Categories
Collaboration

Knowledge base and transfer

Photo by Scott Graham on Unsplash

Knowledge bases are important for teams. They help teams to document the history, current state and planned future of software products and services.

Ideally knowledge bases are maintained in collaboration platforms (eg: SharePoint, Confluence). They are ideally online, searchable and well maintained.

Very often knowledge bases are outdated and not kept uptodate. This is because teams rarely have documentation in their checklist (Scrum: Definition of Done) before marking a task as Done.

This results in a technical debt that is coming back to catch you when you really need it. For example, when a team member leaves, a knowledge transfer needs to happen.

  1. The knowledge transfer would be easy if a knowledge base existed and is constantly kept uptodate.
  2. The knowledge transfer would be easy if every task follows the two person rule (or four eyes principle).
  3. Every active product or service has a technical documentation with setup, configuration, changes and support
  4. Every active product or service has a user manual with usage scenarios and expectations

Now let us see how we can make the above things part of our daily routine.

Step 1:

Every task completed by the team is considered DONE only when two things are satisfied

  1. There is documentation for the product or service feature
  2. There is second person who read and approved of the documentation

Step 2:

Every product or service which goes to Production is considered DONE only when two things are satisfied

  1. There is technical documentation for developers on the team to support and extend it
  2. There is a user (friendly) manual for customers to be able to use the product or service

Even when the above two steps are complete when a person leaves a team a handover needs to be done. Here is a sample list of some of the items we check for handovers.

Handover checklist:

  • Project name
  • Project status (active, inactive)
  • People
  • Priority
  • Handover to (person)
  • Handover activities
  • Due date
  • Status (not started, in progress, completed)
  • Documentation links
  • Notes

Using the handover checklist a team can do a kickoff for the handover where all the projects that need a handover are identified. The important parts are being able to identify the People and the Handover to person.

If a project has only one person on the People, it is already a sign that the handover will be difficult. However with enough planning this can be remedied.

In the kickoff the Handover activities need not be detailed. This can be done between the person doing the handover and the person receiving the handover.

Depending on the priority and the time available the due dates may be timeboxed and a knowledge transfer may be done.

You have a Handover to be done?

With the knowledge base, two person rule, technical documentation and user manual you are ready to handle it with ease.

Categories
Collaboration

Managing Bias with AI

I attended a workshop run by Johanna from appliedAI in the IHK meetup on October 2023.

One of the discussions we had was for managing bias with AI as a HR professional.

Some of the discussions we had were interesting:

  • Will the AI be bias free in ranking the applicants for a job?
    • Are humans bias free when they rank applicants?
  • Is it ok to use AI for the selection procedure if we inform the applicants in advance of the AI engine and version?
    • Will the applicants tweak their applications to get favourable results from the AI engine?
  • Should A/B testing be done with a group of HR team members using AI and another group NOT using AI?
    • Will the applicant ranked by AI be less bias influenced?
  • Can a set of HR rules be established and prompted to the AI to satisfy all or a maximum set of rules while choosing or ranking a candidate?
    • Can the rules be changed as an when new HR rules need to be added or changed?
  • Can the AI engine be asked to explain the decision workflow while deciding or ranking a candidate?
    • Will the decision process itself expose some bias?

I am also looking forward to the EU AI Act on how it will affect the usage of AI in companies:

Link for image on article and slides from AI for the SME sector:

KI für den Mittelstand – die jährliche Konferenz | IHK München (ihk-muenchen.de)

Link to EU AI Act:

Artificial intelligence act (europa.eu)