The latest wave of AI-powered process orchestration is here—and it packs quite the punch.
It’s called agentic orchestration.
Agentic orchestration—defined by specialised AI agents that organisations can deploy to autonomously orchestrate back-office processes on employees’ behalf—represents an exciting new application of AI in the enterprise, which is perhaps why so many vendors are racing to put agentic orchestration solutions of their own.
In February 2025, Tonkean, too, made its agentic orchestration and autonomous AI agents generally available. However, Tonkean’s agentic orchestration technology is built differently. Tonkean Agentic Orchestration layers autonomous AI with deterministic, rules-based automation—meaning, Tonkean agents can work on their own to automate complex processes as well as pursue long-term goals, but only ever within a set of guardrails and rules established by humans. Importantly, and unlike other orchestration vendors, Tonkean also comes with over 200 prebuilt integrations, meaning Tonkean’s agents can autonomously orchestrate work across departments and across every piece of technology an organisation might use—as opposed to solely within one walled garden.
The potential of agentic orchestration for internal teams like procurement, whose key processes are inherently cross-functional and touch many different technology environments, is immense. Telling us all about it is Sagi Eliyahu, Co-Founder and CEO at Tonkean. In this exclusive article, CPOstrategy explores how Eliyahu’s organisation is redefining the possible in procurement via agentic orchestration.
In your words broadly, can you introduce what agentic orchestration is?
Sagi Eliyahu: “Agentic orchestration is how you instrument AI agents for enterprise. It’s an approach that puts agents alongside employees to coordinate workflows, execute tasks, and drive outcomes—all while following configurable policies and guardrails.
“Central to this approach is the ability to combine autonomous AI that can communicate with users and act on its own with deterministic orchestration that follows rules-based workflows to carry out processes across many applications and data sources.
“Agents can operate alone or in collaboration with other relevant agents in a multi-agent architecture and run both as a chat interface and independently in the background. The multi-agent architecture enables companies to distribute the deployment and administration of specialised enterprise agents throughout the organisation, putting control into the hands of the people with the right subject-matter expertise and authority.”
What about Tonkean’s own approach to agentic orchestration?
Sagi Eliyahu: “For AI agents to be truly useful to an enterprise, they must be easily accessible, able to execute work across your entire organisation, and able to autonomously drive business outcomes—while always carefully following your organisation’s policies.
“Tonkean is the only agentic orchestration platform that delivers on all these points. Our biggest differentiator is combining agents with a true orchestration platform. Orchestrating inside one system isn’t true orchestration. Orchestration happens when you work outside—and in turn transcend—the boundaries of individual systems. The value of orchestration in this case, then, is that it allows you to make AI accessible to employees at strategic points in any process and across all your organisation’s departments and tools.
“This is crucial. Unlike other recent agentic and process orchestration offerings, which generally lack the level of control, accessibility, and interoperability enterprises need to derive true transformational change out of AI, Tonkean can integrate with every kind of enterprise technology, from Slack to SAP, cloud applications to on-prem databases and in-house tools. This allows Tonkean Agentic Orchestration to surface intelligent, specialised AI agents directly to employees in the environments where they already work in accordance with how they like to work, and to automate processes that span many different data systems and departments.
“Tonkean Agentic Orchestration is also 100% no-code, meaning internal enterprise teams like procurement can build, deploy and orchestrate agents themselves. (Though the Tonkean library offers ready-to-use agents that make it even easier to start automating complex processes right away, such as Sourcing Specialist Agents, Buyer Agents, Contract Manager Agents, Purchase Intake Agents, AP Specialist Agents, Market Analysis Agents, Compliance Officer Agents, and many others.)”
Sagi, I read you said business processes are not about data or technology, but instead about people. In your view, can you share how Tonkean Agentic Orchestration is cutting through the noise and making life easier for humans in ways different from before?
Sagi Eliyahu: “Making life easier for human employees—by transforming the way humans interact with enterprise technology—has always been a primary inspiration for us as a company. We’ve long felt that, for employees, following internal policies like purchase intake, for example, should feel so easy you don’t even realise you’re following a prescribed process. Rather, the process feels so intuitive—and the technology powering the process so accessible and personalisable and dynamic—that it feels natural with how you’d want to be working in the first place.
“Agentic Orchestration represents a quantum leap in how we’re able to facilitate experiences that feel like that. Now, with Tonkean, any employee can access a full staff of specialised AI experts directly from within whatever application environment they work in—email, Slack, Teams, etc. Employees can assign those agents work—from purchase requests and compliance validation to research and reporting. But internal teams like procurement can also curate guided experiences that are exponentially faster and more seamless than anything they’ve been able to provide before.
“A critical part of being ‘human-centric’ is keeping humans in the loop at those critical moments when important decisions are called for – such human touchpoints can be defined in Tonkean’s no-code process editor, but the agent itself also has discretion to ask the human operator for clarification, direction, and decisions.
“Using a simple process as an example, now, when an employee from marketing wants to buy something but they’re not sure how, all they need to do is call the Tonkean AI Front Door inside Slack and ask. Tonkean will tap whichever agent is most appropriate, depending on the request, and that agent—or whichever other agents that agent decides it needs to collaborate with to give the employee what they need—will guide them to resolution. And the employee will be able to interact with these agents as they would a human administrator, because agents communicate like us. Orchestration and automation handle all the data structuring—you know, the toggling between systems employees used to have to do manually—in the background.
“That’s just one example, but it’s an example of how we’re able to help procurement teams create business processes that truly put people first, in the sense that we no longer have to ask people to conduct lots of frustrating manual work in order to use technology.”
What are Tonkean AI agents doing for its customers today and why is this such an exciting announcement?
Sagi Eliyahu: “With Tonkean Agentic Orchestration, enterprise teams are right now configuring agents to answer questions about policies to ensure compliance, perform actions and query information across all of their organisation’s internal systems, coordinate and execute complex work to deliver on long-term initiatives, and produce personalized experiences with custom user interfaces on the fly.
“Here are a few easy to start with examples we’re seeing a lot of:
- Policy Q&A – Agents are connected to company policies, procedures, and other internal resources to act as the source of truth for employee requests. More than simply answering questions, Tonkean Agents can recommend processes or guide users to intake experiences for key tasks as well.
- Contract review and generation – AI agents perform common tasks related to contracts like NDAs and MSAs, as well as other ubiquitous enterprise documents. For example, some users have created agents that can review a contract, understanding what terms are acceptable and which terms should be redlined. The agent is also capable of accessing your connected data sources and the web to fill in missing information.
- Monitoring connected apps and responding when needed – AI agents track a data source, such as a CLM, CRM, ERP or any other system, and take action when certain conditions are met or updates occur. For example, you can ask an AI Agent to notify you when your vendors pay an invoice, or kick off a workflow to “nag” the vendor if they don’t after a certain time. The agent can send you a Slack message or email—and you can reply back to it for follow-up.
- Integrated process partner – In addition to monitoring data sources and reviewing documents, you can give AI agents a larger set of skills and collaborate with them to carry out core parts of your process. Say, for example, you have a sourcing agent as part of your procurement workflow. You can connect this agent with your core procurement system, including your suppliers, document repositories, etc., and configure a set of skills for the agent that allow it to research and select ideal suppliers. From there, the agent can send an RFI to the supplier and collect the response. In this way, you can effectively rely on the agent to carry out basic tasks in a way that’s fully aligned with your policies.
“Though, to reiterate, this is all just the start. Agents can be configured to be much more specialised—in accordance with specific industries and types of work—and they can collaborate with each other to conduct complex work and achieve audacious goals over time.”
In what ways does agentic orchestration impact compliance and governance?
Sagi Eliyahu: “Tonkean Agentic Orchestration improves compliance, in the sense that it reduces the capacity for human error, but always keeps humans in control. Tonkean AI Agents come strategically scaffolded with deterministic and nondeterministic capabilities. In other words, agents can determine on their own how to meet the responsibilities and achieve the goals you set for them—this is a big part of what makes them so powerful and elevates agents beyond the realm of chatbots or task automation—but, importantly, only ever within the boundaries you establish for them. Humans create and manage the guardrails governing what the agents can and cannot do, when agents can take action on their own, and when they should escalate critical decisions to the right people for review. Meaning, you get the power of best-in-class AI, the improved accuracy of automation, and improved compliance because humans remain in control.”
How do you see AI agents and Tonkean’s own approach evolving over the next few years?
Sagi Eliyahu: “A few areas:
- We’ll continue to invest in the user experiences that provide a frame for human-agent interaction. Most people’s exposure to AI tools has been exclusively through chatbots. Chatbots are useful for some things. But for enterprise processes, what’s really needed is software environments that can act as collaborative canvases between the user and the agent. Creating reports and dashboards on the fly, visualising the steps within a process that’s been initiated, dynamically bringing in actions and functionality from third-party applications – creating a UX structure for agentic workflows that goes beyond the very linear and ephemeral nature of a chatbot.
- You will see Tonkean continue to roll out purpose-built agentic functionality for specific functions: procurement, legal, IT, HR, etc.
- AI moves fast. The major LLM providers roll out new models with new capabilities seemingly every week. We’ll evolve alongside these new capabilities to best incorporate their strengths while helping enterprises adopt them responsibly.”
Is there anything else you would like to share?
Sagi Eliyahu: “We’re really only in the beginning stages of all this, in procurement and everywhere else. Most companies are looking at these technologies but haven’t seen the ROI yet. This year one of our primary goals at Tonkean is to change that.
“That said, we view this new era as something of a full-circle moment for us. We’ve been working on this kind of technology for a long time. It’s sort of a never-ending journey, because our mission is to bring the best technology and tools to every part of the business so people can do the work they actually want to do, unlocking human potential. But some companies are still far behind, and with AI and agents, there’s real potential to leapfrog some of their biggest problems. I’m interested to see how we can help lagging companies skip some early steps and move directly to agents.
“The biggest value of LLMs and AI, in my opinion, is that they communicate like us. When we started the company, I felt technology wasn’t built for people — it was built for data. But business processes aren’t about data; they’re about people. That gap was why we started the company and it remains present nearly a decade later.
“But AI and agents have the potential to leapfrog some of these challenges by allowing us to communicate naturally while orchestration and automation handle data structuring in the background. That’s where we’re headed.”
