Squirro announced general availability of an Agent Catalog in late July: 13 prebuilt, production focused agents covering finance, HR, legal, sales and IT. The detail worth paying attention to is not the number of agents. It is that all 13 share one foundation, the same connections, the same compliance approvals and the same knowledge layer, rather than each being built and cleared separately.
For a regulated enterprise, where every new AI tool historically triggers its own fresh security, legal and data access review, that shared foundation is arguably the more valuable part of the announcement.
What a Reusable Agent Foundation Actually Is
An agent's individual capability, what it can research, draft or action, is the part most product announcements lead with. The foundation underneath it, the data connections it draws on, the compliance approvals confirming it is allowed to access what it accesses, and the knowledge layer it references, is usually invisible unless something goes wrong.
Building that foundation once and reusing it across multiple agents is a fundamentally different approach to building each agent, and its underlying access and compliance case, from scratch. It is closer to how mature software organisations build a shared platform once and let multiple products draw on it, applied to an area, AI agents, that has mostly been treated as one-off projects until now.
Why Vendors Are Converging on This Model at the Same Time
Squirro is not the only vendor moving this direction. SAP's recent consolidation of its Business Technology Platform, Business Data Cloud and Business Transformation Management under a single three layer architecture, context, build and govern, reflects the same underlying recognition: governance needs to be built into the platform layer once, not rebuilt for every agent sitting on top of it.
That convergence, from more than one major vendor, in the same few weeks, is a reasonable signal that this is a genuine shift in how the industry is thinking about agent deployment, not a marketing angle from a single company.
The Risk of Getting This Backwards
Picture an organisation approving five separate customer service agents over five separate security and compliance reviews, each covering essentially the same data connections and the same knowledge sources, because nobody built the shared foundation first. Every agent takes roughly the same time to clear as the last one, and the organisation never gets faster at deploying agents even as it deploys more of them.
This is the trap that a reusable foundation is specifically designed to avoid, and it is also the default outcome for any organisation that treats each new agent request as its own isolated project rather than asking whether the underlying access and compliance work has already been done.
What to Actually Demand From a Custom Build
For any organisation building or buying agents rather than adopting a prebuilt catalogue, the practical question is whether the connections, compliance approvals and knowledge layer are being built as reusable infrastructure or as a one-off attached to a single agent. Asking that question before the first agent is built, rather than after the fifth agent takes as long to clear as the first, is the difference between a foundation that compounds in value and a series of unconnected projects that each start from zero.
The Australian Angle
Reusable foundations matter more, not less, in the Australian market specifically, given the integration cost of connecting AI to legacy ERP, banking core and clinical systems already discussed in earlier research on custom AI development in Australia. A compliance and connection layer built once against those systems, and reused across multiple agent use cases, spreads that integration cost across every agent that draws on it rather than repeating it in full for each one.
What This Means for Your Organisation
What we see across implementation engagements is that the organisations getting faster at deploying agents over time are, without exception, the ones that treated the underlying foundation as infrastructure worth building properly the first time. The ones still treating every new agent as an isolated project are the ones whose fifth deployment takes just as long as their first.
Key Takeaways
- Squirro's newly launched Agent Catalog ships 13 prebuilt production agents sharing one reusable foundation of connections, compliance approvals and a knowledge layer, rather than 13 separate builds.
- SAP's recent platform consolidation reflects the same shift, building governance into a shared platform layer rather than into each individual agent.
- Organisations that clear each new agent through its own separate security and compliance review, without a shared foundation, never get faster at deployment no matter how many agents they build.
- The practical question for any custom agent build is whether the connection and compliance layer is reusable infrastructure or a one-off attached to a single agent.
- Reusable foundations matter especially in Australia, where legacy system integration cost can be spread across multiple agents rather than repeated for each one.
How Trusenta Can Help
AI Agents and Automation builds the reusable connection and compliance foundation this post describes, designed to support multiple agents rather than just the first one.
AI Integration Services connects that foundation to legacy ERP, banking core and clinical systems once, rather than repeating the integration for each new agent.
Enterprise Architecture documents the shared foundation as reusable infrastructure so future agents can be assessed against what already exists.
Conclusion
The prebuilt agents in Squirro's new catalogue are not really the story. The story is the industry quietly agreeing that the foundation underneath an agent, the connections, the approvals, the knowledge it draws on, is worth building once and reusing, rather than rebuilding every time. Organisations still building agents one isolated project at a time are the ones that will feel the difference first.
