Alexander Sutherland

I have heard so many vendors try to convince the government that providing their API to the government is more than sufficient to "meet MOSA". ....so much to unpack with such a perspective. Here are just a few considerations: 1. There is a lot of behavioral content burried in APIs; it must all get exposed (ICDs must specifically mandate that content) 2. Often vendors use a large monolithic API to tie together basic services for multiple components (services need to be pulled out and part of a shared services layer..forcing vendors to plug into a governed API environment) 3. They frequently leverage custom middleware that is NOT standards based. 4. The above is extra problemmatic for software products that are large and "do everything", especially when the product is not modular and based off of government OR commercial driven standards/architectures. #mosa #software #vendorlock

29

1 Comment

#1MinuteTip For years, changing Salesforce org meant knowing exactly where every setting lived — and clicking my way there. Then came the AI coding tools — Cursor, Copilot, Claude Code. Finally I could describe what I wanted in plain English. Real progress. But I was still working inside a developer tool. Now there's a third phase. No IDE. No command line. Just chat. I experimented with adding three fields to the Account object from a Claude chat window. I pasted a short spec — a text field, a checkbox, and a picklist — hit enter, approved the action, and watched all three land on Account. (Check out the video below) What makes this possible is the new Headless 360 MCP server, now in beta. Here's why it matters. The Salesforce MCP servers so far were mostly about data — querying, creating, updating, deleting records. Handy, but they never touched how the org is built. Headless 360 crosses that line. It works with your metadata — the configuration itself. Fields, automation, integration setup. The stuff we normally click through by hand. The part that matters for admins: it runs as you. Your permissions still apply — field-level security, sharing rules, permission sets. If you can't do it in Salesforce, your agent can't either. And every action is audited back to you. Want to try it? Setup → Salesforce Hosted MCP Servers → find Headless 360 → Activate. Then connect it to an AI client like Claude. Want the full step-by-step walkthrough on configuring the Salesforce MCP server? Here is the link to the blog post and step-by-step guide - https://lnkd.in/e23_D3x9 Each phase removed a layer between knowing what you want and getting it done. This one removed the tooling. Which admin tasks would you happily hand to an AI — and which would you never let it touch? #Salesforce #SalesforceAdmin #AI #MCP #Headless360

19

3 Comments

🚀 The Best Salesforce Features Released in 2025 Salesforce’s 2025 releases weren’t about flashy reinvention — they were about making day-to-day work better. While not every update was game-changing, several new features stood out as genuinely useful for admins, developers, and sales teams. Thoughtful improvements, smarter tooling, and quality-of-life upgrades that actually make a difference in real orgs. In my latest article, I break down: • The standout features from 2025 worth paying attention to • Why tools like Agentforce Builder & Vibes feel like a turning point for AI in Salesforce • The admin wins that improve UX • Sales features that support more strategic, data-driven teams If you want a practical roundup — not release-note overload — this one’s for you. 👉 Read: https://lnkd.in/eNjha_Wx What was your favourite Salesforce feature released this year? 👇 #Salesforce #SalesforceAdmins #SalesforceDevelopers #SalesforceFlow #Agentforce

16

When we launched the Salesforce Lightning Design System nearly ten years ago, it was about more than components and tokens — it was about creating a shared language for how we and our ecosystem design and build at and on the Salesforce platform. Today, with SLDS 2 going GA, I can’t help but take a moment to reflect. This milestone is more than a product release; it’s the result of years of listening, learning, and co-creating with an extraordinary community of customers, partners, and colleagues. SLDS 2 is not just a new system — it’s a foundation built for what’s next: AI-driven experiences, enterprise scale, and agentic design patterns that adapt and respond to people, not pages. To my entire dedicated team and everyone who has contributed, challenged us, and believed in what design systems can unlock — thank you. You’ve made this system stronger, smarter, and more human. A design system is never just a set of components. It’s the patterns, guidelines, visual language, tooling, and design experts who have a passion for systems design that inspire makers to innovate and craft exceptional experiences — all supported by a flexible framework that empowers you to make it your own. We’re launching a new customer story alongside SLDS 2 featuring Stephy Hogan and Cassandra Soryung Yu because, ultimately, this isn’t about what we built. It’s about what you will create with it. The real measure of a design system is in the products, stories, and impact it enables — and those belong to you. Here’s to the next chapter of design at Salesforce — and to designing systems that keep evolving right alongside us. Salesforce Designers #designsystems #slds #ux #enterprisedesign

200

9 Comments

Predicting a vibration anomaly at 300 mph isn't just a safety requirement: it's a data integrity challenge traditional monitoring can't handle. In our PDI testing for the ServiceNow Zurich release, we’ve been pushing Agentic AI to its limits, moving beyond simple automation into true autonomy for Maglev infrastructure. Most ServiceNow consulting services focus on basic alerts. But real "Speed to Value" comes from positioning the Workflow Data Fabric Hub as your core. We’re seeing a shift where the AI Control Tower doesn't just flag potential track fatigue; it leverages Agentic Playbooks to reroute traffic and schedule proactive maintenance. I’ve seen the technical scars from "noisy" IoT integrations where sensor floods crash the MID Server. We solve this by implementing robust Machine Identity governance and rate-limiting within ITOM and ITAM. This is how you scale high-speed transit without scaling technical debt. Don't settle for generic AI hype. Work with a ServiceNow implementation partner that knows the Patch 5 friction points and how to build for the Zurich release. Schedule a deep dive with our experts: https://lnkd.in/eEKfRi_u

3

Everyone keeps asking the wrong question. “How do we make our enterprise systems AI ready?” Here’s the honest answer. You don’t. Not today. Not tomorrow. And not by refactoring your Salesforce, ServiceNow, or Snowflake orgs. The amount of cleanup required is absurd. Years of workflows, automations, undocumented dependencies, tribal knowledge baked into production. That is why most enterprise AI projects stall after the demo. It is also why 90% plus of them quietly fail. What we see in practice looks very different. The most complex orgs we work with are not cleaner. They are not more modern. They just stopped trying to fix everything first. Instead of pushing AI into systems that were never designed for it, we built an agentic layer outside of them. You do not fix your systems. You connect them. That layer understands metadata, architecture, ownership, and blast radius across systems. It gives agents and teams shared context so changes can actually happen. Today we released a new Build Agent that works with your team to build across enterprise systems. Not suggestions. Actual building. Converting 1,000 legacy workflows to modern flows in about 10 minutes. With full understanding of what exists, what is connected, and what can and cannot break. Once that layer exists, everything changes. Agentforce implementations stop being scary. Cross system data lineage across Salesforce and Snowflake just works. ServiceNow automations do not require the one person who “knows how it works.” This is what AI readiness really looks like in the enterprise. Not clean systems. The right abstraction layer. Connect your systems. And suddenly, they are AI ready.

33

2 Comments

Are you paying tens of thousands a year to back up your Salesforce org? I've been through the full Salesforce backup journey: manual Weekly Export to a file server, then a commercial solution at $3-5 per user per month, and finally building my own. After running a custom bash/Python backup in production for nearly three years, I rewrote it from scratch in Java with Spring Boot — fixing everything I'd learned the hard way about error handling, file deduplication, and checkpoint resume. Then I open-sourced it. Heimdall is a self-hosted Salesforce backup that runs on AWS for $5-15/month. It uses the Bulk API, stores data in S3 and PostgreSQL, and includes a web GUI for browsing history, viewing field-level diffs, and related records. Some highlights: - ContentVersion deduplication saved 320 GB (38%) on an org with 3.5M files - Checkpoint resume survives crashes and container restarts - The database auto-starts/stops to minimize costs - Record archiving with configurable criteria - One-command AWS deployment via CloudFormation It's AGPL-3.0 licensed — free to use, modify, and deploy for your own org. GitHub: https://lnkd.in/deeRHpA7 Full writeup: https://lnkd.in/dEU4_b_K If you're running a Salesforce org and want to stop overpaying for backups, give it a look. PRs and feedback welcome. #Salesforce #OpenSource #AWS #Backup #Java #SpringBoot #Trailblazer

146

15 Comments