AI is the perfect trigger to get serious about capturing the knowledge across your business
Because AI seems to be the focus of much of the world’s attention, any AI initiative that doesn’t amplify what is smartest about a business must almost by definition be an investment in automating what is sub-optimal, leaving a space for someone else to do it better.
Thus, what knowledge is captured to yield a business’s best result, as well as how it is transformed into clean data capable of delivering it matters, and because most businesses aren’t native to publishing, most also have something to learn about the long-established best practices in managing this process.
As such, what follows is based on my own knowledge from having worked for over 20 years as a professional sub-editor – a key fact-checking, sense-making and quality control editorial role in all professional media – on a variety of professionally produced magazines at ACP (since, Bauer) Magazines, and on the pages of The Australian Financial Review newspaper group in Sydney.
It is also informed by my first-hand experience of being enlisted to help produce high-quality technical documentation within the 200-strong technical development team of a Big Four Australian bank on its transformational undertaking to deliver its services via an array of mobile devices.
As its communication vehicle for this purpose, the bank’s development team used the Atlassian wiki, Confluence, a first-class tool deservedly beloved of technical teams around the world.
And quite apart from what I learned there, in the interests of full disclosure, I also consulted ChatGPT and Bard on the way through, not to write the actual words in this content, but more as a check to make sure I didn’t miss anything important that you as a reader might need to know.
The AI why
- In line with its information needs, culture, organisational objectives and strategy for competitive success, what are your organisation’s AI-system objectives?
- What must be achieved through AI by documenting the knowledge of the organisation?
- What information must be shared to be developed and refined through workplace collaboration?
Audiences
Who must be satisfied by your development and delivery of AI content?
Among these groups may be:
- Leaders, whose support the designers, developers and managers of the system will need.
- Managers who have to sign off content, such that it can be fed to the AI engine.
- Developers who are needed to build on it and design from its outputs and user feedback.
- Workplace users who must use it to perform their daily work, and to rate and feed back their requirements of it as it evolves.
- Customers and partners who will use it externally to buy through it and receive service.
If or where not addressed above, what ways must the system enable the organisation to interact with these strategic stakeholder groups:
- The board, management committee or equivalent
- Shareholders, owners or equivalent
- Executive or senior management team
- Middle management
- Front-line supervisors or first-level managers
- Employees
- Strategic partners or allies
- Major suppliers and services providers
- The wider investment community
- Industry or general media
- The general community or other stakeholders
Content
What are the critical strategic knowledge assets of the organisation?
What content must be captured and with what frequency to grow this?
How will decisions be made about the best ways to document and report on what is captured, and to whom?
What are the most important things that discrete groups need to know in order to do their jobs effectively?
What are the critical processes, procedures, best practices, customer insights, and technical information around which AI improvements must be focused and driven?
Workplace technologies for knowledge capture
What tools will you use, and through which channels will you capture and process content?
How will data be stored and processed through its various stages of development?
How, and by whom, will the system itself be documented for management review?
How will system access and security be managed?
How will system usage and success be measured?
Style rules
How must documents be structured to enable reliable machine learning?
What conventions will be used in individual documents to address meaning, spelling, punctuation, names, currencies, technical measures and conversions?
Who will enforce style and document consistency, and where will its requirements be made accessible across the business?
Contributors and process
What will be the process for capturing new insights and knowledge?
As described above, who will commission new documentation, spelling out its requirements?
Who will be the initial contributors, creating and feeding new content into your system?
At what stage will new content be exposed to its first audiences, and who will be first readers/reviewers of new content?
Which subject matter experts will be needed/invited to review new content, and at which stages of document development will their input kick in?
How will the cycle of review be managed?
Who will act as final style policemen?
Who will sign off content as being complete?
Participation protocols and engagement strategies
How can contributors test their documentation prior to exposing it to wider exposure?
How will users be trained and introduced to the system?
Will users be allowed to comment anonymously about new content?
How may participants be rewarded for their contributions?
Content scheduling and management
How will the workflow and editorial calendar be managed and communicated to encourage contribution?
Who will monitor the correct indexation, tagging and searchability of a document on the system?
How will document structure be enforced to ensure usability and fitness for purpose?
How will version control be managed?
Who will rewrite and check documentation back with the original author or creator to ensure consistency and that what is written is what was intended?
How will references within documents be checked?
How will feedback be captured to facilitate system development, and how will continuous AI improvement be iterated and managed?
Allow me to help you plan to remove some of the data risks to build more reliable AI
With my extensive knowledge of the professional editing and publishing processes, to improve your chances of generating repeatedly reliable, clean AI data, please put my proofreading and writing skills to the test.
Contact me:
By email, at cloudcitizenx@icloud.com
By phone, on 0416 171724
Or via Linked in, at: https://www.linkedin.com/in/grahamlauren/