Everyone's excited to discuss Artificial Intelligence, and even more excited when Deep Learning is brought up. Surely, this is a revolutionary concept. That said, it has surely somewhat dizzied us.

Some seem to think that in the future professions will change and workers will become redundant. Some refer to production workers, others add accountants and other to the list. They are probably right.

Some say that machines will perform many actions previously performed by humans. Driving, for example. They are probably right.

Some say that in a world of AI and Deep learning, Knowledge Management is no longer needed.

They are certainly wrong.

There are many fields for which Deep Learning cannot provide solutions. I believe that these fields' nature is essentially different. Some say that human beings are irreplaceable with regard to human emotion, yet this insight less helps us in taking business decisions in organizations.

I wish to focus on an aspect that is a "black hole" for AI and Deep Learning technologies, the world of small numbers.

The power of Artificial Intelligence is based on deep learning, endless data results analysis, while utilizing neuron network algorithms to learn from the data and information found. The near future will probably feature Deep Learning-oriented machines making decisions better than us. Far better.

However, our business, organizational and personal world involves many decisions that are based on small samples and few stats. Deep learning is less relevant to these areas.

Furthermore, human decisions are not better. As Tversky and Kahneman have taught us in their Nobel Prize winning research, humans systematically tend to make wrong decisions, especially in small scale. Once, when wrongly assuming that behavior in small scale situations is similar to statistic operation typical of large scale. And secondly, when we make mistakes typical of any scale, for example, when we think of a certain solution then suddenly notice that everyone is apparently implementing this solution.

This is where Knowledge Management can come in hand. Knowledge Management reflects the knowledge accumulated through experience and can present us with what has been previously learned. Knowledge Management can serve as a rational anchor by either setting us an insight database, sharing the products of our analyses and activities, or by holding an expert/colleague forum. This anchor can not only shorten processes but also optimize decision making. In areas in which both human intellect and AI cannot be of assistance, Knowledge Management is the natural solution.

Like with the three monkeys we can end with not hearing, not seeing and not talking. But, When AI, human intuitions and Knowledge Management are coupled wisely, these 3 combined factors lead to an optimal solution in almost all situations.

The fourth industrial revolution has arrived. The possibilities of AI and how we will benefit from it is mind boggling and beyond imagination of many. It is said that like second industrial revolution resulted in us getting electrified, the fourth industrial revolution will end in us being ‘cognified’. We are getting into a data and insight driven world and it will be interesting to check the linkage between Knowledge management and Artificial intelligence at this juncture so that we leverage AI in a more meaningful way.

To understand the linkage between KM and AI, let us first understand what exactly organizations do with knowledge. Organizations perform different kinds of tasks and their success and competitiveness depends on the maturity in performing critical tasks, as well as where they stand with respect to industry in this. Tasks are performed by employees and machines, who take input information about the task, process the same based on knowledge (know how and know why) and complete the task. A physician collects symptoms, a professor’s input is what was taught in earlier session of the class, an architect needs requirements from the client etc. Hence for any task there is an input in the form of information, then that information is processed using knowledge and output is created.

In the case of humans, they can process large variation in the input information with respect to a task, even if the input information is not clear, they can remove the noise and if they do not have the relevant knowledge to process the information, they do further study, discuss with others, gain further knowledge and work on the information. They apply both know-how (procedural knowledge) and know-why (causal knowledge) as required. In the case of machines, they are pre-coded with rules (Know-how) on how to process the input information. The types of input information that they can process is very well defined. The knowledge (know-how) created to process the inputs are created by humans and used by both machines and humans.

With advent of AI, this relationship between input, processing and output for machines started changing. AI has enabled machines to create their own know-how to transform input to output. As a result AI can take up a wider range of inputs for a task, create their own know how and give output. Through learning they improve their know how and as a result provide better outputs as they learn. Here do note that, the input range does not change much, but for the given set of inputs, output created improves as a result of learning. 

What does this mean for organizations? As mentioned earlier, success or competitiveness of an organization depends on maturity in performing tasks and how they improve upon it. There is a journey towards efficiency and effectiveness that all organizations are forced to undertake, as a result of market dynamics. Underlying this journey is a continuous decrease in complexity with respect to tasks performed, where more and more variables are identified, their relationships are understood.

How does AI impact the way tasks are performed and the learning cycle?

Positive impacts

• Improved efficiency of tasks: Due to their ability to learn and improve, AI driven technology can help an organization improve its task on a regular basis. Given an approach to performing a task, the AI tools can help reach the most efficient approach must faster.
• Expediting learning: AI based technologies if used prudently can help in fast tracking the learning cycle. This is enabled through generating new data and creating insights from the same in the way tasks are performed.
• Knowledge findability and Employee productivity: One of the most popular use cases with AI has been the ability to find relevant content faster. AI can improve search drastically and give employees the information and the knowledge most relevant to them. This in turn will improve employee productivity and overall productivity.
• Human-machine collaboration and Employee productivity: With AI taking up routine and data heavy activities, employees are able to focus on complex activities, which can directly impact overall productivity of the organization and fast track maturity in performing tasks

Limitations

• Cannot improve effectiveness: AI improvement happens at the know how level and they cannot work with causal knowledge. Hence AI technologies on its own cannot innovate and drastically change the approach to perform a task.
• AI cannot leverage existing knowledge: This is another great drawback of AI. AI is data driven and creates insights from data to improve. It is not able to leverage knowledge generated from other sources, bring them together and create a new know how with respect to the task it is performing.
• Dependency on AI algorithms may at times slow down learning: Because know-how evolved by AI technologies are a mystery when deep learning techniques are used, organizations who extensively use AI in their process, without any clear strategy may find their learning cycle slow down with respect to the specific tasks. This is because they are not able to develop any understanding about the tasks they are performing. They will also become heavily dependent on AI vendors for algorithms to perform those tasks.

Hence for organizations to stay competitive in the long run, we need an approach that considers the strengths and weakness of AI and accordingly leverage knowledge. Unplanned application of AI may actually bring down competitiveness of an organization.

Clients often ask me how to make Knowledge Management (KM) a seamless part of their workforce’s day-to-day operations. They want to know how to shift people’s perceptions from KM as “another thing I have to do in addition to my daily workload” to something that is done naturally as part of their everyday workflow. The idea that there is no immediate, one size fits all solution to effortlessly integrating KM into their culture and work may seem daunting. This goal, however, is no different than getting fit for the summer.

All too often, I am the type of person who jumps on a scale after one workout and wonders, “am I fit for summer yet?!” Having said this, the only times in my life when I’ve ever truly lost weight, gained muscle, or felt more energetic were when I spent months building new habits, like exercising daily, eating a plant-based diet, drinking tons of water, and making time to do yoga and meditation. In a similar sense, within days of implementing a KM solution like a search tool, content strategy, or taxonomy design, stakeholders want to know how these solutions have positively impacted the organization. In order to truly derive this value, however, your organization needs to design KM approaches with your employees in mind. To make KM stick, you have to:

• Motivate your employees to learn and embody new habits;
• Measure the effectiveness of your efforts in a meaningful way; and
• Reward good behavior using incentives that cater towards what drives your employees.

While this won’t happen overnight, investing in a proper integrated change effort will enable you and your organization to be well on your way to making KM stick. Ultimately, you will start to see your knowledge workers creating, sharing, and making good use of their own and one another’s knowledge and information and eventually it becomes an unconscious part of your company’s daily operations.

Understanding

I’m a big proponent of design thinking approaches because they’re based on the fundamental principle that not everything works for everyone, so you have to understand people’s needs, desires, goals, feelings, thoughts, etc. before developing a solution to help them address their daily challenges.

When it comes to fitness, some people prefer individual workouts vs. group classes, designing their own workout program vs. getting a personal trainer, or working out at home vs. going to a gym. When you’re designing KM solutions, ask and observe your end users to determine what would work best for them. Questions to ask could include:

• How tech savvy are they? Do they naturally create, share, and manage content digitally or are they still more paper-based?
• Does their work involve more individual-focused activities, such as research, or are they more collaborative in nature because they’re focused on brainstorming and developing solutions as a team?
• Do they mostly work in the same office or are they physically dispersed with people working from home? Are they part of a local, domestic, or global team?
• How long have they been with the organization? How long have they been in the workforce? How long have they been in their given field?

Understanding the people that you want to adopt the KM solutions is always the starting point for helping them begin to work differently.

Motivating

Motivation is critical for making KM solutions stick because often times people know what to do, but lack the incentive or drive to do it. I know that if I exercised daily and ate like a celebrity, I would probably look like one… or at least look and feel like a better version of myself. What’s prevented me from doing what I should do? Doing what I want to do can feel more rewarding.

Change is hard, so it’s always easier for knowledge workers to revert back to their natural ways of doing things when they are introduced to a new process or technology. For instance, knowledge workers may be accustomed to shared drives with folders within folders within folders, but shifting over to a site that can leverage metadata, as opposed to folder structures, can be challenging, even though it dramatically increases the findability of knowledge and information. Even if the proposed solution or updated process will derive value and save time, well-ingrained habits are often hard to break.

Motivation comes in many forms and different people react to different things. Having said this, taking the time to figure out whether individuals are driven by learning and mastering new skills, recognition for doing good work, or cold, hard cash can help you experiment with ways to incentivize people to practice good KM behaviors. How about offering a reward for the person who creates the most new content in a month or the person who cleans up and archives and deletes the most obsolete information from the intranet or shared drive? When people are rewarded for doing things, it teaches them what to keep doing as well as what’s important to do in order to help their organization succeed.

Embodying

I can watch tons of YouTube videos and read fitness magazines all day, but unless I work out and eat right, I won’t see any results. Similarly, people need to engage in the KM processes in order to mature from a KM standpoint. At EK, whenever we roll out a taxonomy design or content strategy for a client, we almost always design a governance plan to go along with it. Having a governance plan will help ensure that the solution is sustainable, and not just a superficial quick fix. Run through the maintenance workflows, facilitate the governance meetings, update the design based on what you learn from analytics and end-user feedback, use the new enterprise search tools, facilitate the community of practice meeting– just do it! You have to continuously do these things and encourage others to do so in order to get accustomed to doing it.

Measuring

Sometimes the number on the scale isn’t very telling. Knowing that I ran a combined 15 miles this week and feeling my pants fit a little looser could better validate that whatever I’m doing is working. This is the difference between lead and lag measures. Lag measures are metrics that capture the impact of your actions, whereas lead measures track your actions themselves. A combination of both can give you the full picture of the rate at which your KM solutions are being adopted along with the effect they’re having.

Lead Measures

• Number of new articles published.
• Number of Communities of Practice meetings held.
• Time spent transferring knowledge to another team member.

Lag Measures

• Number of unique views on an intranet page.
• Reduced time searching for information.
• Lower bounce back or drop offs from a page due to not finding the right information.

You can capture these metrics directly from the KM systems you use (Content Management Systems, Enterprise Search Tools, Taxonomy Management Tools, etc.), and you can also deploy surveys gauging your end users’ overall satisfaction with the new solutions that have been implemented to help them create, manage, store, and act on the information that they find. It is crucial to measure adoption because the numbers will help guide your future actions by telling you what’s working and not working.

Rewarding

Lastly, and most importantly, reward good behavior! I try not to celebrate good fitness outcomes by indulging in decadent meals, rather, I treat myself to a massage or a shopping spree for new outfits because it motivates me to keep going without negatively affecting the progress I’ve made.

Choose rewards that will increase your KM capabilities. Treat your employees to an event where they can share their ideas for new initiatives, invest in that technology that’ll help further automate their workload, or promote the individuals who have mastered a subject matter and shared their knowledge with others in a meaningful way.

Conclusion

When I am living my best life due to healthy habits, I have more energy, and I am spreading positive vibes. I’m more active and engaged with other people. What does it look like when an organization is adopting KM best practices?

• More people are producing higher quality, useful content.
• Communication is flowing and team members are working towards a common language.
• New technology is being seamlessly implemented.
• There’s a higher rate of social learning and sharing. Individuals are constantly learning and growing.
• Team members are encouraging each other to share knowledge and information without having to be told to do so from the top.
• There’s more creativity and innovation being used to proactively solve complex problems.
• Your workforce is positive, engaged, and envisioning themselves growing within the organizing in the near and long term.

Ditch those crash diets and quick fixes and reach out to EK to learn more about how to make KM stick for the long run.

The Million Dollar Question

If I’m a subject matter expert – and I’m recognized and rewarded for what I know – what is my motivation to share? Is it for the good of humanity, or the business, or my fellow employees? Is it to leave a legacy? Is it just because it’s the right thing to do? For most people, none of these reasons are compelling enough to stop the knowledge hoarding madness.

Breaking the Cycle

In a Harvard Business Review article titled, “How to Prevent Experts from Hoarding Knowledge,” Dorothy Leonard suggests that one of the reasons that experts are reluctant to share is that those who possess this “deep knowledge” have been undervalued in the past. Another cause for knowledge hoarding is that experts have been rewarded for the wrong things and have become part of a “superhero” culture of gurus. It’s not that expertise isn’t valuable, however. “The Ship Repair man Story – Why Experts get paid more?” points out that expertise is precious, and that those that mishandle human capital learn the hard way.  

Engagement and Rewarding the Right Things

Creating a culture of sharing is the goal, and one proven approach is to focus on employee engagement. The exceptionally well researched Gallup Employee Engagement survey provides some guidance in areas on which managers can focus. Most relevant to creating a culture of sharing are the questions that ask: have you received recent recognition, do you feel that your opinion matters and do you think your job is important. High scores on these categories (and efforts to enhance these items) can pay dividends for managers who want his or her experts to share. Systems can also be created to track and report on who are contributing to knowledge libraries. Gamification and rewards can be tied to credible contribution – and that can help prime the pump of collaboration.

Social Innovation can be defined as “the systematic disruption of social norms to effect social change.” As knowledge managers, we are quick to learn that much of the work we do involves not simply designing better processes and using technology advantageously, but disrupting the social and behavioral norms of people in our organizations in ways that enable their successful participation in new processes and technology related changes. Try as we might to introduce smarter and better ways to do business, we could not do business without people - more specifically the socially reinforced behaviors of people.  As a career technologist, knowledge strategist, and the owner of masters of science degrees in both Information Systems and Information and Knowledge Strategy, it was this revelation about behavior that led me to pursue what some might consider an unusual next step toward a Doctor of Social Work.  

Why would someone who once built web applications for universities and helped design decision making displays for Navy commanders decide to enter the field of social work at the doctoral level? Three reasons: (1) To understand how and why social innovation solves problems for our nation’s most vulnerable populations, (2) to offer my skill set to the design of new innovations in social work, and (3) to help knowledge managers and strategists discover a different set of research, science, and frameworks that can lead to innovation in the business world where interest in social enterprise and behavioral values like respect, trust, dignity, and integrity - long-standing tenets of the social work code of ethics - have continued to rise.

As my first term in the University of Southern California’s Doctor of Social Work program comes to an end, I thought I would share a helpful framework called Innovation Dynamics with our knowledge management community. This framework was designed by Andrew Benedict-Nelson and Jeff Leitner and is available in an easily digestible format in their book, See Think Solve: A Simple Way to Tackle Tough Problems. The information in quotations for the remainder of this article are taken from this book unless otherwise noted. In this blog post, we’ll walk through Benedict-Nelson and Leitner’s SEE, THINK, SOLVE approach, applying it to what one might consider a typical knowledge management (KM) problem.  First, we will identify the problem, and then rethink the problem in observable, behavioral terms using their framework.

Imagine:

Your senior leadership is fully on board with a new digital transformation initiative. Not only will you be tasked with moving the organization’s information into a new portal, but you’ll also be moving communication processes that are currently managed in another system, and heavily enabled by email, into a Slack-like tool designed to be used on desktop and mobile. Rather than waiting 24-48 hours for responses to trouble tickets, employees will have access to a searchable database and instant live support via their mobile devices. When this new tool set is finally unveiled, it is a hit! People love the design and the senior leaders love that the organization looks and functions in a more modern way. But you notice something interesting after the first two weeks. Usage has slowed and worse, some departments aren’t using your amazing troubleshooting system at all. Because the portal was a significant financial investment, senior leaders request a report out on its usage at their monthly meetings and you are dreading reporting this frustrating trend. What do you do?

Thanks to user analytics, you know the problem here is that user engagement with our new portal has measurably slowed or ceased. Your job as the knowledge manager is to determine why this is happening and design a solution that will make usage go up so your senior leaders feel like they are getting a return on their investment. Since all systems seem to be functioning as designed, this appears to be a behavioral problem in that users are not interacting as you anticipated they would with the new portal system.

If we use Benedict-Nelson and Leitner’s Innovation Dynamics framework, we can begin to understand how to SOLVE the problem, but first we need to SEE the problem through six innovation lenses. The lenses through which we will make our behavioral observations are as follows:

1. Actors: These are the people or groups of people involved in the problem. First order, second order, and missing actors all play a role here and it is your job to identify who they are. First order actors in our KM problem are people in departments that do not seem to be using the system. Second order actors might be the peers of those people in other departments or the supervisors and managers of those users. Missing actors might be trainers or customer support people who have not reviewed the new tools with the first order actors.
2. History: According to Benedict-Nelson and Leitner, “history is a collection of stories about the problem’s past - the official stories, the unofficial stories, the half-truths, and the you’ve-got-to-be-kidding-me stories.” In our KM problem’s past, people used different systems to get support and do their work before we switched to the new portal. Some people loved the old way of getting things done - maybe the 24-48 hour wait bought them time to do more or maybe they enjoyed a different way of interacting with support. You will have to uncover those stories that influence the behaviors leading to use patterns you are seeing in a particular department.
3. Limits: These are the “formal, explicit rules that influence how people behave in relation to a problem”. Could it be that in a department that isn’t using your new portal there is a rule against the use of mobile devices? Are users only able to access the portal at certain times in the day because they work outside of the office? Is there a supervisor who has created a rule that interferes with the use the new portal? These formal, explicit rules, or limits, are worth investigating, if you want to determine how to change the behavior of low or no portal usage.
4. Future: “The collection of people’s expectations about how a problem will turn out”. In our KM problem, some users think that if enough people keep using the old system they won’t be forced to use the new system because they assume the company will sustain the older resources.  Some users may not believe the information in the portal will be valuable to them and that using it will not make a difference in their own job performance. How do expectations about the future keep your problem in place?
5. Configuration: This is how people make sense of things using labels and categories. In the case of our KM problem, do the people who are not using the new portal organize their tasks or work in a certain way? What can we learn about how they organize themselves or their information that can inform why they are not using the portal? Have they categorized the portal itself as an optional tool or a must have? These are questions of configuration that can illuminate how people make sense of the portal’s use or non use in their work life.
6. Parthood: This sixth and final lens tells us that most problems are often related to, or are a part of, other problems. Might a lack of use of the new portal stem from a problem that some users don’t have computers or mobile devices? Could it be because users in a certain part of the organization have not been empowered to do their work in this new way? Discovering the other problems that might exist in relationship to our KM problem can shed light on how and where we need to change behaviors.

Now that we SEE our KM problem through these lenses, we are asked to THINK about the problem in terms of both social norms and deviance.

• Social norms are “unspoken, informal rules that tell everybody how to behave in social situations.” Here, our social situation is the workplace and we use the six lenses to look at the behaviors in our KM problem to find the norms. For example, if we see the problem (user engagement with our new portal has measurably slowed or ceased) through the lenses of actors and limits, we can identify specific people who are not using our new portal system, and, after speaking confidentially but candidly with those people, learn that that their supervisors have explicitly discouraged users from engaging in the new portal because they themselves are not using it. The social norm here would therefore be that users in department x do not engage in the use of the new portal and the six lenses help us see this is because a supervisor does not use it. This problem might seem “obvious”, but what the innovation framework does is empower us as knowledge managers to explore the problem more deeply, allowing us to engage and observe people in our organization using a more structured set of questions that can help us identify opportunities for innovation.
• Deviance “is a behavior with the potential to subvert a social norm.” Benedict-Nelson and Leitner insist that deviance not only break the rules and disrupt a social norm, but that it change the rules altogether. Understand the social norm, understand the behaviors that could unseat the rules that keep it in place. Once you identify areas and opportunities for behavioral change, you can begin to ideate on solutions and create a deviant.

This leads us to the SOLVE portion of Benedict-Nelson and Leitner’s framework. How we solve our KM problem requires us to create a deviant, not BE deviants in the traditional sense, but design mechanisms that encourage a deviant behavior in an innovation sense. If we take the problem, (user engagement with our new portal has measurably slowed or ceased), use the lenses to SEE a social norm (users in department x do not engage in the use of the new portal because a supervisor does not use it), we can posit that a deviance would be the supervisor changing his or her behavior to increase portal use within their department. But as Benedict-Nelson and Leitner emphasize, we not only want to disrupt the social norm, we want to change the rules holding that norm in place altogether. We want to understand what makes the supervisor NOT want to use the new portal and how can we change the rules around THIS behavior so that ALL supervisors will be incentivized to use the portal. One deviant innovation might be designing a recognition system that rewards supervisors directly for high departmental use and collaboration, maybe at those monthly meetings where you have to report out to the senior leadership. Perhaps another deviant innovation would be designing a questionnaire or conducting an interview that helps you tailor the new portal to each supervisor’s specific needs, addressing the WIIFM (what’s in it for me) more directly.  

Whatever your proposed innovation, or deviance, it is likely to get the support you need to execute it if you can show that you have investigated your problem using an innovation framework that has been applied, proven, well-researched, and costs nothing to your organization. I hope you will consider Benedict-Nelson and Leitner’s Innovation Dynamics and SEE THINK SOLVE when diagnosing your next major KM problem.