Academy
6. Managing History and Relaunching Researches
Find past sessions in History, review their status, and use Try Again to refine a failed or thin investigation.
Every Deep Research session you launch is automatically saved in History. This dedicated view stores the original question you asked, the sources you selected, the current status of the session (running, completed, or failed), and the full sourced report once it's ready. You don't need to keep the tab open while a research runs — sessions execute in the background, so you can navigate away and return to History at any time to check progress or reread a finished report. Think of History as your research archive: nothing is lost, and every investigation remains accessible for review or follow-up.
Each session in History displays a clear status so you can quickly triage your past work. A completed session holds the full cross-checked report with inline citations. A failed session — or one that returned thin, unsatisfying results — is not a dead end. Open that session and you'll find the Try Again option. This lets you relaunch the investigation with a reformulated question or a different set of sources. The key insight: very broad questions tend to dilute results. When retrying, narrow the scope — name specific entities, define a time period, and state the angle you care about. This precision is what drives report quality.
Before hitting Try Again, review the original report (even a thin one) to spot what's missing. Then reformulate: swap vague terms for named entities, add a date range, or switch source types — e.g., swap web for academic providers if you need peer-reviewed data. Fewer, well-chosen sources often beat Select All for focused topics.
Source selection matters just as much on a relaunch as on a first attempt. If your original research used only web sources and came back shallow, adding academic or specialized providers may surface deeper material. Conversely, if a broad Select All query was slow and noisy, trimming to two or three targeted providers can tighten the report. Remember that cost scales with the number of sources and synthesis depth, so a focused retry can also be more economical. If certain providers require API keys you haven't added yet, head to Settings to enter them and unlock those sources for your next attempt.
Once a relaunched report meets your needs, remember the output is portable. Copy the rich-text report — with its inline citations intact — into Documents, ChatPro, or any workflow. The sourced format keeps your material verifiable wherever you reuse it.
Open your Deep Research History now. Find a past session — ideally one that returned limited results or failed. Open it, review the original question and sources, then hit Try Again. Reformulate the question to be more specific (add an entity name, a date range, or a sharper angle) and adjust your source selection before relaunching.
History is your research command center: every session is preserved with its question, sources, status, and full report. When a session fails or feels thin, use Try Again to relaunch with a sharper question or better-matched sources. Precision — named entities, time frames, targeted providers — is the single biggest lever for improving results. And every completed report can be reused across OUPI with its citations intact.