced KM practice, with events called “Marketplaces” where the transfer of best practices between “givers” and “takers” was organized.

Technology has only enabled these practices to exist on a permanent, large scale, without having to move people around, but it’s still just the scaling up of a shared, existing practice.

Another alternative, which in my opinion was the only way to make it work, was to integrate social tools into processes , because discussing a subject as part of one’s professional activity, to collaborate or solve a problem openly (the famous “working out loud”) was the best way to get knowledge expressed and captured.

Unfortunately, with very few exceptions, businesses haven’t had the courage to go that far. They were afraid of changing the way things were structured and making the process compulsory, yet when they did (for example, when I worked with the sales department at Dassault Systèmes France), the ROI was very fast. (The twilight of enterprise 2.0 and the emergence of process socialization and Enterprise 2.0 and social business : what’s next ? (Part #2 : tools)) to name just two of the many articles I’ve written on the subject).

In short, the promise of social media in KM has been a failure (but I’ll certainly talk about it again sometime soon), because we’ve never been able to make the approach take off in the absence of managerial courage, appropriate processes and the right culture. Any one of these elements would have sufficed, but most of the time we didn’t have any of them.

A very long introduction, but necessary to understand what follows.

Back to square one, then, until the arrival of AI, which raises legitimate expectations.

AI and KM: an obvious partnership

A priori, AI and KM form a natural alliance. While KM aims to optimize the use of organizational knowledge, AI excels in the analysis of complex data and the creation of novel connections. Together, they can overcome the historical limitations of KM, particularly in the management of tacit knowledge as described above.

First and foremost, AI can be seen as a catalyst for the efficiency of KM approaches.

Traditional knowledge bases often suffer from problems of relevance, redundancy or obsolescence (Data debt hampers AI investments, sustainable processes drive business value). AI offers solutions by :

  • Cleansing and structuring data via advanced semantic analysis tools.
  • Identifying gaps or duplications in existing content.
  • Providing dynamic, contextual recommendations of people and information using machine learning algorithms.

For example, an AI-enhanced internal search engine can provide accurate, personalized and up-to-date answers, speeding up decision-making and problem-solving. More on this below.

Making tacit knowledge explicit: a challenge for AI

Tacit knowledge, often considered as “what we know but find hard to express”, is essential to the functioning of organizations. It is based on experience, intuition and interpersonal relationships, which makes it all the more difficult to capture, as its holders must have the opportunity and desire to release it. But AI opens up promising prospects:

1°) Identify hidden experts: By analyzing employees’ digital interactions, contributions and behaviors (emails, documents, internal forums), AI can map often invisible networks of expertise.

2°) Facilitating the extraction of tacit knowledge : Tools based on natural language processing can transform meetings, conversations or informal exchanges into usable content. Virtual assistants, for example, can automatically summarize discussions or capture “best practices” evoked during collaborative projects.

3°) Create adaptive knowledge bases: By combining AI with knowledge management solutions, it becomes possible to document tacit knowledge in the form of interactive guides or reusable artifacts, making it easier to pass on to new employees.

4°)Stimulate collective reflection: AI-powered platforms can organize workshops or virtual brainstorming sessions to help teams make implicit knowledge explicit through collaborative scenarios.

This is the major evolution from the “web 2.0 era”: we no longer try to motivate employees to go to a tool and say what they know, but we take knowledge where it is, where it is expressed not by desire but by necessity. This reminds me of an initiative I saw with the search engine Sinequa (The implicit social network according to Sinequa) and it makes me think of what the French start-up Ask For The Moon is doing, even if it does make me wonder about the confidentiality issues linked to these new approaches, which are bound to raise questions of trust.

Fostering a culture of continuous learning thanks to AI

Another potential promise of AI lies in fostering a culture of continuous learning by :

  • Offering personalized recommendations based on the specific needs of each employee.
  • Creating adaptive learning paths based on behavioral data.
  • Providing instant access to essential problem-solving information in real time.

We can therefore legitimately expect quick skills upgrades and enhanced organizational agility. That is, if the promise is kept, because history is littered with technological promises that have not been kept, usually due to human factors.

Challenges ahead: AI can’t solve everything

Despite its promise, AI is not a miracle solution. It requires a suitable environment and raises a number of issues. We can’t use technology to solve problems that are primarily human in nature.

  • Data quality: AI can only be effective if the initial knowledge bases are well structured and relevant. By virtue of the famous “shit in shit out” principle, if humans enter erroneous data or fail to update it, AI won’t perform miracles.
  • Algorithmic biases: Biases in the data can limit the reliability of results. If humans do nothing and AI learns from them, don’t expect it to improve things – quite the contrary.
  • Human adoption: Employees need to be trained in the use of AI tools, and a culture of knowledge-sharing needs to be fostered, even if, as we’ve seen, it’s now possible to fetch information where it’s needed. They will also need to understand that the process is designed to help them, not to extract their knowledge in order to replace it.

In short, AI can amplify KM efforts, but it cannot compensate for poor governance, a lack of clear strategy and human factor issues.

KM and AI: a promising future, but first and foremost a human one

AI can profoundly transform knowledge management, in particular by making tacit knowledge more accessible and exploitable. However, this transformation will only be fully realized if AI is integrated into an approach centered on humans, who are ultimately the only holders and users of knowledge.

Once again, as is often the case when it comes to technology, it’s all a question of context, and I’d like to remind you of a principle that can be applied to just about any technology :We should not expect an application to work in environments for which its assumptions are not valid.

Organizations should therefore take care to :

  • Framing the use of AI with clear governance to align technology and strategic objectives.
  • Valuing people: AI is a tool that should be used to support human expertise, not replace it.
  • Cultivate an organizational learning mentality: Encourage exchanges, experimentation and knowledge sharing.

Bottom line

AI has the potential to solve some of KM’s historical challenges, particularly in the area of tacit knowledge. But its implementation must be carefully planned and supported. Effective knowledge management is first and foremost a human affair: understanding, sharing and valuing collective knowledge.

This is all the more true as the application of AI to KM is moving in the direction of capturing information where it is, with, as already mentioned, issues linked to confidentiality and trust. It’s all about how people perceive the process….

By combining the technological capabilities of AI with a learning-oriented organizational culture, businesses can build an agile, innovative and sustainable KM.

Can AI save KM? Yes, provided it becomes an ally, not a substitute, for human knowledge.

Photo by Stockphotos.com

Bertrand DUPERRIN
Bertrand DUPERRINhttps://www.duperrin.com/english
Head of People and Business Delivery @Emakina / Former consulting director / Crossroads of people, business and technology / Speaker / Compulsive traveler
Vous parlez français ? La version française n'est qu'à un clic.
1,756FansLike
11,559FollowersFollow
26SubscribersSubscribe

Recent