A Deeper Look Inside the System of Action for Services Businesses

Earlier today, Certinia unveiled the System of Action for services businesses, backed by the largest expansion of our Veda AI suite to date. You can read the press release here. With this post, I wanted to dive deeper into what a System of Action is, what it requires architecturally, why we built it the way we did, and what our commitment to 100% agentic coverage means for the people who use Certinia every day.

For years, traditional PSA has been the industry standard for automating back-office administration for services businesses. Someone still had to read the data, make the call, and enter the next step by hand. That's no longer enough to stay competitive. Our System of Action closes that gap. It reasons through the same project, financial, and customer data, applies judgment, and carries the next action forward itself. This is why we're committed to delivering 100% agentic coverage across the entire services lifecycle — from sales through delivery. Our System of Action transforms PSA from something that automates the work to an intelligent partner that acts on your behalf, clearing the administrative work off your plate so you can spend more of your time on the client value only you can deliver.

What is a System of Action?

A System of Action is technology that understands context, applies judgement, adapts and learns, and then acts on trusted data, turning software from something people manage into something that manages the work for them. 

Unlike traditional automation, which executes predefined steps, a System of Action reasons through changing conditions and coordinates outcomes across an entire process before deciding what to do. For services businesses, Certinia delivers this through three components working together: foundational context, compounding intelligence, and agentic action, powered by Veda.

A System of Action depends most on context and judgment — understanding changing conditions, weighing consequences, and coordinating outcomes across an engagement. Autonomy is part of that picture, but on its own it is only automation with partial confidence.

That’s why the real product of professional services is accountable judgment. A staffing decision, a proposal, a revenue recognition entry: someone is accountable for each of these, whether a person makes the call or an agent does. So we designed for that judgment on both sides: AI that reasons from the full context of a situation before acting, and people working from that same trusted foundation so their own judgment improves in turn.

The Architecture Behind It

The System of Action is built on three layers working together.

Foundational context starts with structured data, the customer, project, and financial records that live in professional services, customer success, financial management, and CRM. To that we add the unstructured reality of daily work: calls, meetings, chat threads, and documents, much of it captured through our recent acquisition of Moonnox. We also capture telemetry, how the system is being used, and metadata, so agents improve based on real usage and feedback rather than static rules alone. Without that complete picture, an agent can summarize an email, but still can’t exercise business judgment.

Compounding intelligence fuses that data with more than 17 years of Certinia's proprietary services domain expertise, business logic, best practice, and reasoning built up across tens of thousands of customer engagements, while also learning each customer's own business rules and best practices so their compounding intelligence reflects what makes their business distinct, not just Certinia's baseline expertise. This is also where we apply deterministic reasoning rather than leaving decisions to probabilistic guesswork. Large language models are inherently probabilistic. That's a reasonable trade-off for drafting an email, but isn't acceptable for closing a set of books. When the work touches accounts payable, accounts receivable, or financial close, the standard is not 99.9% accurate. It has to be completely 100% correct. So we layer deterministic reasoning on top of the LLM, creating a closed loop where every outcome refines future judgment instead of adding more uncertainty.

Agentic action is delivered through three modalities working together: Veda agents that operate within Certinia or Salesforce Agentforce, agentic workflows built directly into the software experience so there is no legacy, human-only path left behind, and Veda Intelligent Actions (VIAs) that extend the same capabilities into whatever tools a team already uses through Model Context Protocol (MCP), including Claude, Microsoft Copilot, and others.

Together, these three layers are what let Certinia move from a system of record to a System of Action: trusted data, judgment built on that data, and action that carries the judgment through to the work itself. In practice, that means agentic capabilities that run the day-to-day business and accelerate sales and delivery across the entire services lifecycle, from the first proposal through delivery, renewal, and expansion. 

Certinia is the only purpose-built System of Action that runs the business and accelerates time-to-value for your clients, uniting all three layers into one connected system rather than stitching AI onto separate tools for each function. The system itself can carry work forward on its own, letting organizations reach the fastest possible time-to-outcome at the lowest cost of service, without sacrificing the judgment a services business depends on.

Certinia's System of Action architecture, spanning Professional Services, Customer Success, and Financial Management.

100% Agentic Coverage Across the Services Lifecycle, Defined

One line from our announcement is worth repeating here because it is easy to misread: our commitment to 100% agentic capability coverage by year end.

This doesn’t mean removing people from the work. It means that anything a person can do today through Certinia's software experience across the services lifecycle — closing a deal, planning and staffing a project, reconciling a ledger, managing customer health — will also be available agentically, through an agent, a workflow, or a Veda Intelligent Action. Coverage is about choice, not replacement. Some organizations will keep a person in the loop for particular decisions. Others will let an agent carry a task through end to end. Both should be equally well supported, and neither should hit a capability gap the other does not.

We are advancing rapidly on this front. While the growth in the sheer number of agentic capabilities and intelligent actions is impressive, the more impressive part is how much of the services lifecycle those numbers now cover. In April, we launched with 10 Veda agents and 64 VIAs, focused mainly on the earliest stages of an engagement: proposals, staffing, and early delivery work. As of today, that has grown to 24 agents and 135 VIAs, spanning sales and planning, delivery and execution, and renewals and expansion; agentic capabilities that run the day-to-day business and accelerate work across the full lifecycle, not just a single function. By year end, we are targeting more than 40 agents and more than 300 VIAs, the trajectory behind the 17 new purpose-built agents we've committed to in today's news. Each new agent and action is chosen to close a specific gap in that lifecycle coverage. The goal is to close every remaining gap between what a person can do in Certinia, and what an agent or action can do alongside them, across sales, delivery, finance, and customer success alike.

Three Ways We Deliver This to Customers

No two organizations adopt AI the same way, and we don't expect them to. In practice, we deliver the System of Action through three paths, and most customers end up using a combination of all three.

The first is agentic SaaS: for people using Certinia every day inside a project plan, a Gantt chart, or a staffing view, the interface itself is already infused with generative and agentic capability. This can be a helpful path for heavy daily users.

The second is Veda agents, running either natively on Certinia's own AI platform or embedded in Salesforce Agentforce for organizations standardized there. Picture a resource manager trying to staff an urgent training request. Rather than forcing a bad match, the agent can explain exactly why a request cannot be filled as asked, offer to adjust a filter such as availability, and then reason through the trade-off between two remaining candidates based on their specific skills rather than availability alone. That is the difference between automation and judgment.

The third is Veda Intelligent Actions delivered headlessly through MCP, for people and tools that live outside Certinia's own interface entirely. For example, a project manager can ask Claude what milestones need attention across their book of business, and Claude will use Veda Intelligent Actions to pull the relevant data directly from Certinia, reason over it, and answer in place, with no separate login required. A consultant can do the same thing to file a timecard from inside Claude, without ever opening Certinia directly.

Salesforce’s recent Claudeforce announcement with Anthropic strongly validates this architecture and signals where the entire ecosystem is headed. While Claudeforce provides a smarter front door into Salesforce, Certinia’s System of Action supplies the domain expertise, unified data, and governed actions that make that front door genuinely operational for running a business. That is why we built VIAs to be headlessly accessible for Claude and other AI models, allowing teams to use their preferred interface to execute the workflows that actually run their services business. 

None of these paths compete with each other. They meet people and teams where their work already happens.

Claude creating a Certinia timecard, demonstrating Certinia's Model Context Protocol (MCP) support for headless, bring-your-own-AI experiences.

What This Looks Like in Practice

The architecture only matters if it changes outcomes, so here is where things stand today. Certinia already has thousands of active Veda users in production, and that number is growing every week as more customers move from piloting Veda to running it at scale. Siemens is running thousands of projects through the platform. Diabsolut has seen a 70% reduction in documentation time. And across our project management customers, we are seeing close to an extra half week of productivity reclaimed per person; time that goes back into client-facing work rather than administrative overhead.

These are still early results relative to where the roadmap is heading, and that is the point. A System of Action is not a single release. It is a foundation that becomes more capable with every project, every timecard, and every decision it supports.

If you want to learn more about the architecture, new agents, intelligent actions, and deployment options we announced, check out today’s announcement. If you’re ready to see the System of Action in practice, join our live webinar Insight to Impact: Introducing the System of Action for Modern Services on September 30. And if you’re attending Dreamforce this year, come talk to us in person — check out our full schedule at Certinia.com/Dreamforce. I hope to see you there!

Frequently Asked Questions