A wave of product updates and partnerships shows vendors embedding generative and agentic AI into ERP, subledger and procure‑to‑pay workflows — accelerating automation while adding confidence scoring, role‑based permissions and model connectors to address auditability and security concerns.
According to Solutions Review’s weekly round‑up for the week of 15 August 2025, a fresh wave of product updates and partnerships underlines how AI is reshaping enterprise resource planning and process automation. Vendors ranging from niche applied‑AI startups to hyperscalers announced features designed to accelerate automation, tighten governance and make it easier to bolt best‑of‑breed models into core ERP and financial workflows — even as questions about trust, security and auditability remain front and centre.
Certinia’s Summer ’25 release exemplifies the push to embed agentic AI into service‑centric ERP. The company says the update brings generative capabilities to project forecasting and resource planning, and introduces purpose‑built assistants — a Staffing Agent to accelerate resource allocation and a Customer Success Agent to create account summaries and communications. According to Certinia’s own blog, the release also adds RAID (risks, assumptions, issues, dependencies) tracking, tighter estimate controls and dashboards that surface objectives within account health scores — features pitched at mid‑market professional services and customer success teams that need faster decision cycles without sacrificing governance.
Parallel moves show vendors sequencing automation deeper into financial processes. Sage announced a suite of AI‑powered tools for Sage Intacct aimed at automating subledger reconciliation, line‑level matching and procurement controls, and confirmed vendor payments integration powered by MineralTree. The company frames these additions as productivity and compliance aids for finance teams managing growth and increasingly complex audit requirements.
Integrations that connect orchestration platforms to industry finance stacks are also proliferating. Cleo said it has teamed with Open ECX to combine Cleo Integration Cloud’s EDI, API and managed‑file capabilities with Open ECX’s Procure‑to‑Pay platform and AI‑driven document extraction. The partners say the joint solution will digitise invoices and payments for construction and other sectors, reduce manual touchpoints and provide pre‑built connectors to major ERPs including SAP, Oracle and NetSuite. “Together, we’re helping businesses eliminate manual touchpoints in their most critical business processes, reduce latency, and gain full visibility into their transactional lifecycles,” Cleo CEO Mahesh Rajasekharan said in the announcement.
That last point — visibility and trust — is precisely what several suppliers are trying to address. Fisent Technologies introduced a Confidence Rating capability for its BizAI process‑automation platform that surfaces semantic and multi‑model similarity scores and routes low‑confidence outcomes for human review. Fisent positions the feature for regulated industries such as finance and insurance, emphasising model‑agnostic deployment and audit‑friendly explanations to reduce false negatives and speed regulator‑facing reviews.
The need to govern external AI models while preserving flexibility is also driving platform work. Oracle’s NetSuite has launched an AI Connector service that implements the Model Context Protocol to allow customers to bring their own or third‑party models into NetSuite workflows. NetSuite’s documentation outlines role‑based permissions, OAuth 2.0 authentication, SuiteApp installation steps and recommended mitigations — signalling that vendors expect administrators to exercise granular control over what models can access and do inside ERP systems.
That same theme is visible at cloud scale. Oracle and Google Cloud confirmed an expanded partnership to make Google’s Gemini multimodal models available to Oracle Cloud Infrastructure customers through OCI’s Generative AI service, starting with Gemini 2.5 and surfaceable via Vertex AI and, in time, Oracle Fusion Cloud Applications. Thomas Kurian, CEO of Google Cloud, said the integration will make it easier for Oracle customers to begin deploying AI agents that support developers and streamline data integration tasks. Oracle and Google say customers will be able to use Oracle Universal Credits to access Gemini models; the companies pitched the move as increasing enterprise model choice while emphasising security and enterprise controls.
On the enterprise observability and resilience front, ScienceLogic updated its AIOps tooling with enhancements in predictive analytics and intelligent automation to prevent incidents across hybrid environments. Rockwell Automation published the cybersecurity chapter of its 10th annual State of Smart Manufacturing report, drawing on feedback from over 1,500 manufacturing leaders and noting that more than 60 percent of security and IT professionals plan to adopt AI and machine learning for security in the next 12 months — a marker of how rapidly defenders are turning to the same technologies attackers exploit.
Other items in the round‑up point to specialised marketplaces and healthcare pilots. SuperOps launched an Agentic AI Marketplace in partnership with AWS to give managed service providers access to deployable autonomous agents, and announced a global “SuperHack” developer challenge to build real‑world IT agents. In healthcare, Qualtrics said it will deploy generative‑AI agents with Stanford Health Care to automate routine interactions and act on experience feedback to improve patient, provider and staff journeys.
Taken together, the announcements sketch a market moving from experimentation to pragmatic deployment: vendors are not only putting generative models into ERP and BPM canvases but are also adding controls — confidence scoring, role‑based gates, authentication protocols and audit trails — that will determine how quickly finance, manufacturing and regulated sectors embrace agentic tools. The balance firms seek is clear: unlock the efficiency and insights AI promises, while maintaining the traceability and security that auditors and boards expect.
For practitioners weighing adoption, the immediate tasks are practical: map high‑value processes that can tolerate automated decisioning, insist on model‑agnostic integration layers and require explainability and human‑in‑the‑loop gates where compliance risk is material. Vendors are racing to provide those controls, but as these weekly updates show, the maturity of governance will be the deciding factor in which projects scale from pilots to production.
Source: Noah Wire Services
Subscribe to Industry Updates
