In Q3 2026, Graip.AI took a step forward to making day-to-day work with agents clearer. You can now see where every thread stands, count on agents that remember earlier conversations, and capture data from the whole request instead of the attachment alone. Read on to see how each update works and what it means for your team.
New Features and Improvements
Work View
Agents Lab now includes a Work View that shows the status of every current thread. Once agents handle real volume, the first question is always the same: what is running, what is waiting, what is done. Work View answers it on one screen, so teams can check progress instead of assuming where a request stands.
Improved Document Viewer
The document viewer inside agents has been upgraded, continuing the work of embedding IDP (Intelligent Document Processing) into Graip.AI Agents. Reviewing and correcting an agent’s document work is a daily task for the people who use it, so the quality of that screen is felt more often than most features.
Agent Long-Term Memory
Agents now have long-term memory, so context carries beyond a single thread. Without memory, every interaction starts from zero, and you supply the same context each time. With it, an agent builds on what came before. Repeat work gets quicker, and the agent behaves consistently across a working relationship instead of session by session.
Email Extraction Across All Header-Field Pipelines
All header-field extraction pipelines now read the email itself, not only the attached document. Business documents rarely arrive alone. They arrive in an email that carries part of the context. Reading both means the request is processed as one thing, and nothing gets lost between the message and its attachment.
Spotlight Use Cases
SAP CPQ Business Requirements Automation
The Challenge: In an SAP CPQ migration project, a large share of the effort goes into capturing requirements that already exist but are locked inside legacy code and configuration. Teams spend hours reading scripts and audit results by hand and rewriting them as business requirements before the new system design can even begin.
The Solution: Graip.AI produces business requirements directly from legacy system scripts and audit results. Its agent reads the existing code and configuration, extracts the underlying business logic, and drafts the requirements automatically, roughly eight times faster than manual capture. Subject-matter experts no longer start from a blank page. They review, correct, and approve what the agent drafts.
The Outcome: Requirements capture is no longer the bottleneck at the start of a migration, and the team spends its time on review instead of manual work.
Finance Integration Finalized
The Challenge: Self-billing invoices arrive in varying formats and have to be read, checked, and keyed into the finance system before they can be processed. This is slow, repetitive work that scales badly with volume.
The Solution: Graip.AI delivered an automated IDP solution for self-billing invoices, integrated end-to-end with a client’s finance process. The solution extracts and validates the invoice data and passes it straight through to the downstream system, in line with the SAP Self-Billing Cockpit process.
The Outcome: The finance integration is complete and live for the client. Self-billing invoices flow from intake to posting without manual re-keying, with people involved only where judgment is needed.
Putting It All Together
Q3 takes the manual effort out of working with agents, whether they are capturing migration requirements or processing self-billing invoices. You can track every thread in one place and rely on context that carries over between conversations. Requests arrive complete, so your time goes to decisions instead of double-checking.