Three back-office items this week, each verified and linkable. We are running three strong items rather than padding to five. For the screening framework behind our reads, see the Three-Signals Test.
1. Rockland FCU automates loan QC and publishes actual numbers
Rockland Federal Credit Union, a $3.8 billion Massachusetts institution rebranding as Arise Financial, deployed Kintera AI’s workflow platform on its indirect auto loan post-closing quality control process. The claims: live in 10 days with no IT integration, per-file review time down from 20 minutes to 2, full cross-referencing of loan file data points, and $250,000 in documented annual savings. Our take: this is a textbook signals workflow: repetitive, procedural, and short per item. It is also a vendor press release, so treat the 10-day figure as a best case measured after the SOPs already existed in writing. The transferable part is the precondition: Rockland could automate the process because it was already written down. If your QC procedures live in someone’s head, writing them down is step one, and it needs no vendor.
2. Arizona Financial buys the platform after building the org chart
Arizona Financial Credit Union, $3.9 billion with 180,000 members, selected Creatio’s CRM and workflow platform to unify sales and service operations and coordinate branch and back-office teams, with AI-driven automation of routine administrative tasks on the roadmap. The detail worth noticing here is the title of the person quoted: vice president of AI and development. Arizona Financial put an accountable AI executive in place before the platform decision, and it named measurable success criteria (employee adoption, operational efficiency) up front. A selection announcement is not a deployment, so file the outcome claims under pending. The org chart came first, though, and that part is already real.
3. Tech CU builds the knowledge layer before the AI layer
Technology Credit Union, $4.5 billion with 170,000 members, selected eGain’s AI Knowledge Hub to centralize governed knowledge across its contact center, digital banking, branches, and back office before expanding AI on top of it. Tech CU’s framing: AI output is only as trustworthy as the knowledge powering it, so the knowledge gets governance first. Unglamorous, and probably the move that determines whether its member-facing AI works two years from now. A large share of AI chat failures trace back to the knowledge base underneath: the model confidently serves an outdated fee schedule because nobody owns the fee schedule. Tech CU is putting ownership on the answers before it buys more ways to serve them.
The pattern
All three institutions sit in the $3.8B to $4.5B band, and all three are moving on operations before member-facing novelty. Rockland documented the process first. Arizona Financial hired the owner first, and Tech CU is governing the knowledge before it expands anything that answers members. If you are picking your own first workflow, the 90-day pilot plan covers the path from shortlist to a scale-or-kill decision, and the back-office automation pillar holds the full screening kit.
Get this roundup in your inbox every Thursday: subscribe to the AiForCU newsletter.