The future of shopping, if you listen to its more breathless salesmen, is nearly frictionless. An agent knows what you want, finds it, buys it and politely vanishes. But between desire and delivery sits a crowded mechanical room: product descriptions, inventory counts, pricing rules, warehouse choices, split shipments, returns and old software with a long memory. Mike Micucci has spent a career in that room. He is cheerful about it, which may be the first useful clue.
Micucci is the chief executive of fabric, the enterprise commerce software company. His current vocabulary includes agentic commerce, product-data confidence and AI order management. His résumé begins well before any of those phrases could clear a marketing meeting. It runs through HP, Netscape and Commerce One; through a collaboration startup called GroupSwim; through eleven years at Salesforce; and through a stint as chief operating officer of Automation Anywhere. The products changed. The difficult part stayed oddly consistent: make complex systems useful to actual people.
That continuity matters. Micucci has seen several revolutions arrive carrying fresh acronyms. In 2001, as vice president of solutions strategy at Commerce One, he was discussing XML links between giant companies and Microsoft's then-new .NET platform. Today he talks about agents moving across retail systems. The distance between those conversations is enormous. Their basic promise - let software coordinate work across organizational borders - is close enough to be family.
An exit that became an entrance
In 2005, Micucci co-founded GroupSwim, an early business collaboration platform. The name now sounds like a particularly ambitious calendar invitation. The idea was serious: give companies an online place to share knowledge, find expertise and work together. By early 2009 the company had found product-market fit and attracted two acquisition offers. Salesforce bought it that December.
For many founders, the acquisition announcement is the final scene. For Micucci, it became a change of set. GroupSwim's technology fed into Chatter, Salesforce's enterprise social network, which launched in 2010. He stayed. He then pitched a way to connect communities around Salesforce customers, partners and employees. That idea grew into Community Cloud.
The progression was unusually tidy in retrospect: collaboration became community; community met commerce. Micucci has said that a customer who built commerce capability into Community Cloud in 2014 helped stimulate the product direction. By 2016, Community Cloud and Commerce Cloud were one business unit. He eventually became CEO and executive vice president of Salesforce Commerce Cloud, responsible for B2C and B2B products used by major brands and hundreds of millions of shoppers.
The ascent also turned a founder into a student of scale. At GroupSwim, a customer could call the CEO. At Salesforce, dashboards translated millions of customer actions into operating signals. Micucci has described separating usage measures from business measures: daily and monthly users on one side; gross merchandise value, contract value, order value and churn on the other. The romance of cloud software, it turns out, comes with a fine collection of denominators.
Internet products
Markets and collaboration
Community and commerce clouds
Automation and agents
The operator takes the whole floor
In May 2021, Micucci joined Automation Anywhere as chief operating officer. The remit widened from one cloud business to global operations and company strategy. Automation Anywhere was moving its robotic-process automation platform toward the cloud while adding AI and process intelligence. Micucci helped oversee that turn, including the period in which the company acquired process-discovery specialist FortressIQ.
He left after a little more than two years, publishing a note that dwelled less on his own next move than on the colleagues who had made the work possible. That pattern appears often in his public writing. Product launches come with lists of engineers, marketers, customers and partners. His announcement that he was joining fabric thanked the board, the team and Faisal Masud, the chief executive he was replacing. An executive post can be a small monument to its author. Micucci tends to install extra doors.
Fabric brought him back to ecommerce in August 2023. He called it his dream job. The attraction was partly familiar territory and partly unfinished business. Retailers had assembled modern-looking storefronts on top of systems that were often siloed, expensive to change and fragile during peak demand. A shopper might see one brand. Behind the screen, inventory, merchandising, payment, order and fulfillment systems conducted a tense committee meeting.
The order is the story
Micucci's diagnosis begins after the moment most commerce advertising ends. Checkout is merely when a retailer makes a promise. Order management decides whether the promise survives. Can an item be found in a store when a warehouse is empty? Should one order become two shipments? Can a customer change delivery without calling support? Does the associate see the same inventory as the website?
In December 2024, fabric announced AI Order Cloud after more than a year of research, development and work with design partners. Micucci described the order-management system as the center of the brand, but argued that many such systems still followed an old financial-control model. Fabric's approach put agents and analytics closer to the operational data. An employee could ask about split shipments or fulfillment performance, receive an explanation and then be guided toward a rules change that once required programmers and months of work.
Ask what the merchant or customer must accomplish before admiring the architecture.
Checkout creates the promise. Inventory and fulfillment decide whether it is kept.
Add useful services to the existing stack when a total replacement would create needless risk.
Agents need structured facts, current availability and confidence, not decorative prose.
The practical tone is worth stealing. Micucci does not argue that every retailer should detonate its technology stack. Fabric can add discrete services such as inventory or promotions, or provide a broader unified system. The point is to shorten the path to a result. By 2025, he and retail veteran Jay Topper were talking about capturing return on investment in 90 days instead of years. This is less cinematic than transformation. It is also easier to put on a calendar.
Now the customer may be software
The next turn in Micucci's career arrives with a peculiar customer: an AI agent acting for a human shopper. Agents do not browse the way people do. They ingest attributes, compare claims and look for evidence that a product fits a request. A poetic product page can charm a person. A machine still needs the dimensions, materials, compatibility, price and availability in forms it can interpret.
That is why Micucci calls product data the front door for AI discovery. Fabric's Product Agent monitors how catalogs appear in answer engines, enriches product information and assigns confidence to the result. In spring 2026, Micucci emphasized the sequence: normalize first, enrich second, expose a clear signal about quality. “Real structure. Real enrichment. Real signal,” he wrote. It is a three-beat slogan for work that most shoppers will never knowingly see.
He also treats AI as more than a new chat box. At NRF in January 2026, Micucci said fabric had spent 18 months taking a first-principles approach to commerce in an agentic world. He pointed to Google's Universal Commerce Protocol as a meaningful direction. The larger ambition is interoperability: agents that can understand product truth, operate across commerce systems and help people act without concealing how confident the system ought to be.
There is a neat irony here. The more conversational shopping becomes, the more disciplined the underlying records must be. An agent can produce a lovely sentence about a jacket. It cannot make the jacket exist in a warehouse. It can recommend next-day delivery. It cannot rescue a stale inventory count from becoming a disappointed customer.
By August 2026, Micucci was sharing observations from a year of customer work on AI shopping. He said fabric had seen AI-referred shoppers add products to carts 30 to 70 percent more often than other traffic, with one brand converting that traffic at twice the rate. The numbers came with a warning against vanity metrics. Retail leaders, in his telling, need to understand where products surface, make catalogs answerable and connect visibility to business results.
A long memory for the next wave
Micucci's advantage is not that he arrived earliest to AI. It is that he remembers what happened after earlier breakthroughs met the enterprise. Integrations mattered. Governance mattered. Customers asked for features no roadmap had anticipated. Branding and pricing proved harder than expected. The winning system was rarely the one with the finest stage demonstration; it was the one still behaving during a holiday spike.
His career also offers a less obvious model of ambition. Founding a company did not prevent him from thriving inside an acquirer. Running a major cloud business did not prevent him from taking a wider operating job. Leaving that job did not prevent him from returning to a familiar industry with a different thesis. Each move accumulated another view of the same machine.
The aspiration at fabric is large: build an enduring commerce company for a world in which agents help discover, decide and operate. Yet the daily proof will remain stubbornly ordinary. Is the product described correctly? Is it available? Can the retailer fulfill the promise profitably? Can a person intervene? Can everyone trust the answer?
Those questions will not stop a conference hall in its tracks. They do something more useful. They keep the box moving toward the right door. After twenty-five years of new interfaces, Micucci is still working on the machinery. The machinery, with its terrible timing and excellent job security, is finally becoming the headline.