Steve Endow's Business Central Podcast

I Need Coffee - Episode 220 - Weekly BC Review! (August 14, 2026)
1. Claude Code Auto-Approval
   - Anthropic now recommends the auto-approval setting in the VS Code extension.
   - Host previously avoided it after bad experiences with GitHub Copilot.
   - Switched after constant trivial “Can I run this?” approvals became meaningless.
   - Caution: in auto mode the agent can independently decide to commit, push, and open a draft PR even when you only asked a question. Stay vigilant.

2. .NET MAUI Passphrase Generator Rewrite
   - Legacy .NET Framework passphrase generator (manual typing of long random passwords is painful) rewritten in .NET MAUI for Windows x64 / ARM64 (future macOS potential).
   - Fable/Claude one-shotted the rewrite in ~45 continuous minutes, fixed a race-condition hang-on-close, switched to proper cryptographic RNG, and bluntly criticized the original code quality.
   - Host’s rationale for multi-word passphrases: longer length can still provide strong search-space resistance despite lower entropy per character; practical cracking attempts on an RTX 3070 Ti were unsuccessful for realistic multi-word + digits + symbols combinations.

3. SpreadsheetML for Formatted Excel Exports from Business Central
   - Customer needed complex calculated reports exported as ready-to-use Excel (titles, dates, headers, freeze panes, proper column widths).
   - Claude used SpreadsheetML (Open XML) from AL — something the host had never heard of — and delivered a fully formatted file in ~2 minutes (commit + draft PR).
   - Users loved it; no more manual reformatting. Changes are now trivial (“make this cell purple”).

4. BC Telemetry Onboarding Form
   - Structured form to capture everything needed before enabling telemetry: tenant ID, version, licenses, companies, key user IDs + telemetry GUIDs, etc.
   - Recommended practice so analysis is purposeful instead of “we turned it on… now what?”.

5. ERP Academy
   - Wrapping up Week 4 of the 18+ month program (10 hours/week self-study + Monday evening group calls).
   - Breakout rooms discussed core business functions (finance, sales, purchasing, warehousing, etc.). Ronald continues heavy curriculum and facilitation work.

6. Sandbox Point-in-Time Restore Tip
   - Useful when customers repeatedly request clean sandboxes for testing.
   - Leverage Azure SQL point-in-time restore; timestamp a known-good state and restore into a new environment name as needed.
   - Limits: 10 restores per environment per month; requires a free environment slot (temporarily consumes two slots).

7. Email Delivery Mystery Solved (Barracuda)
   - Weeks of troubleshooting bounced emails to a customer who insisted “it’s not us.”
   - Claude DNS/MX analysis revealed Barracuda spam filtering at the edge. Customer had zero visibility or control.
   - Host wrote a 15-page diagnostic procedure for the customer’s IT. Adding a sender exclusion rule (for unclear reasons) finally resolved it.

8. Full Account Takeover via Auth Bypass
   - Host’s account on an old marketplace site was fully taken over (password + TOTP reset). Site later admitted a complete OAuth/authentication system weakness affecting many accounts.
   - Another data point in the rising wave of compromised web applications.

9. AIO Theory – Documentation as Marketing Asset
   - Emerging hypothesis: well-written user guides and documentation may outperform the main marketing website for AI-optimized search (AIO) results.
   - Some AI answers already cite the doc site and specific Blue Dragonfly apps.
   - Consequence: prioritising additional high-quality user guides for better AIO visibility.
   - Broader learning: modern SEO/AIO is largely about proper structure, metadata, and crawlability rather than keyword stuffing. Structured pricing data (one-time vs monthly vs annual) is a concrete example of the level of detail required.

10. Latent Bug Discovery – Unwired Setup Toggle
    - While updating a user guide, the rigorous verification prompt revealed that a setup toggle (“Allow changes to approved documents”) present since v1.0 had never been wired into any application code.
    - Survived months of development, multiple developers (host + Franklin), AI code reviews, and releases.
    - Lesson: assume every existing app may contain “skeleton” features; perform full code reviews of older codebases with no assumptions.

11. $9.5 Billion High-Precision Report
    - Complex PTE report: ≥236 dimension values, >50 rows, >50,000 calculations, totaling $9.5 billion.
    - Had to match customer’s historical Excel reports to the penny despite rounding differences.
    - Key insight: Excel’s 15 significant-digit limit (not just decimal places). Business Central’s higher precision was actually *too* precise and produced different rounding.
    - Added configurable percent-rounding precision; report now matches historical numbers exactly. Customer is discovering and correcting their own prior misclassifications.

12. Pangram AI Detection & Authenticity Discussion
    - Subscribed to Pangram (first AI-detection tool the host finds reasonably accurate). Browser extension now tags posts on X/LinkedIn.
    - LinkedIn is saturated with high-percentage AI-generated content; Microsoft marketing emails also score partially AI.
    - Open philosophical question: where is the line of authenticity? Meeting notes and wireframes feel acceptable; personal “I learned this lesson…” stories or LinkedIn engagement posts generated by AI feel disingenuous. Marketing content is more tolerated than personal voice.

13. Crypto-JS Open-Source Security Failure
    - Once-popular JavaScript crypto library (16.4k stars, 15.6 million weekly npm downloads) is discontinued, known-insecure, and still widely used.
    - Clear cryptographic implementation flaws; README lacks strong warnings. Classic illustration that “open source = secure” is outdated.

14. Team Updates
    - Joselyn: Passed MB-800 — now two certified Business Central consultants in Berlin, El Salvador (population ~15–16k). Continuing financials work.
    - Ronald: Leading ERP Academy + second English Academy cohort; starting Continia Document Capture implementations.
    - Fatima: Organizing next El Salvador Doug meetup (Spanish, Wed 19th); resuming El Coyolar Christmas festival planning (team is pushing hard for pupusas for ~200 people).
    - David: Continuing data-import challenges.
    - Franklin: App bug research and PTE work.
    - Daniela: Driving SEO/AIO, analytics, and content production.

I Need Coffee - Episode 219 - Weekly BC Review! (August 7, 2026)

1. Utah Trip & Home Maintenance (1,600+ miles / ~2,575 km)
   - Primary trigger: misbehaving sprinklers. Turned out to be a reset timer + unplugged wires; only one physical valve issue. Avoided digging 1-inch residential PVC.
   - Saw Tanya; worked from dining table, strategic planning, fresh cherries.
   - Temporary folding desk + daughter’s old IKEA chair at Airbnb proved uncomfortable → designing a sturdier, larger, closet-folding desk in SketchUp.
   - Utah “data center” upgrade: retired paint-bucket server rack; installed proper shelf/mounting plate. Also installed a clean two-zone radon abatement system (common in high-radon areas; excellent local install that preserves future basement finishing).
   - Family dinner at YGF Malatang (Chinese-style hot-pot buffet) after an 11-hour drive.

2. Marketing & SEO / AIO Progress (with Daniela)
   - Full-time marketing associate is driving accountability (“chop chop”).
   - Claude-drafted SEO + AIO project plan after scanning the site against best practices.
   - Key insight: modern SEO is mostly housekeeping (proper structure, meta tags, structured data) rather than keyword stuffing. Search engines already detect pricing but flag missing frequency (one-time / monthly / annual).
   - Set up Google Search Console and Bing Webmaster Tools; submitted sitemap.
   - Built monthly scorecard: 10 questions run across top AI search engines (Perplexity, Claude, Copilot, Gemini, GPT, etc.) and results matrixed.
   - Host presenting “Marketing Without a Marketing Team” (low-budget guerrilla approach that avoids ~$60k cash outlay) at Partner Vibe 2.0.

3. Azure Artifact Signing
   - Organization identity verification finally approved after 14 calendar days.
   - Certificate profiles are ephemeral (create → use once → expire in ~3 days), similar to Let’s Encrypt model.
   - $10/month; strongly recommended for anyone in US / Canada / UK / EU who needs code signing. Cleaner and far cheaper than traditional providers.

4. Business Central Web API Key Incident (Security)
   - Third party was given BC web API keys / app registration credentials with essentially zero security practices (keys at high risk of exposure to crawlers/bots).
   - Host treated it as a cybersecurity incident, pushed the partner to inform the customer, revoke keys, and delete the app registration.
   - Creating a formal incident report + standardized Web API Authorization form (customer + third party must sign; security procedures required before keys are issued).
   - Blog post published: “Reducing the risk of Business Central web API secrets through education” (includes the form and process).
   - Goal: document everything so the host cannot later be accused of negligence for failing to escalate.

5. Azure DevOps Quirk
   - Variable groups linked to Key Vault stopped working in one old project.
   - Root cause buried deep in Azure Resource Manager service connections at the project level. “Rotate secret” instantly fixed it. Reminder that Azure DevOps still has obscure corners.

6. Internal Tooling – Blue Dragonfly Bulk Data Import
   - Consulting team exhausted by Configuration Packages (slow, error-prone, painful corrections) and only partially satisfied with Insight Works power tool.
   - Decision: build an in-house general-journal-focused bulk import utility (will expand to master data, fixed assets, etc.).
   - Claude produced a combined ~14-page FRD/PRD. Franklin to vibe-code v1 once feedback arrives.
   - Classic “reinventing the wheel” calculus has flipped in the AI era: low cost to try building something better tailored to exact needs.

7. Claude Language Degradation / AI Slop
   - Extended rant on current Claude output: dense, unnatural phrases (“ground truth loadbearing first-class feedback-instrumented auto-enriched permissions posture,” “dormant schema,” “seed-free graceful degradation,” etc.).
   - Non-native English speakers often find it completely opaque; some others claim it “makes sense.”
   - Explanation circulating (Kungchen / Guidedon): shift from RLHF (human preference) to RLVR (Reinforcement Learning with Verifiable Rewards). Models are optimized for verifiable correctness by other models and are actually penalized for human-likable language (“alignment tax”).
   - Growing discussion that agents may develop an internal language subset separate from human-facing output.
   - Real-world example: public PR from a multi-billion-dollar software company full of unreadable AI-generated text that no human appears to have reviewed.

8. Entra ID / MFA Neglect
   - Partner still using SMS/voice verification (Microsoft has been pushing to disable this since before Sept 2025).
   - Microsoft blocks SMS and voice MFA to El Salvador numbers by default.
   - Passkeys become the default experience; Microsoft-provided SMS/voice authentication retires 1 Feb 2027.
   - Red flag on the partner’s overall security hygiene.

9. Other Observations
   - GitHub relative dates (“two weeks ago”) instead of actual timestamps remain terrible UX (must hover to see the real date).
   - Context switching across multiple Claude sessions / projects is still cognitively expensive.
   - Hand-coding is increasingly “supervising” AI-generated change requests and development plans.
   - People regularly burning through $200/month Claude Max plans are almost certainly using tokens inefficiently; deterministic scripts + smaller models (Haiku etc.) should handle bulk work.

10. Cybersecurity – Hardware Wallet RNG Failure
    - Popular “high-security” Bitcoin hardware wallet shipped firmware that replaced its RNG with a flawed open-source Python library (pseudonymous author). Fell back to deterministic test RNG.
    - Foundation of the security pyramid (true entropy) was sand. Wallets were pre-identified, sorted by balance, and swept starting hours before public disclosure.
    - Attacker appears unsophisticated (funds sent to traceable wallets rather than properly laundered). Classic lesson: cryptography is only as strong as its random-number generation.

11. Team Updates
    - Joselyn: Financials / GL work; continuing MB-800 study.
    - David: Passed MB-800; fighting customer/contact/vendor import and posting-group issues.
    - Ronald: New cohort for Berlin English Academy + ongoing ERP Academy curriculum (quizzes, enrichment, Monday calls).
    - Franklin: PTE bug fix (classic pattern: internal testing misses what a consultant/customer finds in five minutes).
    - Fatima: Experimenting with Claude + Teams Planner + VTT transcripts to auto-generate meeting notes and project docs.
    - Daniela: SEO/AIO execution and preparing first marketing campaign.
    - Eloar Christmas festival planning kicked off with local organizers.

12. Upcoming Events
    - Partner Vibe 2.0 — Provo, Utah, 21–23 September. Host presenting on marketing without a marketing team. Contact via steveendo.com for possible discount code.
    - Summit NA — Nashville, 11–15 October (summitna.com).
    - Dynamics Con Regionals: Columbus (Aug), Montreal (Sept), Frankfurt (Oct), Houston (Nov) — verify exact dates as some listings are inconsistent.

I Need Coffee - Episode 218 - Weekly BC Review! (July 24, 2026)
1. App & Product Work
   - Corporate Direct Debit BAI2 import marketplace app: on-the-fly enhancements after customer clarified real requirements (A *and* B, not just A). BAI2 parsing itself is now easy with AI; process/edge cases are the hard part.
   - New BAI2 user guide auto-generated, versioned (v005), committed to repo, and published nightly by the documentation site.
   - Business Central ↔ bill.com integration (with MEM support): new feature added via Fable/Claude to work around a bill.com limitation. Full docs and user guide updated.
   - Customer PTEs reviewed with partner: complex line-level sales price lists (customer price group, multiple levels — not standard BC lowest-price logic) and BOM explosion that retains parent-child relationships, original assembly price, and quantity cascading. Feature-complete and well received.
   - bill.com marketplace app sandbox cleanup: scrubbing customer data for a usable demo/test environment (hard with most SaaS platforms).

2. DevOps & Infrastructure
   - Switching from BC Container Helper / Algo-for-GitHub (painful artifact/image downloads, 30–60+ min pipelines, >2,000 GitHub Actions minutes burned in <2 weeks on one app) to Cosmo Pack / Cosmo Alpaca (target ~5 min runs). Now enabled on GitHub Enterprise orgs.
   - Code-signing certificate move: DigiCert reissue triggered organization re-verification (messy process, outdated Skype references, wrong product type in emails, high cost ~$700). Evaluating Azure Artifact Signing ($10/month, available in US/Canada/EU/UK, GitHub + Algo integration, automated identity verification via Microsoft Authenticator + third-party ID service with face/driver’s license scan).
   - New ERP Center of Excellence website (site #4): rebuilt with Claude Code CLI + Astro 6 + CSS + GitHub Actions → Cloudflare Pages. Slideshow, more photos/content (still trimming), sub-200 ms loads, auto-publish in ~60 seconds. Original 2023 Namecheap free builder site retired.

3. AI Tooling Observations & Quality
   - Fable/Claude highly effective for complex feature work + docs once prompted properly; Anthropic limit resets and team credits helped usage.
   - Claude often produces dense technobabble / “AI meta language.” Explicit prompts for plain English, no jargon/metaphors/quips dramatically improve results (occasional residual “Honest answer, Steve…” style slips remain).
   - Marketing research paper example: first AI draft was abstract-heavy and audience-mismatched; rewritten in simple language for a new marketing associate / ESL reader.
   - Persistent missing feature across AI tools: no timestamps on messages/prompts/responses. Host implemented a workaround via Claude settings.json hooks (user-prompt-submit + stop) + PowerShell, forcing the LLM to emit timestamps. Works better in CLI than VS Code extension; still a duct-tape solution.
   - Quality-noticing exercises: Joshua Bell Stradivarius subway video (most people walk past excellence); Blue Dragonfly logo deliberately distorted ±15% aspect ratio — once seen, similar squash/stretch errors appear everywhere on signs, menus, vehicle wraps, etc.

4. Bugs & Technical Discoveries
   - Obscure latent bug (Franklin): removing a permission set causes one specific event subscriber to silently fail while everything else works. Investigating registration library vs. permission-set interaction.
   - Learned about Permission Set Extension objects (BC 2021 Wave 1). Caution advised for extending base/ISV sets; safer for own library apps.
   - Microsoft Teams Planner task chat: appears broken (messages not visible until refresh / one-way visibility). Described as “duct tape all the way down” on an eventually-consistent distributed service that has been passed around Microsoft orgs.

5. Cybersecurity Notes
   - Nextcloud site compromised via WordPress vulnerability (public PSA: WordPress + plugins have had >2,000 vulnerabilities in recent years; treat as high-risk for anything important). Host recommends static sites (Astro → Cloudflare) instead.
   - Researcher “fried rice” / Chaoan / Shiao Shu (who previously found Claude source code in an npm package) is using Kimmy K3 (open-weight, minimal guardrails) to discover zero-days rapidly (e.g., Redis, Telegram Desktop/iOS one-step-from-RCE, calculator pop in minutes with multi-agent setups).
   - GitHub paid a $100k bounty (via HackerOne) for an RCE on github.com.

6. Team Updates
   - Joselyn: continuing financials / GL data-load config packages; sat MB-800 again (failed on purchasing/inventory sneaky questions); studying fixed assets, posting groups, deferrals.
   - David: check report work + continued studying.
   - Ronald: heavy lift on Berlin ERP Academy curriculum (filling gaps in AI-generated skeleton materials — adding visuals, screenshots, videos, task-level assignments for LMS); also running second round of Berlin English Academy for E4CC English scholarship graduates (practice speaking).
   - Franklin: permission-set bug research; starting a relatively small PTE tomorrow; user-guide updates.
   - Fatima: supporting Berlin ERP Academy; finalizing next El Salvador Doug meetup (around Aug 19); Teams Planner exploration.
   - Daniela: marketing pivot from traditional campaign/metrics approach to gradual brand-building; aiming for first campaign next week. Office visitor: a mouse crawled on her during a call.
   - Overall office wildlife in Berlin, El Salvador includes cats, dogs, birds, mice, etc.

7. Upcoming Events & Personal
   - Vive 2.0 — Provo, Utah (September); host speaking (two sessions). ~45–60 min south of Salt Lake City.
   - Summit / conference — Nashville, October 11; book hotel/flights/register.
   - Dynamics Con regionals: Ghana (July 25 — check remaining availability), Columbus (August 25 — double-check dates).
   - Personal: early Friday ~12-hour / 700+ mile drive to Utah (Heber / Park City area) for house maintenance + time with Tanya Henderson. Forecast high 90s °F at elevation; scenic drive past Jordanelle Reservoir.

I Need Coffee - Episode 217 - Weekly BC Review! (July 17, 2026)
- Neighborhood animal chaos
  Helped trap a stray cat with severe mange for the local shelter (TNVR process). Separately, a pack of large collared dogs has been roaming the area twice in two months, drawing multiple sheriff’s cars, motorcycle cops, and even a helicopter after one of the dogs bit someone.

- ERP Academy launch
  Created a full 18+ month Business Central / ERP training curriculum that starts from zero knowledge and goes through manufacturing plus a hands-on practicum. Built primarily with Claude/Fable (roughly 400 pages of material). Aimed at nearshore talent in El Salvador (Edgar, Fabricio, Jazelle, Melvin, Ronald, Fatima, and others). LMS is set up; first lesson and call start Monday. Goal is to have people job-ready for a large company seeking nearshore resources in 18–24 months. Plan includes periodic AI-assisted refreshes as Microsoft updates Business Central.

- Grok privacy issue
  Reports that Grok was uploading entire local repositories—including untracked files, credentials, and environment files—despite users being opted out. Speaker is avoiding Grok for now.

- Microsoft tenant billing confusion
  Received an unexpected invoice for Business Basic on a newly created isolated tenant used for GitHub Enterprise + Azure Key Vault. Tracking down the correct tenant ID took significant time.

- Custom PTE proposal & sales friction
  Sent a proposal for a custom Business Central extension. Will present at Vibe 2.0 in September on making the sales and proposal process more efficient. Noted how hard it is to scope custom work on short customer calls; prototypes are often required first.

- Dynamics GP ISV trial lesson
  A partner purchased the highest-capacity version of a high-volume posting product for a customer that only posts about 35 transactions per day. Strong reminder to always run a proper trial before buying the top-tier license.

- Marketing & website work
  Call with Daniella on marketing campaigns. Used Claude to add prominent pricing (including a $1,000-off promotion) to a product page in minutes. Also refreshed the five-year-old Privacy Policy and Terms of Service, added modern legal clauses, and switched to fully cookieless analytics (Cloudflare + exploring Umami) so no GDPR cookie banner is needed.

- Claude Code CLI tip
  Disabled accidental mouse-click selections with the environment variable CLAUDE_CODE_DISABLE_MOUSE_CLICKS.

- Automated documentation site
  Built a dedicated docs / user-guide site for the marketplace apps using Astro + Starlight. It is fully automated: every night it checks the app repositories for updated user guides and publishes them. Already live for the first two apps. Described as previously feeling like an unrealistic “super yacht” project that is now running with almost zero ongoing effort.

- Standardizing the AI development process
  Inspired by Waldo’s article on top-down AI frameworks. Pushing the small team toward one consistent process instead of each developer inventing their own approach to documentation, PRDs, user guides, etc. Current pain points include excessive Git branching by Claude, PR overload, and constant manual approval of actions.

- Local voice-to-text
  Settled on Open Whisper running locally (Nvidia model) on both the main machine and a Windows VM for RDP sessions. Basic transcription works well with only a short delay.

- GPT deletion incidents
  Multiple reports of GPT aggressively deleting files (including an entire home directory via rm -rf), production databases, and even all customer subscriptions on a Shopify store. OpenAI attributed some cases to full-access mode without proper sandboxing.

- Microsoft Marketplace (formerly AppSource)
  Official rename. App update was submitted and finally published around midnight, allowing a same-morning install for a customer. Removed trial record limits after they caused incomplete test data; now relying on online registration and license validation instead. MVP status renewed for the 13th year.

- App logic improvement
  Payment-to-invoice matching logic changed from invoice date to due date; Claude suggested adding a plus/minus tolerance window, which was accepted.

- AI project-management experiment
  Forwarded a real proposal email to an agent (OpenClaw) with instructions to create a Teams Planner opportunity, attach the document, summarize hours, assign it, and set a follow-up date. It worked in one shot. Next step is exploring Microsoft Work IQ + Teams Planner MCP server so Claude can handle richer multi-step project tasks.

- Upcoming events
  - Vibe 2.0 – September in Provo, Utah (speaking on marketing with zero budget/team).
  - Summit NA – October 11–15 in Nashville (book early; hotels and flights are expensive).

- Important warnings
  - Do not fully outsource critical thinking or bug hunting to AI; it is weak on edge cases, multi-step workflows, and external data.
  - Customer claims of “we only ever have one simple scenario” are frequently wrong when real data appears.
  - One company standardized on Microsoft AI tools and blocked everything else before licenses were ready, grinding work to a halt.
  - SonicWall SMA VPN appliances have a critical unauthenticated RCE (CVSS 10.0) that fully rootkits the device beyond recovery—hardware must be destroyed. Similar risks exist on other SonicWall SSLVPN products.
  - Cursor automatically executes any file named git.exe when a project is opened (known issue for many months).

- Team updates
  - Ronald – posting the ERP Academy curriculum.
  - Franklin – enhancing the partner dashboard (auto-pushing minimum supported BC versions) and investigating a hard-to-reproduce transaction error.
  - Fatima – helping with the academy, August meetup agenda, and testing project-management processes.
  - Daniella – building the marketing plan and campaign; learning UTM parameters and Umami analytics.
  - Joselyn & David – studying for the MB-800 exam (fixed assets, posting groups, and deferrals are hot topics).
  - Active MCP servers in use: Business Central, Microsoft AL Code, and Stefan Maron’s AL dependency server.

I Need Coffee - Episode 216 - Weekly BC Review! (July 10, 2026)
1. AI Development Practices & Lessons

Development Process Improvements:
Uses PRD → Detailed Design Doc → Comprehensive Development Plan (52 pages in latest example).
Adds wireframe walkthroughs for early validation.
Implements review gates, proofs of concept, and phased execution.

Key insight: One-shotting complex apps with AI is risky. Needs heavy guidance, nurturing (especially with Opus), and context.
Experimenting with Fable for development.
Learned hard lesson on prompting: Giving AI only code leads to hallucinations and poor root cause analysis. Needs full context (docs, rules, history, sample data).

2. Berlin ERP Academy Progress

Full curriculum built: 10 phases, 22 documents, 441 pages.
Timeline: 18–20 months for complete beginners (zero ERP knowledge) to become BC consultants + manufacturing experts.
Includes training guides, quizzes, labs, gates (exams/practicums).
Team (Fatima, Ronald) loading into LMS; using Notebook LM heavily for infographics, slides, and study aids.
Daniela (new marketing associate) will also participate to gain BC knowledge.

3. New Side Project: AI-Powered Project Management

Goal: Systematize project management using AI agents (OpenClaw, Claude, etc.) integrated with Teams Planner via MCP.
Vision: Auto-generate tasks, risk assessments (ISO 31000 style), status reports, budget vs actual, blockers, etc.
Fed by requirements docs, estimates, timesheets.
Enthusiastic about the potential but notes tools are still janky/slow.
Enlisting Fatima to help build this framework.

4. Business Central Development & Version Issues

Multi-company demo tenant trick: Use evaluation company setup wizard in unlicensed demo environments to create multiple companies.
BC Version Circus (28.x):
Avoid developing on 28.1, 28.2, 28.3 — major issues (Application Test Library fails, symbol download problems, update date resets).
Stick to 28.0 for stability.

DevOps/GitHub pipeline struggles with complex dependency scaffolding.

5. Marketing & Daniela Update

Daniela started as Marketing Associate.
Learning fundamentals from scratch (value propositions, Business Central vs ISV differences).
Will participate in ERP Academy for domain knowledge.
Steve’s approach: Building marketing capability in-house with minimal budget instead of expensive agencies ($13k/quarter quotes).

6. Conference Updates

Partner Vibe 2.0 (September 2026, Provo, UT):
Steve speaking on two sessions: “Marketing Without a Marketing Team” and AI-assisted project delivery.
Blue Dragonfly sponsoring (cocktail/mocktail + gamer prize).

Summit NA – October 2026, Nashville.
Other mentions: Dynamics Con Regionals (Ghana July, Columbus August).

7. Team Excellence Updates

Joselyn: Heavy financials, config packages, data loads.
David: Complex contacts migration (GP → BC).
Ronald: Continia Document Capture expert; formal project plans; Academy curriculum support. Starting implementations soon (6–7 week timeline).
Franklin: Sales Price List PTE + assemblies/kits; bug diagnosis; AI-generated user guides.
Fatima: Academy LMS setup; planning El Salvador DUG meetup in August; project management initiative support.
Daniela: Marketing fundamentals training.

8. Other Notable Items

Customer thank-you for responsive support.
Azure Foundry: Not recommended for frontier model dev work (quota, access, and model limitations).
Windows hardening tool: “Harden System Security” by Violet Hansen (recommended, available in Microsoft Store).
New Bill.com integration user guide in progress.
Upcoming tasks: Vibe code new ERP Center of Excellence website, finalize proposal, enhance integrations.

9. Closing Thought

Artwork highlight: “Fly Quietly” by Masaki Sasamoto (2016).
Encouragement: Go build/vibe code something, but be aware of AI’s limitations and the need for strong processes.