Representative End User Clients
Representative Automation Clients
Representative Software Clients
KEYWORDS: Kinaxis, Supply Chain Orchestration, Agentic AI, Data Fabric, Forward Deployed Engineering (FDE)
Overview
At the Kinexions 2026 conference in Las Vegas, a theme that has recurred in supply chain conversations once again arose. The market is taking steps toward orchestration via the use of AI, though in an uneven fashion, as is reasonable in such a step change. Kinaxis communicated a similar vision, led by Chief Executive Officer Razat Gaurav, who took the helm in January 2026. Gaurav expressed this position on the strength of the company’s consistently regarded supply chain planning solutions. Specifically, the company is using its AI-native Maestro platform as an execution command center designed to continuously sense, reason, and then execute supply chain decisions autonomously as necessary. It’s an interesting vision with a strong premise that will ultimately be decided by a market determining how many orchestrating “ingredients” are necessary for the soup that needs to be made.
Kinaxis is using its AI-native Maestro platform as an execution command center designed to continuously sense, reason, and then execute supply chain decisions autonomously as necessary.
A few key observations from the week’s event include:
- Kinaxis is making a concerted effort to move beyond AI potential by highlighting instances of skill-specific agents deployed within its customers’ operational execution flows.
- To facilitate the uptake and scale of AI, the company provides forward-deployed engineering services, both to reduce initial friction and to provide a means for developing specialized agents that can address specific supply chain needs of each client.
- Customer stories and proof points were central to the narrative, while also reflecting the myriad stages of maturation across markets.
Tracking Progress from a Year Ago
I always find it helpful to re-examine the ideas, solutions, and themes from a company’s last user conference to see what stuck, what progressed, and, perhaps, what stalled. I find that it’s a very practical way to apply a measuring stick. My colleague, Colin Masson, provided a fantastic breakdown of a major announcement for the 2025 event, which was the partnership between Kinaxis and Databricks. He broke down the implications of bringing the supply chain and industrial data fabrics together. So, when I was in a small group discussion with Gaurav at this year’s event, I asked him about the progress made on the partnership. Though he is new to the organization, he graciously answered one of the countless topics he is likely being asked about in detail. He did confirm what others said and showed throughout the week, as the companies combine to put data into motion with the aim of driving outcomes. Or, as Gaurav noted, "Supply chains don't struggle because of a lack of data; they struggle because decisions are disconnected. What's changing is that AI is no longer just informing decisions; it's becoming part of how decisions are made and executed."
Databricks is also being employed as a critical component of Kinaxis’ capability to orchestrate, a primary goal of its vision for establishing competitive differentiation through autonomous AI workflow and decisions.
Kinaxis is leaning on Databricks infrastructure to extend the supply chain company’s ability to feed and improve its machine learning solutions using live data. Makes perfect sense, and now they are able to demonstrate this capacity working in operational use cases and customer environments. Databricks is also being employed as a critical component of Kinaxis’ capability to orchestrate, goal of its establishing a primary vision for competitive differentiation through autonomous AI workflow and decisions. Databricks unifies and presents structured and unstructured data from both within and outside of the enterprise. Kinaxis (and its partners, it is important to note) can tap this data, apply additional insights, and then deliver semantic and ontological intelligence using AI agents on a single platform versus an integrated compilation of acquired or fit-for-purpose execution applications. A year in, Kinaxis has now moved beyond the articulation of the vision with Databricks to show how it can be executed and instances where aspects are operationalized.
Skill Building, Speed to Outcomes, and Standardization
While the main announcements were a collection of the recent and new, they were delivered with the intent of presenting a clear message. The theme from this year’s event specifically targeted success in initial adoption and the speed of doing so. As such, they individually and collectively provide a solid reflection of moving from “can do” to “do.”
Embedded Collaboration
The major capability announcement signaled a shift from traditional pre-defined software features to engineered solutions. To help customers operationalize AI, Kinaxis launched forward deployed engineering (FDE) services. Though the language harkens to the approach coined by Palantir (and that was based on a military deployment technique) and now used by numerous technology companies (including Databricks), the native engagement service embeds Kinaxis engineers into their clients’ operations to co-build and deploy context-specific agents directly into workflows. They communicated an approach that co-builds pragmatically, based on the “physics of enterprise operations.” This injection of specialization into the area of need is designed to improve initial development, value realization, and expansion of use to scale.
Agent Skills and Planning Compression
In March 2026, Kinaxis announced performance results from integrating NVIDIA’s cuOpt GPU-accelerated optimization engine directly into Maestro. They again shared the output with their customer base at the event. The test was done on a semiconductor planning model containing nearly 50 million decision variables and 40,000 SKUs. The performance report indicated that the engine delivered up to 12 times reduction in total end-to-end calculation time. Daily planning cycles were also stated as compressed from hours to minutes. Kinaxis has included a library of “agent skills” within Maestro with the purpose of facilitating these types of performances with other customers. Beyond the cuOpt example, agent skills reflect the addition of more precisely designed AI within Maestro for specific problem-solving. I’m convinced we will see considerably more of this approach from Kinaxis and others, as supply chain is just one example of industrial process areas that require far more expertise than generic reasoning.
Standardization to Improve Data Use and Security
Similar to the cuOpt announcement, Kinaxis shared with its customer base the Genpact Kinaxis MCP Server, built on Amazon Bedrock AgentCore framework. Again, this was announced a few months prior but worth bringing to the attention of those not in attendance. The capability leverages model context protocol (MCP) to apply standardization for data access and extraction to improve the speed and precision of data queries. They noted an example of the capability that enabled a client to reduce what was typically a 12-week project to just three days. Additionally, it adds security and auditability to workflow authorization and control, with session isolation, performance telemetry, human-in-the-loop approval gates as examples. Ultimately, it’s another step in the details required to remove friction points related to the deployment and continued use of AI.
Other Noteworthy Mentions
- Emphasis on Storytelling: On the surface, this might appear to be, well, surface. However, it’s a reflection of the need in the market for companies like Kinaxis to tell effective stories about how move from the possibility of autonomy and end-to-end orchestration to the reality of it. To that end, Kinaxis Appoints Kristin Russel as Chief Marketing Officer.
- Remix Maestro Hackathon: Adding directly to the above on the need to show how possibility becomes reality, the company announced a three-week customer and partner program designed to build and deploy AI agents using Maestro Agent Studio.
- Customer Awards: As the company has done for each of eight years, it announced its 2026 customer recognition awards at this year’s event. Winners included Reckitt, General Motors, Lupin, Jabil, and Cardinal Health.
Customer Stories Take Center Stage
This is my favorite part of any user event, and Kinexions was no exception. The true test of any provider’s performance lies in the measurable business value it delivers and the willingness of customers to publicly share results. Customers of Kinaxis and its partners showed clear results from various uses of the company’s solutions, such as improved cycle times, manual process elimination, waste and cost reduction, and improved response. And yes, some of these were directly a result from the use of AI. The following are a few customer perspectives worth noting.
Cardinal Health
This Fortune 15 multinational health care services company delivered a crash course in the fruits of sound thinking when it comes to modernization, as presented by SVP, Global Supply Chain and VP Global Medical Products, Peter Bennett, and VP Global Medical Products, Peter Jenny (yes, the two Peters, one stage). They focused on the company’s global medical and products distribution business, which operates in 50 countries and across 10,000 trade lanes, supported by integrated manufacturing sites. Beginning with a massive acquisition spree in 2018, they outlined how the organization came to realize that what they perceived as a technology problem really wasn’t. It was about understanding what long-term viability looked like, defining the outcomes necessary to that, and letting that lead them back to strategy-driven simplification.
They began not with technology but with the people involved in their business, the personas of those roles, the associated disparate processes, and then the systems that supported that disconnected environment. Pulling experts from across the business, they undertook a series of what they refer to as business implementations, run as related initiatives under a single vision, for reorganizing customer management processes, supporting decision autonomy, and simplifying technology architecture. Central to that was advanced planning capabilities using Maestro (then Rapid Response), with support that began and is maintained at the CEO level, enabling exponentially more control over end-to-end supply chain decisions. As evidence of their progress, they stated their service level rating improved double-digit percentages. Not to diminish all the detail that went into that, but the takeaway was that it was accomplished by beginning with market signals, using them to define the necessary outcomes and success metrics, mapping that to the relevant people and processes, and then determining if, where, and what technology would be applied to support change. It’s not easy, probably not always exciting and cutting edge, but it can change the competitive trajectory of the business.
Unilever
Global Head of Customer Operations, Graham Sommer, of the $55 billion global consumer products giant that operates in 150 markets, gave an updated and rather elegant view of the company’s organized progression toward their modernization of market responsiveness, including their use of AI agents to do so, seven of which he directly discussed. The customer operations team has two mandates that cover growing the business and providing a superior customer experience. My takeaway from this simple explanation of purpose is that it provides the group with a specific set of market signals from which to develop capability, providing a massive head start for where to begin, why, and to what effect. It was refreshing in its clarity and specificity. As you might guess with the complexity of their supply chain, they began with integrated planning that incorporated demand, forecasting, supply, factory, and materials, as well as logistics management.
They also did a deep dive into customer problem statement interactions, realizing they had a lot of commonalities across an amazingly disparate customer base. In doing so, they were able to identify a significant set of common issues that all clients needed to address in shades of a similar fashion. Using that as a springboard, they developed agents to support their efforts, even adding them to the org chart. In a nod to the diversified and dispersed nature of the customer base they serve, they have extended agentic capacity through the development of localized market versions. Again, this is another example of something I think we’ll see more of, particularly as various players in the value chains grapple with smoothing the cost of AI use.
The end result architecture is vastly simplified and direct in purpose, down from 150 systems to 10 core “business products” and from 300 to 20 integrated ecosystems that are specifically aligned to support a highly defined, market-first business strategy. While this transformational approach is not fully deployed across the entirety of Unilever yet, it is in production across certain areas of the business. They communicated a host of measurable improvements in forecast accuracy, dispatch rate, product creation cycle, finished goods inventory, business waste, and logistics costs.
Qualcomm
During his conversation. Brent Wilson, Senior Vice President of Global Supply Chain Operations of semiconductor stalwart Qualcomm, noted that transformation, in their case for S&OP, requires an organizational and cultural shift where supply chain is recognized as a "team sport.” He pointed to the company’s cross-functional teaching group and an event they named their “supply chain Olympics” as collaborative examples. As the company transitions to the cloud, it is leveraging Kinaxis to connect demand, capacity, and execution planning. From a market perspective, he also outlined how the company is betting on AI being a central feature of the user interface experience. With this quick sidenote mention, he demonstrated that no matter the subject, AI is a grappling point in all of them.
Conclusion
The Kinaxis event brought the energy, dialog, possibility, and reality of markets undergoing profound change. The company has outlined its approach to the market and its customer base rather clearly. They have a solid foundation, and their customers are taking advantage of it, particularly in building some of the core intelligence and structure required to employ transformative business strategies more broadly. While significant headwinds still exist, especially with AI’s use, they are not specific to Kinaxis by any means. I’ll dive into the broader market signals I’m seeing emerge over the last number of months, which will probably feel like a decade long, given the pace of change in today’s hyperconnected markets.