Platform Engineering Reality Check: From Big Dreams to Practical Steps

Platform engineering journeys often start with ambitious visions of seamless developer experience and full self-service capabilities. The reality, however, is typically far more incremental.

In this talk, Paula will share three real world case studies highlighting the tangible value platforms can deliver while also showing that successful platforms are not built overnight. They begin with practical, sometimes unglamorous or boring steps, and the most impactful initiatives are guided by clear metrics, continuous user feedback, and a strong focus on developer needs.

Drawing on examples from across different industries, this session describes the common patterns, trade-offs, and pitfalls teams encounter on their platform journey. Attendees will leave with practical advice on where to start, how to measure progress, and how to turn big platform dreams into achievable, sustainable outcomes.

Transcript

1. @PaulaLKennedy Platform Engineering Reality Check: From Big Dreams to Practical Steps paulalkennedy @paulalkennedy.bsky.social paula@syntasso.io Paula Kennedy (She / Her)
2. @PaulaLKennedy Key Takeaways • Who? What? How? • Build, ship, learn (perfection is the enemy of progress) • Celebrate the wins, drive demand for your platform Your platform dreams will become reality!
3. @PaulaLKennedy Paula Kennedy (She / Her) • Co-founder and COO, Syntasso • Organiser: ◦ FastFlowConf UK ◦ Kubernetes Community Days UK ◦ London Platform User Group (LoPUG) • Champion: ◦ OpenUK Ambassador ◦ CNCF Ambassador ◦ Team Topologies Advocate CREATED BY Who am I ?
4. @PaulaLKennedy What is Platform Engineering? Source: https://www.gartner.com/en/articles/what-is-platform-engineering
5. @PaulaLKennedy What’s the challenge? Photo by JD Hancock
6. @PaulaLKennedy What’s the challenge? Source: https://landscape.cncf.io/
7. @PaulaLKennedy What does good look like? Source: https://www.youtube.com/live/jJSo3kdflpA?si=BJHsDk20PrilKqlS “Platform as a Product is what happens when you apply a product mindset to your internal platform!” Paula Kennedy
8. @PaulaLKennedy What does good look like? • Who are you building the platform for? • Build something: Minimum Viable Product • Get feedback • Iterate
9. @PaulaLKennedy Case Study 1: NatWest Group Source:https://www.youtube.com/watch?v=79o52DAZT_k&list=PLj6h78yzYM2MeuSNq pcDl-qdMeCe6vp6p&index=5 😭 THE CURRENT PROBLEM: • Dev teams had high cognitive load and inconsistent deployments • Slow, manual processes, weeks to stand up a Kubernetes cluster
10. @PaulaLKennedy Case Study 1: NatWest Group Source:https://www.youtube.com/watch?v=79o52DAZT_k&list=PLj6h78yzYM2MeuSNq pcDl-qdMeCe6vp6p&index=5 😭 THE CURRENT PROBLEM: • Dev teams had high cognitive load and inconsistent deployments • Slow, manual processes, weeks to stand up a Kubernetes cluster 🌈 THE DREAM END STATE: • Standardised, self-serve environment • Accelerated Developer Experience
11. @PaulaLKennedy Case Study 1: NatWest Group Source:https://www.youtube.com/watch?v=79o52DAZT_k&list=PLj6h78yzYM2MeuSNq pcDl-qdMeCe6vp6p&index=5 😭 THE CURRENT PROBLEM: • Dev teams had high cognitive load and inconsistent deployments • Slow, manual processes, weeks to stand up a Kubernetes cluster 🌈 THE DREAM END STATE: • Standardised, self-serve environment • Accelerated Developer Experience 👣 THE FIRST STEP: • Created MVP with Backstage, Kratix and Flux using GitOps • Reduced time to deploy Kubernetes cluster on AKS from 10 weeks to 35 mins
12. @PaulaLKennedy Case Study 1: NatWest Group Source:https://www.youtube.com/watch?v=79o52DAZT_k&list=PLj6h78yzYM2MeuSNq pcDl-qdMeCe6vp6p&index=5 😭 THE CURRENT PROBLEM: • Dev teams had high cognitive load and inconsistent deployments • Slow, manual processes, weeks to stand up a Kubernetes cluster 🌈 THE DREAM END STATE: • Standardised, self-serve environment • Accelerated Developer Experience 👣 THE FIRST STEP: • Created MVP with Backstage, Kratix and Flux using GitOps • Reduced time to deploy Kubernetes cluster on AKS from 10 weeks to 35 mins 🚀 WHAT COMES NEXT: • Iterate on platform to manage more workloads • Begin exploring AI platform patterns
13. @PaulaLKennedy Case Study 2: The Access Group Source:https://www.syntasso.io/post/how-the-access-group-scaled-ai-driven-developm ent-with-platform-engineering 😭 THE CURRENT PROBLEM: • Hosting & managing the increasing number of apps being created by AI • Need consistency, security, scalability and reliability
14. @PaulaLKennedy Case Study 2: The Access Group Source:https://www.syntasso.io/post/how-the-access-group-scaled-ai-driven-developm ent-with-platform-engineering 😭 THE CURRENT PROBLEM: • Hosting & managing the increasing number of apps being created by AI • Need consistency, security, scalability and reliability 🌈 THE DREAM END STATE: • Platform to provide “constructive constraints” to empower AI development • Expand usage of “Evo Builder” to develop applications faster across org
15. @PaulaLKennedy Case Study 2: The Access Group Source:https://www.syntasso.io/post/how-the-access-group-scaled-ai-driven-developm ent-with-platform-engineering 😭 THE CURRENT PROBLEM: • Hosting & managing the increasing number of apps being created by AI • Need consistency, security, scalability and reliability 🌈 THE DREAM END STATE: • Platform to provide “constructive constraints” to empower AI development • Expand usage of “Evo Builder” to develop applications faster across org 👣 THE FIRST STEP: • Run a global hackathon to test the platform capabilities • “Feedback loop has genuinely impressed the team”
16. @PaulaLKennedy Case Study 2: The Access Group Source:https://www.syntasso.io/post/how-the-access-group-scaled-ai-driven-developm ent-with-platform-engineering 😭 THE CURRENT PROBLEM: • Hosting & managing the increasing number of apps being created by AI • Need consistency, security, scalability and reliability 🌈 THE DREAM END STATE: • Platform to provide “constructive constraints” to empower AI development • Expand usage of “Evo Builder” to develop applications faster across org 👣 THE FIRST STEP: • Run a global hackathon to test the platform capabilities • “Feedback loop has genuinely impressed the team” 🚀 WHAT COMES NEXT: • Since hackathon, platform is now hosting over 2,000 applications • Drive usage of AI, use platform to safely manage and scale
17. @PaulaLKennedy 😭 THE CURRENT PROBLEM: • Slow, manual processes, high platform friction • Extended development lead times, high operating costs & compliance risks Case Study 3: Financial Services Org Source: written case study coming soon (hopefully)
18. @PaulaLKennedy 😭 THE CURRENT PROBLEM: • Slow, manual processes, high platform friction • Extended development lead times, high operating costs & compliance risks 🌈 THE DREAM END STATE: • Platform to support growth strategy - global scaling, market expansion • Reduce technical debt, lower operational overheads Case Study 3: Financial Services Org Source: written case study coming soon (hopefully)
19. @PaulaLKennedy 😭 THE CURRENT PROBLEM: • Slow, manual processes, high platform friction • Extended development lead times, high operating costs & compliance risks 🌈 THE DREAM END STATE: • Platform to support growth strategy - global scaling, market expansion • Reduce technical debt, lower operational overheads 👣 THE FIRST STEP: • POV to deploy one application: reduced time from 4 weeks to 6 minutes • Learn from steps taken, iterate Case Study 3: Financial Services Org Source: written case study coming soon (hopefully)
20. @PaulaLKennedy 😭 THE CURRENT PROBLEM: • Slow, manual processes, high platform friction • Extended development lead times, high operating costs & compliance risks 🌈 THE DREAM END STATE: • Platform to support growth strategy - global scaling, market expansion • Reduce technical debt, lower operational overheads 👣 THE FIRST STEP: • POV to deploy one application: reduced time from 4 weeks to 6 minutes • Learn from steps taken, iterate 🚀 WHAT COMES NEXT: • Second stage pilot to prove repeatable patterns, deliver consistently • Prove patterns, scale platform to more teams globally Case Study 3: Financial Services Org Source: written case study coming soon (hopefully)
21. @PaulaLKennedy Key Takeaways • Who? What? How?
22. @PaulaLKennedy Key Takeaways • Who? What? How? • Build, ship, learn (perfection is the enemy of progress)
23. @PaulaLKennedy Key Takeaways • Who? What? How? • Build, ship, learn (perfection is the enemy of progress) • Celebrate the wins, drive demand for your platform
24. @PaulaLKennedy Your platform dreams will become reality! Photo by JD Hancock
25. @PaulaLKennedy Want to know more? ME & ROSIE WHAT I DO! syntasso.io paulalkennedy @paulalkennedy.bsky.social paula@syntasso.io
26. @PaulaLKennedy References / links All photos by JD Hancock: photos.jdhancock.com Blogs / websites: https://www.syntasso.io/ https://www.gartner.com/en/articles/what-is-platform-engineering https://www.syntasso.io/post/platform-engineering-orchestrating-applications- platforms-and-infrastructure https://landscape.cncf.io/ https://www.syntasso.io/case-studies/natwest-uses-kratix-to-reduce-developer- cognitive-load-and-enable-platform-cocreation https://www.syntasso.io/post/how-the-access-group-scaled-ai-driven-developm ent-with-platform-engineering Videos: “Crossing the Platform Gap, Paula Kennedy, Devoxx UK, 2022 https://youtu.be/pAk5GReIs90?si=90g8kStzkALnYz5Z “The Ins and Outs of Delivering your Platform as a Product“, Paula Kennedy, 2022, https://www.youtube.com/live/jJSo3kdflpA?si=BJHsDk20PrilKqlS Conferences / Meetups: https://devopsdays.org/events/2026-london/welcome/ https://kcduk.io https://www.meetup.com/london-platform-user-group-lopug/ https://www.fastflowconf.com/
27. @PaulaLKennedy THANK YOU! HELPING PLATFORM TEAMS THRIVE