Jeremy Vyska | #MSDYN365BC MVP
banner
bc.jeremy.vyska.info
Jeremy Vyska | #MSDYN365BC MVP
@bc.jeremy.vyska.info
950 followers 120 following 190 posts
๐Ÿ‡ธ๐Ÿ‡ช/๐Ÿ‡บ๐Ÿ‡ธ in Gothenburg Sweden. #MSDYN365BC MVP. He/him. Autistic/ADHD. Dad-joker. Problem solver. Lgbtqia+ Supporter. This is my business world profile and I'll mostly be engaging in work topics here.
Posts Media Videos Starter Packs
Pinned
๐Ÿš€ After building BC extensions and seeing the same development challenges repeatedly, I'm excited to share our approach to comprehensive BC development intelligence.

๐Ÿงต
Pretty low feedback so far, will likely push it through today and then when it's live, people will probably have all sorts of opinions ๐Ÿ˜‰
Reposted by Jeremy Vyska | #MSDYN365BC MVP
Docs, AL Language changelog, BCLE videos, and what not ... so much to learn about what's new in BC 27 for developers!

But don't worry - I've combined it all-in-one in my new blog:

๐——๐—ผ๐—ฐ๐˜€: ๐—ช๐—ต๐—ฎ๐˜โ€™๐˜€ ๐—ป๐—ฒ๐˜„ ๐—ถ๐—ป ๐—•๐—– ๐Ÿฎ๐Ÿณ ๐—ณ๐—ผ๐—ฟ ๐—ฑ๐—ฒ๐˜ƒ๐—ฒ๐—น๐—ผ๐—ฝ๐—ฒ๐—ฟ๐˜€
nataliekarolak.wordpress.com/2025/10/02/d...
Docs: Whatโ€™s new in BC 27 for developers
Summary of all technical changes in Business Central 2025 release wave 2 (BC 27 / runtime 16.0) mentioned in the docs and beyond.
nataliekarolak.wordpress.com
With the tech changing every week, one thing stays the same.

I hate waiting on progress bars.

@bctechdays.com
I'm most excited about the opportunity for all of us to test this together and refine it based on real BC development scenarios. There's a lot we can learn from community feedback.

What BC development challenges would benefit from this kind of systematic guidance? I'd love to hear your thoughts.
โ€ข Layer system for company/team customization
โ€ข One-click VS Code integration through MCP

๐ŸŽฏ Ready to help test it? Find out more here:
nubimancy.com/2025/09/22/...

๐Ÿงต
โœจ What we've included so far:
โ€ข 100+ BC knowledge topics from ALGuidelines and community best practices
โ€ข 14 AI specialists for different BC development areas (architecture, performance, security, etc.)
โ€ข Structured methodologies for common workflows (optimization, testing, integration)
๐Ÿงต
After 30 days of AI-accelerated development, we've built BC Code Intelligence - a system that brings BC domain expertise directly into VS Code. This works well for our development patterns, though there's definitely room for community testing and improvement.

๐Ÿงต
๐Ÿš€ After building BC extensions and seeing the same development challenges repeatedly, I'm excited to share our approach to comprehensive BC development intelligence.

๐Ÿงต
I'm curious about your experiences: Have you seen patterns where generic AI guidance actually made things worse? What's worked best in your domain?

I've even added instructions about how you could play with it to test different models - tell me what you find!
I've open-sourced the complete testing framework because I'd rather see more people succeed with AI knowledge engineering than keep this approach to myself.

Full methodology and results: nubimancy.com/2025/09/09/...

๐Ÿงต
The results surprised me. Generic programming knowledge actually _hurt_ performance - we're talking a 4% regression. But when I got the knowledge engineering right? One module went from 64 seconds to 81 milliseconds. That's 746x faster.

๐Ÿงต
Like many of you working with AI tools, I keep hearing about "feeding knowledge to AI," but I hadn't seen systematic testing of whether it works - and more importantly, which approaches work better than others. So I built a controlled experiment with five different knowledge engineering tiers.

๐Ÿงต
I just finished an intense week of testing something I've been curious about: whether different approaches to "knowledge engineering" actually improve GitHub Copilot's effectiveness with Business Central development.

๐Ÿงต
This wraps up my 3-part AI transformation series - from skepticism to organizational enthusiasm.

The approach is replicable, but infrastructure and governance matter more than the AI itself.

Full post: nubimancy.com/2025/08/26/...
If you're working on AI adoption in your org, curious about your experience:

Have you hit the "individual success but team adoption struggles" wall?

What's worked for making AI accessible to everyone, not just the early adopters?

๐Ÿงต
The formula that emerged:

Without all three elements = individual productivity improvements at best

With all three elements = teams actively requesting AI integration on every project

๐Ÿงต
The most surprising part? Once we had all three elements working together, team transformation happened in about one week.

Despite technical hiccups (VS Code versions, MCP setup), they went from "can we make this work?" to "can we roll this everywhere?"

๐Ÿงต
Business Governance solved the "AI enthusiasm vs business reality" problem.

AI agents would jump into implementation even when requirements had obvious gaps needing stakeholder input. Quality Gates force pause points for business validation.

๐Ÿงต
Knowledge Architecture was the game-changer though. We created ".aidocs" folders - documentation designed specifically for AI consumption, not humans.

AI agents need context about WHY decisions are made, not just WHAT to implement.

๐Ÿงต
Technical Infrastructure meant zero-friction access. If team members need to become AI experts or configure complex environments, adoption stalls at early adopters.

We built a "Bootstrap" system - one VS Code task activates AI guidance in any repository.

๐Ÿงต
The breakthrough came when I stopped thinking about AI adoption and started thinking about AI infrastructure.

Three specific elements made the difference:

Technical Infrastructure + Knowledge Architecture + Business Governance

๐Ÿงต
Here's what I discovered: Individual AI mastery and organizational AI transformation are completely different challenges.

You can be great at using AI personally and still completely fail at scaling those benefits to your team.

๐Ÿงต
This is Part 3 of my AI transformation series - and the gap between individual AI success and team transformation? Way more complex than I expected.

๐Ÿงต
Three months ago I was sitting in conferences wondering "why does everyone think AI is transforming business when it's basically fancy autocomplete for me?"

Today my team actively wants our AI infrastructure rolled out to every project.

๐Ÿงต