Page 231 - CyberSecurityEssentialsEnglish

FIGURE 5-9: THE MATURITY LADDER OF AN AI-SUPPORTED WORKFLOW

What to notice: Maturity does not begin with an agent or with automation. It begins with

documentation, then standardization, then measurement, then improvement.
Source: Prepared by the author.

? Question to consider: where did the time actually go?
Did AI improve the full cycle time, or did it make the draft faster and shift the burden to
review and correction?

● Remember: Section summary
• Key idea: Improvement needs shared assets and performance data.
• Practical decision: Keep the templates, prompts, and versions, and compare the whole
cycle against a baseline.
• Mistake to avoid: Celebrating generation speed while ignoring review, rework, and

errors.

5.12 Extended Case Study: A Monthly Briefing Pack for Management
The previous situation: in the first week of each month, management asked for a short

briefing on the performance of an internal program. The inputs arrived from three sources:
an indicator table, notes from delivery teams, and emails containing problems or requests
for exceptions.

Each time, a different person gathered the material and wrote the report in their own way.

Sometimes the current month’s figures were used with notes from an earlier month,
sometimes a recommendation appeared with no clear source, and sometimes a follow-up
on a previous decision was forgotten because the earlier minutes were never turned into an

action log.

The problem was not the team’s ability to write. It was the absence of a fixed way of working.
AI shortened texts well, but it did not know which file was the approved one, it did not always

230