Celigo

We polled 300+ IT professionals about their # 1 concern about AI agents. The top answer was governance and guardrails. Above integration complexity, cost, and adoption. We dig into this and more in this month's Integration Bits. Also included: a live demo of Celigo Ora diagnosing an error in Celigo's production environment + 10 minutes of live Q&A from our AI launch webinar. Questions asked and answered include: ⏺️ Will Celigo's AI capabilities allow LLMs to be trained on my data? (answer: absolutely not) ⏺️ How do guardrails work inside Agent Builder so it doesn't do something it's not supposed to? ⏺️ What if my workflow is part rule-based, part AI? Integration Bits gives you the latest and greatest in AI and automation each month. Subscribe so you never miss an issue.

Were you at #DirectionsNA this week? Here's what our team heard on the floor. "My clients want to build integrations through normal language/conversation." "How can I keep new integrations from turning into custom projects?" "I need to deliver faster without my margins taking the hit." Microsoft's commitment to Business Central is bringing more customers to the platform, and those customers are arriving with bigger automation ambitions than ever. They want workflows that run end-to-end: across Shopify, Salesforce, their 3PL, their billing platform, their data warehouse. But when every one of those connections becomes a bespoke project, the math stops working. The AI pressure makes it more urgent. Because AI doesn't work on fragmented data and manual workflows. Partners who haven't solved the integration foundation yet are the ones who'll struggle to give clients a real answer when the AI question comes up. The partners we saw getting ahead of this at Directions aren't building harder. They're building smarter: reusable frameworks, prebuilt connectors, automation that clients can actually manage themselves post-go-live. If these were your conversations this week too, let's keep them going. Shoutout to our team on the ground for another great conference!

I ran our AI demo on a real production error, live on a public webinar. Here’s what happened. It was a full circle moment. In 2019, we were running a high-powered, GPU-filled machine in the lab to classify error messages with NLP. Last week, I pulled up a live production error in front of hundreds of customers and let Celigo Ora work through it in real time. No safety net. 😅 It diagnosed a Jira/Smartsheet sync failure in about a minute. Checked whether the flow had ever worked (it had, for at least the previous 21 days). Pulled Smartsheet documentation, and came back with a specific fix. Then, it asked me to confirm before doing anything. That last part is intentional. We're not handing the automation keys fully over to AI yet. Ora's capabilities are human-in-the-loop by design. I wrote up what happened: the demo, the live poll results on what enterprise teams are actually worried about with AI agents, and the Q&A that followed. https://lnkd.in/gz3937tg Curious what AI governance looks like for your organization right now?

Happy 6th birthday, Celigo U! 🎂 Six years ago, we set out with a simple belief: Whether you work in finance, IT, HR, or customer support, everyone deserves access to the knowledge needed to automate and scale their processes. Celigo University (now Celigo U) was built to make that a reality. Today, Celigo U offers free, self-paced training for any user with a Celigo account, covering platform fundamentals to advanced developer tools. Want live guidance? Celigo Liftoff is a training program that gives you hands-on experience with the support of a live instructor. And for those who want to take it a step further? The Celigo Builder Core Certification lets you put your skills on the record. We're proud of how far Celigo U has come and even more excited about what the next six years hold. The platform keeps growing, and so does the learning experience built around it. Not a Celigo user yet? You can still explore with the free Introduction to Celigo course: https://lnkd.in/gBawahgA No account required. Here's to six years of leveling up together. Start learning at training.celigo.com 🥳

A month after ShopTalk Spring 2026, here are 5 ecommerce takeaways that still hold true: 1. Operationalizing AI is still the hard part. 2. Fragmented stacks are a strategic liability. 3. Agentic commerce is already happening. 4. Peak demand is your integration stress test. 5. Rules-based automation and agentic AI belong on the same platform. In this guide, Ronen breaks down his learnings and observations from the conference. Also included, a clip of Jan on the main stage. 👏 Link in the comments!

Yesterday, Matt, Celigo's CPO, led a webinar demo on Ora. He pulled up a live error in Celigo's own Jira-to-Smartsheet flow and walked through exactly how Ora diagnoses, reasons, and recommends fixes. No prior knowledge of the flow required. The on-demand recording is now available. Sign up to get more demos, poll insights from your peers, an exclusive look at future offerings, and an energetic Q&A. Link in comments!

The view from Jamil's gondola is nice. The conversation at booth J70 will be even better. 🚠 Retail leaders are under real pressure right now: AI initiatives that promise operational uplift stall because the underlying stack isn't connected. Many AI solutions are just chatbots, not true agents. And that gap is where customer experience breaks down. We're sharing Celigo's latest capabilities at Retail Tech Show London to address this, including: Ora: a natural language interface built directly into Celigo, so any team member can build, manage, and troubleshoot automations without an IT ticket or specialist knowledge. Agent Builder: a low-code environment for constructing agents that actually take action across your systems, with human-in-the-loop approvals and governance built in. Together, they're how retail operations teams move AI from pilot to production. Come find us at booth J70 to learn more.

If you're in customer support or commerce ops, you've probably noticed: ⏺️ Your AI can draft a response. But it can't actually fix the order. ⏺️ Your AI can suggest a product. It can't pull live inventory to back it up. That's the gap between AI as a chat layer and AI as an execution layer. Here's what it looks like when agentic AI is actually connected to your systems: 🔍 Order Issue Resolution Instead of a rep jumping between your OMS, fulfillment tool, and CRM, an AI agent retrieves the order, checks shipment status, applies escalation logic, and creates a support ticket with full context attached. 🛒 Guided Buying Instead of a static recommendation engine, an AI agent connected to your product catalog and commerce systems helps buyers narrow options in real time, based on live operational context. Watch this demo to see how Celigo operationalizes both workflows. If you'd like to read a step-by-step walkthrough, the link is in the comments.

The conversation around enterprise AI has shifted. It's no longer about whether AI works. It's about whether you can trust it to work consistently at scale. Here's how Paula Melo, VP of Operational Excellence at Feedzai, put it in MIT's report with Celigo: the models are probabilistic, but the outcomes need to be deterministic and reliable. That gap between what AI can do and what the business can actually depend on is where most AI initiatives stall. We covered how IT and Ops leaders are closing it in our newsletter. Check it out - then download the report for the full story. https://lnkd.in/giitJEZh

At our Operate at the Speed of AI events in Philly and Boston, we sat down with IT and Ops leaders who are evolving their automation foundation before AI demand overwhelms their teams. Thanks to everyone who joined for two great discussions! Next stop: 📍 The First Lady Riverboat, Chicago, IL 🗓 May 13, 2026 Join us for: ⏺ An interactive workshop on the platform approach to AI ⏺ An exclusive look at the Celigo AI roadmap ⏺ Peer conversations with leaders solving the same challenges Spots are limited: save yours below. https://lnkd.in/g36buzjU