The AI-savvy operating model - Matthew Skelton, Conflux - Agile to Agility conference

Slide 1

Slide 1 text

The AI-savvy operating model towards a humane, effective, financially-transparent way of working for AI-enhanced knowledge work Matthew Skelton, Conflux - co-author of Team Topologies Agile to Agility conference | 2024-12-04 K38 Photo by Barbara Zandoval on Unsplash

Slide 2

Slide 2 text

2 ❤

Slide 3

Slide 3 text

3 Matthew Skelton holistic innovation Originator of Adapt Together™ by Conﬂux Co-author of Team Topologies matthewskelton.com

Slide 4

Slide 4 text

Small, scrappy startups that use GenAI well will be able to out-innovate larger incumbents [initially[. So what would an eﬀective “AI-savvy” operating model look like that used AI to reduce time-to-value for larger organizations? 4

Slide 5

Slide 5 text

(we already know) 5

Slide 6

Slide 6 text

6 ● Empowered teams ● No hand-oﬀs ● Ongoing stewardship ● Clear boundaries ● Active knowledge diﬀusion

Slide 7

Slide 7 text

The near-future AI advantage (?) My perspective on AI (and work) How to trust and organize AI The AI-savvy operating model 7

Slide 8

Slide 8 text

The near-future AI advantage (?) 8

Slide 9

Slide 9 text

9 “AI-savvy” operating model

Slide 10

Slide 10 text

10 Traditional AI: pattern matching, huge data volumes, temporal correlation Generative AI: next-token guess, content generation, option generation

Slide 11

Slide 11 text

11 “Learning” but no understanding with AI

Slide 12

Slide 12 text

12 Savvy: don’t pretend or forget that AI doesn’t understand ⚠

Slide 13

Slide 13 text

13 Savvy use cases for AI: ● Speed up content generation (incl. code) and workﬂows ● Augment human decision-making and options ● Agent-based processes ● …

Slide 14

Slide 14 text

14 Do like AI and look backwards!

Slide 15

Slide 15 text

15 Cloud birthed SaaS, leaving older models behind (likewise GenAI)

Slide 16

Slide 16 text

16 Agentic AI - devise 230 new product variations and: ● Generate code, deploy ● Test with synthetic users ● Find the best combination ● Data-driven product ﬁt in hours A/B testing on steroids!

Slide 17

Slide 17 text

17 but…

Slide 18

Slide 18 text

18 LLMs approaching limits?

Slide 19

Slide 19 text

19 Serious security problems

Slide 20

Slide 20 text

20 Data Science Data Eng ML Ops

Slide 21

Slide 21 text

21 https://web.devopstopologies.com/ 2013

Slide 22

Slide 22 text

My perspective on AI (and work) 22

Slide 23

Slide 23 text

23 1998 My ﬁrst “AI” system: backpropagation neural network BSc, University of Reading

Slide 24

Slide 24 text

24 https://www.linkedin.com/in/promarkbishop/

Slide 25

Slide 25 text

25 2000 MSc in Neuroscience at the University of Oxford Research into dyslexia and Alzheimer’s disease

Slide 26

Slide 26 text

26 2001 Software for MRI brain imaging machines

Slide 27

Slide 27 text

27 2011 Organizational architecture and new techniques & tools for adopting cloud and Continuous Delivery at Trainline (UK)

Slide 28

Slide 28 text

28 https://blog.matthewskelton.net/2013/10/22/what-team-s tructure-is-right-for-devops-to-ﬂourish/ 2013

Slide 29

Slide 29 text

29 https://web.devopstopologies.com/ 2013

Slide 30

Slide 30 text

Team Topologies Organizing business and technology teams for fast ﬂow Matthew Skelton & Manuel Pais IT Revolution Press, September 2019 Order via stores worldwide: teamtopologies.com/book 175k copies sold to date 30

Slide 31

Slide 31 text

Slide 32

Slide 32 text

32 Delivering at speed means we need to organize for fast ﬂow - “sociotechnical”

Slide 33

Slide 33 text

How to trust and organize AI 33

Slide 34

Slide 34 text

34 Assume we have a collection of AI agents responding to … some kind of input from humans…

Slide 35

Slide 35 text

35 Savvy: don’t pretend or forget that AI doesn’t understand ⚠

Slide 36

Slide 36 text

36 Humans set the context, goals, guardrails, and execution constraints in a repeatable and traceable way (programming)

Slide 37

Slide 37 text

37 How would we be able to trust a set of AI agents?

Slide 38

Slide 38 text

38 How are we be able to trust a set of people?

Slide 39

Slide 39 text

39 ● Guardrails ● Ongoing domain context ● Good boundaries ● Ongoing stewardship and responsibility

Slide 40

Slide 40 text

40 Teams of humans and AI agents with context, guardrails, stewardship, boundaries, etc…

Slide 41

Slide 41 text

41 Good boundaries for: ● Context windows ● Security ● Resilience

Slide 42

Slide 42 text

The AI-savvy operating model 42

Slide 43

Slide 43 text

43 Cloud removed IT infrastructure as a blocker

Slide 44

Slide 44 text

44 AI removes typing, generating options, trying out combinations, etc. as blockers

Slide 45

Slide 45 text

45 Humans understand the organizational intent and specify guardrails & goals

Slide 46

Slide 46 text

Principles from Team Topologies provide guidance for knowledge work 46

Slide 47

Slide 47 text

47 Multiple, independent ﬂows, fractally

Slide 48

Slide 48 text

Respect Conway’s Law (aka ‘sociotechnical mirroring’) 48

Slide 49

Slide 49 text

Clear ongoing stewardship of services and systems 49

Slide 50

Slide 50 text

Stream-aligned teams have end-to-end responsibility for a service (You Build It, You Run It) 50

Slide 51

Slide 51 text

Platforms improve ﬂow and reduce extraneous cognitive load 51

Slide 52

Slide 52 text

Teams are small (~9), slowly changing, with ‘aligned autonomy’ 52

Slide 53

Slide 53 text

Teams are empowered to sense and adjust boundaries to improve ﬂow on a frequent basis 53

Slide 54

Slide 54 text

54 With knowledge work, we’re fundamentally concerned with the ﬁdelity of representation of intent (in code, writing, etc.), so ongoing domain knowledge, clear boundaries, and stewardship are essential.

Slide 55

Slide 55 text

55 Architecture for fast ﬂow resembles an ecosystem of loosely-coupled independently-viable services with clear boundaries and ownership aligned to the ﬂow of business value.

Slide 56

Slide 56 text

56 3EO: Entrepreneurial Ecosystem Enabling Organizing https://www.boundaryless.io/3eo-framework/

Slide 57

Slide 57 text

How might AI tools help leadership decision-making? 57

Slide 58

Slide 58 text

TeamOS 58 teamos.is Disclosure: Matthew Skelton has invested personally in the company behind TeamOS

Slide 59

Slide 59 text

59 TeamForm teamform.co

Slide 60

Slide 60 text

60 CodeScene codescene.com

Slide 61

Slide 61 text

61 “The work is delivered in many small changes that are uncoordinated to enable ﬂow. … Management’s job is to provide context, prioritization and to coordinate across teams. Lending resources if needed across teams to unblock things. … It works well within a high trust culture.” Adrian Cockcroft https://mastodon.social/@adrianco/111174832280576410 Technology strategy advisor, Partner at OrionX.net (ex Amazon Sustainability, AWS, Battery Ventures, Netﬂix, eBay, Sun Microsystems, CCL)

Slide 62

Slide 62 text

If we have clear boundaries for ﬂow, with limited interactions, how do we create alignment? How do we learn from each other at pace? 62

Slide 63

Slide 63 text

63 Multiple, independent ﬂows, fractally

Slide 64

Slide 64 text

64 Active diﬀusion of knowledge across team boundaries

Slide 65

Slide 65 text

66 Future: AI-powered tools to detect duplication, uncertainty, waste, innovation, etc. and diﬀuse learning

Slide 66

Slide 66 text

67 internal conferences guilds Communities of Practice lunch & learn public blogs

Slide 67

Slide 67 text

68 https://internaltechconf.com/ Internal Tech Conferences Victoria Morgan-Smith and Matthew Skelton

Slide 68

Slide 68 text

69 “This initiative around internal conferences has been the single most eﬀective thing to align business and technology that I have seen in this organization” – Murray Hennessey, CEO, (UK retail co)

Slide 69

Slide 69 text

70 https://internaltechconf.com/ Internal Tech Conferences Victoria Morgan-Smith and Matthew Skelton It’s o p te !

Slide 70

Slide 70 text

71 “The way that the Conﬂux crew used their active knowledge diﬀusion approach to seek out and champion good practices was a real revelation to us at TELUS and helped to shift thinking around how we innovate and share successes.” – Steven Tannock, Director, Architecture (Platform Technology & Tools) at TELUS Digital

Slide 71

Slide 71 text

72 Thriving organizations, delivering at speed™ Create alignment, trust, and engagement across your organization whilst delivering at pace with fast ﬂow. adapttogether.info

Slide 72

Slide 72 text

73 Fast flow organizations need active knowledge diffusion across flow boundaries to create trust, alignment, and learning

Slide 73

Slide 73 text

The near-future AI advantage (?) My perspective on AI (and work) How to trust and organize AI The AI-savvy operating model 74

Slide 74

Slide 74 text

Slide 75

Slide 75 text

76 ● Empowered teams ● No hand-oﬀs ● Ongoing stewardship ● Clear boundaries ● Active knowledge diﬀusion

Slide 76

Slide 76 text

Let’s do what it takes to empower teams of humans + AI agents to make decisions quickly, safely, and compliantly with high-ﬁdelity domain knowledge and visibility of the data sources & results, plus the agency to address problems caused. 77