If a model can build a working prototype in an afternoon, what is left for SaaS platforms to defend?
This question was top of mind among product leaders at World of Workato (WOW) 2026. In the session: Your Data, Your Moat: Orchestrate, Protect, Defend, moderator Kunal Khandekar, Director of Customer Success for Workato Embedded Platform, asked three customers how they built an AI moat, and why yesterday’s integration decisions now make their AI and MCP plans work. Here’s what they said and what you can borrow.
On the panel:
- Alexey Korotich, Chief Product Officer, Wrike: a collaborative, enterprise-grade project management and work management platform designed to help teams track tasks, manage complex projects, and streamline workflows.
- Andy Armstrong, Head of Business and Market Strategy for Credit Analytics, S&P Global: a credit risk analytics provider that uses quantitative models and data to score the creditworthiness of public and private companies worldwide.
- Umang Bhatt, Principal Product Manager, Epicor Software: an industry-specific, enterprise-grade ERP software provider for manufacturers, distributors, retailers, building suppliers, and automotive businesses.
Become the platform your customers can’t live without: Workato ONE for SaaS Platforms
Saying “yes” to every integration
Each company came to Workato for different reasons, but faced the same pressure: customers needed their systems connected, and internal teams couldn’t build and maintain every integration themselves.
Epicor customers wanted to connect third-party apps, run their own automations, and stay in control. The systems varied by region: customers might use Salesforce in the US, HubSpot in Europe, and different HR systems in Australia and New Zealand. Epicor created a connector and gave customers access to the Workato platform through Automation Studio.
At S&P Global, the requests came through RFPs, especially in Europe. Buyers wanted S&P’s data through a connector, with a faster path to value than a point-to-point API project.
Wrike workflows span multiple systems. Connecting Salesforce opportunities to time tracking in ERPs meant the possible integrations quickly multiplied. Wrike chose Workato for its security, reliability, and operations, then launched Wrike Integrate, with curated recipes customers can use as is or customize.
What Workato made possible:
- Epicor: Shopify and WooCommerce orders reach the ERP in about a week instead of months. Sales can say yes when an RFP asks about an integration.
- S&P Global: A supplier risk analyst can choose a data package, tailor it, and send it to SAP Ariba.
- Wrike: Workato handles complexity customers would otherwise face, connecting Wrike to the systems where work starts and ends.
The requests differed, but they pointed to the same strategic need: each platform had to fit into the systems its customers already used. Integration helped make that possible without requiring each company to build every connection on its own.
“ERPs can survive siloes, but we cannot thrive. We have to have integrations available.”
— Umang Bhatt, Epicor
Your moat is what your platform makes possible with data
Across the panel, a pattern emerged: data alone rarely creates a lasting advantage. The advantage comes from turning it into something customers can use—insight in the moment a decision is made, visibility into how work is progressing, or expertise that helps them run their business.
Andy illustrated the difference between having data and creating value from it. Financials, sanctions lists, headcount, and adverse media are widely collected and sold. S&P’s edge is the proprietary analytics it layers on top, and its ability to deliver those analytics into the systems where decisions happen.
Wrike starts with a different kind of data: what it calls the telemetry of work. Every action taken by a person, system, or agent in a workflow creates a signal. Customers can use those signals to see where a process stands and improve it. Over time, that activity can feed a flywheel Wrike is working to launch, with early wins from customers.
Epicor’s differentiator is decades of ERP expertise. Umang described the company’s approach as putting that expertise in customers’ hands rather than fencing it off.
Together, these examples suggest a broader view of the moat: it doesn’t have to be a wall around data or intellectual property. It can come from applying data and expertise where customers need them – and earning their trust in the process. Alexey said Wrike sees AI as a multiplier for its IP, not a reason to guard it from AI. Umang echoed that outlook: customers already trust Epicor with their operations, and the company’s job is to guide them well.
“The more data is in the system, the smarter your agents become. And the smarter your agents become, the more data they create back into the system.”
— Alexey Korotich, Wrike
Roll out AI with guardrails that fit your business
None of the three companies described AI as a single launch. Each is expanding what its AI can do in stages, while setting boundaries around data and actions.
Epicor calls its approach crawl, walk, run:
- Crawl: an agent that answers questions about help documentation.
- Walk: an agent that helps customers run 14 standard workflows, such as quote-to-cash.
- Run: customers building MCP servers on Workato to connect other apps and write into the ERP, alongside a headless ERP planned for 2027.
Epicor draws a line between access and action: its agent reads, while Workato writes. On customer data, Umang said Epicor won’t use one customer’s data for another, and that the boundaries are still being defined.
Wrike’s path moves from a copilot to agents it calls digital labor, then toward Workmates, its next generation of AI built to take on roles like project manager. Self-healing workflows are a longer-term goal.
S&P showed how the same questions play out in a regulated business. Its Credit Memo Builder cuts a task that took an analyst six to eight hours down to minutes, while supporting the analyst rather than replacing them. Andy said it combines AI and deterministic workflows. S&P is still experimenting, while keeping a sharp line between what comes from S&P and what comes from a model.
The approaches differ, but the shared challenge is to make AI useful while keeping its actions and outputs reliable, traceable, and appropriate to the business.
“It has to be reliable. It has to be traceable. It has to be authentic.” — Andy Armstrong, S&P Global
“Can’t enterprises build this themselves?“
An attendee raised the question behind many SaaS roadmaps: as building gets easier, why wouldn’t enterprises create their own systems?
Alexey said expectations are changing, even at Fortune 10 companies. Nontechnical users are excited that MCP lets them build on a platform’s data, and Wrike is investing in that possibility. But a working demo is only the beginning. The solution still has to be reliable, scalable, and secure. Enterprises look to vendors to provide that foundation, and they won’t switch just for the appeal of a shiny new tool. Trust takes years to earn.
Umang sees the same gap between a proof of concept and a system people can depend on. Someone still has to scale and maintain it. Epicor’s customers make physical goods, not software, and as AI news spreads, they keep asking Epicor how it can help.
The panel’s answer points to a changing role for SaaS platforms. Customers may build more of their own tools on top of platform data, while still relying on vendors for the systems, expertise, and trust that make those tools usable in real operations. As you review, check that “Wrike is investing in that possibility” and “Epicor how it can help” match what the panelists actually said.
Platforms earn trust through execution and control
Wrike, S&P Global, and Epicor connected their products to the systems customers already rely on before they began building their AI strategies. Those connections now support what comes next: analytics and work telemetry that inform decisions, and agents that can act across workflows.
Together, their examples show how SaaS platforms can stay essential: make systems work together, turn data and expertise into useful action, and give customers control over how AI operates.
Execution depends on integration. Control depends on governance. Workato ONE for SaaS Platforms gives product teams one governed foundation for the workflows, agents, and integrations they ship.
Explore Workato ONE for SaaS Platforms
