Franciszek Georgiew — Owner at XXII Ventures
Talk to your company like you talk to Siri
September 20, 2024
14:05
AI in Practice
🇬🇧 English
AI in Practice
Artificial Managers or Intelligent Managers? What Will Come From The Mix of AI And Management
About
The impact of AI on developers’ work is already spectacular. But there is another worth considering dimension of AI’s impact - and it is on either management or ways of working. In this speech, I explore the ways in which ChatGPT already transforms how IT teams and companies are managed, how it changes different roles and impacts agile software development methods. I show how ChatGPT can radically improve collaboration and communication, redefine certain practices, and streamline processes. This presentation also includes a plot twist that will reveal more fundamental changes that are still ahead of us all.
Watch the full talk
Watch this WaysConf session, then continue with related talks or explore the current programme.
From the recording
Talk in brief
A reflective challenge to managers who want AI to generate more work artifacts without reconsidering why those artifacts exist. Rather than listing automated tasks, the talk proposes five areas to rethink: how much management is content creation, how juniors gain mastery when basic work is automated, what certificates mean when a model can pass tests, whether AI fixes root problems or adds another layer, and how benefits change after mass adoption. Examples such as generating long emails only to summarize them show that automation can preserve an inefficient system. The recommended stance is experimental but outcome-led: use AI where it improves learning or communication, while simplifying the process before accelerating it.
Key takeaways
- 01
Reframe managers as content creators
Emails, tickets, boards, presentations, and chat messages reveal how much management work consists of producing communication artifacts.
Watch from 2:00 - 02
Protect the path from basics to mastery
Automating beginner tasks may remove the repetition through which people learn rules well enough to challenge and transcend them.
Watch from 4:00 - 03
Rethink evidence of professional skill
When a model can pass a certification test, organizations need stronger ways to verify applied understanding at scale.
Watch from 6:01 - 04
Solve the cause before automating symptoms
Generating and summarizing the same bloated communication may add work instead of improving the underlying conversation.
Watch from 8:03 - 05
Expect adoption to change the advantage
A tool that saves one person time can create more incoming work once everyone uses it, so strategy must evolve with adoption.
Watch from 10:00
Read edited transcript highlights
These concise notes were edited from automatic captions and checked against the talk structure. They are not a verbatim transcript.
A generated answer was the wrong talk
The original plan was to ask a language model for a list of ways AI would affect managers and present the result. Although that could produce competent material, it would avoid the more useful work of examining the system around management. The revised talk therefore uses AI as a prompt to question responsibilities, learning, assessment, communication, and adoption.
Watch from 0:00Routine work is also a learning scaffold
Experienced professionals often welcome automation of repetitive tasks, yet those tasks are where beginners practice the foundations of a craft. Mastery usually progresses from following rules to challenging them and eventually working beyond them. If the first stage disappears, the entry barrier may rise and the divide between people who understand the system and those who only operate tools may widen.
Watch from 4:00Testing may no longer demonstrate knowledge
A language model was able to reach a passing score on an advanced professional certification exercise after iteration. That result does not prove the model can perform the role, but it does weaken the test as evidence that a person possesses the underlying skill. Training providers and employers must reconsider how applied judgment can be assessed when answers are easy to generate.
Watch from 6:01Automation can create a circular workload
One manager can use AI to expand a short intention into a long email, while the recipient uses AI to reduce that email back to bullet points. Both have adopted the tool, but the communication process has not improved. A direct conversation or a concise original message may solve the actual problem with less work, so the first design question should concern the root cause.
Watch from 8:03Efficiency changes when everyone participates
Email once removed trips to the post office and shortened response times, creating obvious value for early users. When the whole organization adopted it, inbox volume consumed much of that saved time. AI may follow a similar pattern: present advantages are real, but managers need feedback loops that reveal how their own usage and the surrounding workload evolve as adoption becomes universal.
Watch from 10:00Explore WaysConf 2026
