Kinaxis Maestro Agents are AI agents that Kinaxis embeds in its Maestro supply chain planning platform. Kinaxis announced them on 2025-10-17 as “now available to Kinaxis customers and embedded natively in Kinaxis Maestro”, with Maestro Agent Studio in limited availability.[1] Maestro Agent Studio followed on 2026-02-05.[2] The capabilities are of interest to licence managers because Kinaxis describes them as a source of new revenue, which suggests they are commercially separate from the base platform, yet it publishes no metric, unit, price or entitlement terms for them. This article collects what the Kinaxis website, Trust Center and results releases say, and marks what they leave open.
Editions
Maestro Agents
Maestro Agents are described as context-aware digital co-workers that analyze live data, flag issues and recommend next-best actions inside Maestro, within human-in-the-loop guardrails.[1] The commercial position can be read from three Kinaxis statements. In the third-quarter 2025 results, Kinaxis said the initial agents create “the opportunity for a new revenue stream for Kinaxis”.[3] In the first-quarter 2026 results, it reported “good early-stage demand for our Maestro Agents, with new paying customers in Q1”.[4] In the second-quarter 2026 results, it said approximately 10% of its installed customer base was on a paid or trial subscription to Maestro Agents at 2026-06-30.[5] Together these statements say that agents are subscribed, that some subscriptions are paid and some are trials. They do not say which agents exist, whether agents are sold individually or as a bundle, or what quantity limits apply. See Maestro Agents.
Maestro Agent Studio
Maestro Agent Studio gives supply chain teams “a no-code way to compose AI agents grounded in their real operating context”, built into Maestro and designed to work with large language models including OpenAI’s GPT and Google Gemini.[2] Agents built in the studio run inside Maestro’s concurrent environment with human-in-the-loop oversight.[2] The Planning One page presents a path in which customers start with conversational AI embedded in Maestro, extend to agent-led workflows, and then use Agent Studio to create agents around their own rules.[8] That ordering suggests tiers of capability but the page does not describe them as tiers of licence. See Maestro Agent Studio.
Kinaxis stated that later in 2026 it planned orchestrator agents that coordinate multiple agents, secure connections between Maestro Agents and external agents and systems, and expanded shared data context.[2] These are statements of intent. By the second-quarter 2026 results Kinaxis reported that it had launched an Orchestrator Agent and reusable agent skills, as well as Maestro Advanced Solver Studio.[5] A licence manager should treat a roadmap statement as outside any entitlement until the Order Form lists the capability.
Advanced Solver Studio
Advanced Solver Studio adds Gurobi-powered optimization for complex, constraint-heavy decisions, with models orchestrated through governed Maestro workflows and results delivered back to the planning workspace.[9] A separate release describes GPU-accelerated optimization using NVIDIA cuOpt integrated within the Maestro platform.[11] Neither source says whether the customer receives a separate licence from Gurobi or NVIDIA, or whether the components are sublicensed through Kinaxis. See Maestro Advanced Solver Studio.
Forward Deployed Engineering
Kinaxis introduced Forward Deployed Engineering on 2026-06-02 as an engagement model in which Kinaxis co-builds agentic solutions with customers on Maestro.[7] It states that FDE “will be available through Kinaxis account teams for selected customer engagements”.[7] The release frames the model as moving from features to outcomes, from project to product and from go-live to ownership, and says that solutions developed in an engagement establish a foundation that customers and partners can extend.[7] It does not say who owns the intellectual property of a co-built solution or how it is priced. Those points belong in the Professional Services Agreement and Statement of Work. See Forward Deployed Engineering.
Metrics
No metric is published for any AI capability. Kinaxis does not state whether agents are counted per user, per agent, per run or per unit of consumption. No Kinaxis document on kinaxis.com names a unit, and the catalogue records no agent metric. The only quantities are in the second-quarter results: approximately 10% of the installed base on a paid or trial subscription at 2026-06-30.[5]
Data use, governance and customer control
The Trust Center has a section on AI security. It makes the following statements, each of which Kinaxis says reflects current practice and not a contractual commitment.[6]
- Kinaxis “does not use customer data in Maestro to train public or shared AI models”, and any large language model usage occurs within a customer-dedicated environment.[6]
- Customer environments are isolated, and AI capabilities do not introduce cross-customer data access.[6]
- Maestro’s AI features process only customer-authorized data already available in the customer’s environment, solely to deliver the requested functionality.[6]
- “Maestro AI capabilities are not mandatory”: customers control which features are enabled, who can access them and how outputs are used.[6]
- Maestro and Maestro Agents do not independently execute actions outside customer-configured workflows, permissions and controls.[6]
The Trust Center states that its content “is not intended to create any contractual commitments or warranties” and that the signed agreement governs.[6] A customer that relies on the no-training statement should ask for it in the SaaS Agreement or the Order Form. The standard DPA lists the conditions on which Kinaxis may use subprocessors, including notice of a change and a 14-day objection period.[12] The Kinaxis subprocessor list on the Trust Center is covered in Kinaxis hosting, data processing and services. OpenAI, which Agent Studio is designed to work with, is not named in that list; Google appears only as a cloud hosting provider, and the other entries are data-centre, security monitoring and content delivery providers.[6]
Programs
- Remix. In June 2026 Kinaxis ran a customer and partner hackathon, called Remix, in which teams built agents with Maestro Agent Studio over a three-week period with Kinaxis mentors.[10] The post states that teams need not be AI experts and describes the competition as a learning programme. It says nothing about licensing the agents built.
- Trial subscriptions. The second-quarter results refer to trial subscriptions to Maestro Agents.[5] No trial terms, duration or conversion rules are published.
Out of scope
- Agent pricing, units and quotas, and any usage-based overage terms. Kinaxis publishes none.
- The terms of OpenAI, Google, Gurobi and NVIDIA products that Maestro components use or interoperate with.
- Partner-built agents and the agent marketplace that Kinaxis said would arrive in 2026.[1]
- The core Maestro applications, covered in Kinaxis Maestro platform and applications.