Tuesday, October 18, 2022

An Ecosystem Approach to Blend Learning and Systems

In the previous blog post, I touched on the need to blend systems and learning for a successful Knowledge Management program.  When I teach Knowledge Management, I talk a lot about methods, tools, approaches and the like, but in the real world of KM in organizations, there are very few people who speak that language, so the KM practitioner needs to work with practical examples, preferably focusing on processes that have an immediate and clear impact on the business' performance. 

Here is a task example:  

Jane is a relatively new employee. She is climbing the learning curve really fast, with some help.  She is working on a proposal team and she has been given the task of writing the past experience section.  She has never done this task before.  She needs to find information about relevant company past experience that can be used in the proposal. She has been given some general advice about where to look (past proposals, the repository of project documents, etc...) and the names of a couple of past projects and proposals she should look up because they are likely to be relevant.

This is not a simple task of retrieving some existing documents and doing a cut-and-paste job.  It requires some understanding of the proposal itself to transform existing information into an accurate, yet tailored rewriting to meet the specific needs of the proposal.

In the process of completing this task, Jane will gain valuable experience in at least three key areas:

  • Searching company databases
  • Combining/recombining information to create a new, tailored version of that information
  • Receiving and integrating feedback from proposal team members

Assuming well organized Knowledge Management systems, Jane will be able to spend a limited amount of time locating and retrieving relevant documents from the databases, and more time applying critical thinking to create the tailored version of the document.  

Assuming a learning-driven organization, Jane will optimize her learning by a) attending training for this  task (if available) and/or reviewing existing written guidance for this task as soon as he/she is given the task; b) receiving the support of a task-specific mentor throughout the process; c) reflecting in action; d) reflecting on action. 

Reflecting in action might mean that while Jane is search for a past proposal, she comes across another proposal that looks similar and might also be of interest. She pauses briefly and realizes that the proposal team lead who gave her the list of other proposals to look up may not be aware of this information she has come across.  She needs to decide whether to stick to what she has been asked to focus on or dig a little deeper into what she has come across.  Reflection in action may result in a quick insight that results in a slight change in direction, a decision made while she is completing the task.

Reflecting on action happens after the task has been completed and allows a look back to reflect on what has already happened and what could have been done differently.  In a learning-driven organization Jane will participate in the proposal team debrief of After Action Review (AAR), which will cover the entire proposal process. However, there is a lot that Jane can learn about her own task and how she handled it that will not necessarily be discussed in the team debrief. If Jane didn't realize that other projects might be relevant while she was gathering information (during the action), she might have that insight when she pauses to reflect after the action, perhaps in preparation for her participation in the team debrief.  

Assuming the proposal team debriefs are done consistently across the organization, the collection of debrief notes is a gold mine for analysis, to identify pain points and improve processes, existing guidance documents, and training, and continuously update information found in the "systems".  In most cases, systems are not self-sufficient and do not update themselves automatically with fresh information.  Jane, or someone on the proposal team, is going to be responsible for uploading the final version of the proposal and past experience section to the appropriate repository and add the necessary metadata according to a well-thought out taxonomy.  

Assuming Jane brings up her insight about other projects to the team debrief, it might be added as a good practice in the existing guidance.  It could be integrated as guidance to the proposal lead:  When guiding the past experience writer to specific projects and proposals, don't limit yourself to projects and proposals you know personally.  And it could be integrated as guidance for the past experience writer to "use best judgement when looking up information and don't limit yourself to the proposals and projects you have been told to look up."  

We want information to be available at the click of a button.  That gives us more time for processing that information, applying critical thinking and transforming information into actionable knowledge.  Most of the technology we use today to help organize our information so that it is easily accessible is helping with efficiency and speed in retrieval.  The technology does not do the thinking for us.  With AI and machine learning, some of that initial thinking, parsing and filtering of information sources will be more automated. 

Continuing with Jane's task example,  the proposal lead and others on the team may guide Jane towards specific existing proposals or projects that are likely to be relevant to the task.  These recommendations are based on one or two individuals' knowledge of prior projects and proposals.  In a larger organization, that could be very incomplete knowledge.  There is significant potential for missing out on relevant information.  Jane, during the search of the databases could encounter new information, but it would be totally dependent on her to engage in reflection in action and to pro-actively bring it up to the team.  That simple action will be dependent on the organizational culture and psychological safety the team.  Proposal teams can be high pressure, and even if the organizational culture promotes psychological safety, there can be subcultures that are more intimidating.

In the near future, the organization could use AI and Machine Learning to efficiently and automatically read the RFP, identify key components, then search the company's databases for all the relevant documents, and either spit out a list of all documents, sorted by relevance, or even draft a summary of the information that could serve as Jane's first draft.  Obviously, there is still need for a human applying critical thinking to decide how to adjust this first draft, but over time, assuming machine learning works as it should, the first drafts would become better and better.  The AI might even learn the organization's writing style if it is given training data based on the organization's existing materials.

Jane's task was relatively simple, yet exposed many connections between people and systems.  Expanding the task to the broader proposal process would expose hundreds of tasks and associated connections between systems and the people who maintain and use them.  And the proposal process is not totally isolated from the rest of the organization since as we saw from Jane's task, it is connected to past project experience and organizational knowledge. There is a broader, dynamic organization learning ecosystem.

All this requires a comprehensive framework for thinking through the tasks that Jane and other employees across the organization have to complete, establishing clear processes, defining roles and responsibilities, setting up user-friendly, integrated systems, and an overall governance that allows for seamless embedding of learning processes and systems. 

No comments: