---
type: "post"
id: "office-intelligence-for-long-running-ai-work-3-4eba7c"
slug: "office-intelligence-for-long-running-ai-work-3-4eba7c"
title: "Office Intelligence for Long-Running AI Work (3 minute read)"
description: "Office Intelligence frames the unit of knowledge-work research as a continuing environment in which agents move from requests through evidence and artifacts to review and the next task."
retrieval_nugget: "Office Intelligence frames the unit of knowledge-work research as a continuing environment in which agents move from requests through evidence and artifacts to review and the next task."
published_at: "2026-10-20"
updated_at: "2026-10-20"
record_date: "2026-10-20"
date_kind: "published_at"
topics: ["long-running-agents","knowledge-work","agents"]
source_urls: ["https://officeintelligence.github.io/"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:office-intelligence-for-long-running-ai-work-3-4eba7c"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:bed2b308eb92250cd79a32e4d4e79518e9ba6c80ac27bbb291b2e86f4f8c0c05","reviewed_at":"2026-10-20T14:09:03.498Z","source_evidence_count":1,"verified_claim_count":3,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":3,"implication_count":2,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"Office Intelligence treats continuing work environments—with evidence, review, and retained context—as the research object for knowledge-work agents.","observed_facts":[{"text":"The site organizes the program around environments, evaluation, training, and harnesses.","source_urls":["https://officeintelligence.github.io/"]},{"text":"The described task-time workspace combines authorized files, tools, organizational state, and task context; deliverables are checked against task-specific criteria and evidence links.","source_urls":["https://officeintelligence.github.io/"]},{"text":"The site presents an OfficeTown reconciliation example with an editable workbook deliverable and supporting records.","source_urls":["https://officeintelligence.github.io/"]}],"mechanism":"A task-time workspace combines authorized files, tools, organizational state, and task context; deliverables are checked against task-specific criteria and evidence links, while execution traces and working memory can inform later tasks.","why_now":"Long-running work makes continuity, evidence, and review part of the agent interface rather than optional prompt context.","implications":["Evaluation can focus on whether an agent produces an inspectable artifact under task-specific constraints, not merely whether it answers a prompt.","Retained work traces and working memory may connect assignments, but their contribution to later-task quality remains to be demonstrated."],"evidence_boundary":"The public site describes a research program, environments, and examples; it does not establish validated performance in real organizations or prove that context reliably transfers across tasks.","watch_conditions":["Watch for published evaluations showing whether retained work traces improve later-task outcomes without weakening evidence quality or reviewability."],"related_records":[],"new_branch_reason":"No prior New Runtime record is supplied, so this starts a distinct analytical record on continuing work environments as an agent research and evaluation object."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/office-intelligence-for-long-running-ai-work-3-4eba7c/","markdown":"https://newruntime.com/posts/office-intelligence-for-long-running-ai-work-3-4eba7c.md","json":"https://newruntime.com/posts/office-intelligence-for-long-running-ai-work-3-4eba7c.json"}
related_patterns: ["ai-native-orgs-move-to-review-and-orchestration"]
related_shifts: ["fixed-teams-to-executable-mandates"]
---

# Office Intelligence for Long-Running AI Work (3 minute read)

## Retrieval answer

Office Intelligence frames the unit of knowledge-work research as a continuing environment in which agents move from requests through evidence and artifacts to review and the next task.

Office Intelligence frames the unit of knowledge-work research as a continuing environment in which agents move from requests through evidence and artifacts to review and the next task. The site describes a research program spanning environments, evaluation, training, and harnesses. Its thesis is: Office Intelligence treats continuing work environments—with evidence, review, and retained context—as the research object for knowledge-work agents.

The mechanism is: A task-time workspace combines authorized files, tools, organizational state, and task context; deliverables are checked against task-specific criteria and evidence links, while execution traces and working memory can inform later tasks. The site also gives an OfficeTown example involving a reconciliation role and a reviewable workbook. The why_now is: Long-running work makes continuity, evidence, and review part of the agent interface rather than optional prompt context. The first implication is: Evaluation can focus on whether an agent produces an inspectable artifact under task-specific constraints, not merely whether it answers a prompt.

The evidence boundary is: The public site describes a research program, environments, and examples; it does not establish validated performance in real organizations or prove that context reliably transfers across tasks. The first watch condition is: Watch for published evaluations showing whether retained work traces improve later-task outcomes without weakening evidence quality or reviewability. If those results appear, the project’s framing can be assessed as more than a research design. No prior New Runtime record is supplied, so this starts a distinct analytical record on continuing work environments as an agent research and evaluation object.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Long Running Agents](https://newruntime.com/topics/long-running-agents.md): Explore the long-running-agents topic hub.
- topic hub: [Knowledge Work](https://newruntime.com/topics/knowledge-work.md): Explore the knowledge-work topic hub.
- topic hub: [Agents](https://newruntime.com/topics/agents.md): Explore the agents topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)