The Imperfect Architect - How to survive as an architect in the digital age.

Software architects often have to navigate a complex environment, wrestling with legacy systems, battling organizational politics, explaining technical choices to management, dealing with team dynamics and - above all - chasing the perfect architecture. We've learned that building trust is paramount, and that the "last responsible moment" is the best time to make a decision but apparently also just one letter away from the “least responsible moment”. But we should have no fear as AI Agents are here to fix all this - right… right?

Join this talk as we explore how to thrive as an architect in this difficult environment. From leveraging political capital to navigating organizational resistance, to effectively communicating architecture decisions. We'll discuss how to build trust, learn to lead with a situational mindset, adapt your toolkit, and accept that the most resilient architectures are often perfectly flawed. Because for the imperfect architect, nurturing collective intelligence and continuous learning isn't just a goal - it's the only way forward.

1. The Imperfect
Architect
Devoxx 2025 Robin Custers

2. 2

3. 3
My dad is an
IT architect
He is just
in calls all
the time

4. 4

5. The Architect

6. 7
2012
The Mayan Apocalypse Gangnam Style

7. 8

8. 9

9. 10

10. “Do you want to be our new architect?

11. “You will get more responsibility

12. 13
That I am now responsible for
everything that went wrong?
Does having more responsibility
actually mean

13. How to be a good IT architect?

14. “
An experienced architect
Someone who has messed up so many
times that they just know what works and
what doesn’t

15. arkhitekton
CraftsmanLeader

16. Technical Decisions

17. 60K
Year 1
50K
Year 2
40K
Year 3
35K
Year 1
45K
Year 2
55K
Year 3

18. Going to the movies
19
€ 20
Would spend remaing €20
on a ticket?
No

19. Going to the movies
20
Would spend remaing €20
on a ticket?
Yes

20. 21
€ 20
The lost ticket The lost bill
€ 20
No Yes

21. 22
System 2
•Slow and deliberate
•Logical and analytical
•Monitors and overrides
System 1
System 1
•Fast, automatic, intuitive
•Mental shortcuts, patterns,
past experiences
•Rapid judgements
•Emotional and prone to
bias
The Human Mind

22. 23
How to make better decisions?

23. 24
Confirmation Bias

24. 25
Trade-Off Analysis

25. 26
Pareto Principle
Avoid too much focus on edge cases

26. 27
Last Responsible Moment

27. 28
Tracer Bullets

28. Child Benefit System
Determine
monthly amount
Manage
personal data
Pay
Generate and
send documents
Manage debts
Integrations with
external systems
Accounting
Reporting
SecurityPerformance
…

29. Child Benefit System
Determine
monthly amount
Manage
personal data
Pay
Manage debts
Integrations
with external
systems
Accounting
Reporting
Security
…
Generate and
send documents
Performance

30. 31
De-risk early!

31. 32
Innovation Addition

32. 33
TDD
BDD
DDD
FDD
Three Letter Abbreviations
FashionDrivenDevelopment

33. Provide the best
ROI
Within the given time & budget constraints

34. 35

35. arkhitekton
CraftsmanLeader

36. 37
Team

37. Production
Application
Application
Application
Application
mvn clean install
<plugin>
<!--Used to automatically drop (if any) and create a
database prior to running integration test cases. -->
<groupId>org.codehaus.mojo</groupId>
<artifactId>sql-maven-plugin</artifactId>
Application

38. System

39. Directing
“Enthusiastic Beginner”
Little to no experience
Very excited & motivated
Give clear and specific instructions,
supervise closely
Coaching
“Disillusioned Learner”
Basic experience
Lacks confidence or motivation
Leader makes decisions, encourage and
explain “why”
Supporting
“Capable but cautious”
High experience
Lacks confidence or motivation
Collaborate, shared decision making, power
shifts to team member
Delegating
“Self-reliant achiever”
High experience
High confidence and motivation
Give autonomy, guide when asked
Trust
Agreement
Low High
Low
High
Directive behaviour
Supportive
behaviour
Situational Leadership
You can do it. You’ve got this.
This is what you need to do.
This is how you need to do it.
I’m here if you need me. Let me
know what I can do to help you
to achieve our goals.
This is the goal we should reach.
Can you make sure we get
there?

40. 42
Organisation

41. 43
Politics

42. 44
Mainframe EOL application Modernnew application
Architect’s

43. I don’t believe in the proposed solution.
I don’t think it is a good solution.
It will not work.
Can you explain why?
We really made this with our best people.
Can you please explain why?

44. It does not fit the requirements.
I don’t have time for that!!!
You made the solution.
You should fix it! I have more important things to do!
Ok, what do you mean exactly?
Can we go over it together?
Ok but then tell me what
I should fix in the solution.

45. It is not just technical…
the people
dimension

46. You will meet different persons…
Bu Trust
Agreement
Low High
Low
High
Bedfellows Allies
OpponentsAdversaries
(*) Peter Block
-The Empowered Manager
Agreement

47. “The accumulated goodwill, trust, and influence that
you possess, which can be leveraged to achieve
objectives or influence decision-making”
Political Capital

49. 51
If you want to influence
Keep the balance of your political
capital bank account in mind

50. “
An experienced architect
Someone who has messed up so many
times that they just know what works and
what doesn’t

51. 53
Takeaways!
• When deciding always evaluate different options and
their trade-offs.
• Avoid FDD and provide best ROI!
• Lead your team with a situational mindset
• Trust gives you influence
• Mind the balance of your political capital bank account
• Use common sense

52. 54
The perfect architect is like the Loch Ness Monster
Rumoured to exist, but never actually been seen.

53. 55
Have a great Devoxx!
Robin Custers
Architect
https://www.linkedin.com/in/robincusters