Companies are building parts of the idea: shared knowledge, management-meeting support and agents that turn messages into work. The selected sources do not establish a complete company brain with every conversation and an AI CEO.
Twelve selected examples: nine energy companies or sponsors, two named practices outside energy, and one software product offer. “Company brain” is our comparison lens, not a shared product name or an industry adoption count.
Quantum Firm Intelligence / Claude Enterprise / QEC
Energy private equity / Operating components; broader system being built
Quantum has articulated a firm-wide memory system, with existing AI and data platforms beneath it.
Its 2026 report names Claude Enterprise as QIQ’s successor. QEC joins financial, operational and market data; reported source categories include board records and investment outcomes. The report describes recurring fund-operation workflows and automated inquiries. Undated Lakebase foundation: a Databricks story describes a shared relational layer for geological, production and financial deal data; this does not establish an AI decision-maker.
Quantum Firm Intelligence is the broader goal: retain the context behind investments and apply it to later decisions.
Email, chat and meeting capture are not established. Neither complete decision follow-through nor delegated investment authority is demonstrated. Usage, quality, full costs and measured returns are undisclosed. The report generally uses a December 2025 information cutoff. Lakebase boundary: query and table counts are not active users, investment returns or proof of AI use; publication/adoption dates and full costs are unknown.
This is the clearest stated firm-intelligence ambition in the selected energy sources. Ask how a past assumption, its decision and its eventual outcome remain linked.
Quantum 2026 Stakeholder Report
Quantum Capital Group · 2026; exact publication day not stated · PDF pp. 11–12: data, Firm Intelligence, Claude/QIQ, QEC and firm workflows; p. 68: information-date disclaimer
Issuer report reviewed 9 October 2026. Current components and the broader goal are distinguished.
Transforming energy deal evaluation with Lakebase
Databricks · Unknown · Building a consistent foundation; Lakebase relational capabilities; adoption and FAQ paragraphs
Full original vendor customer story read by editor. No publication date shown; search age and footer year are not rollout dates. Vendor/customer claims, not independent validation.
Microsoft 365 Copilot / Knowledge Agents
Oil & gas operator / Customer-reported use
Aker BP joins meeting assistance with shared knowledge agents and routes difficult questions to experts.
Microsoft’s June 2025 customer story describes Teams transcripts and summaries, plus live agents answering questions and helping staff reach experts. Copilot Studio agents include human-in-the-loop features.
The story describes wider licensing and more business agents as the next stage.
Transcripts are not proof of a complete communications archive or decision ledger. The reported 96% active-user figure has an unclear denominator. Net return and autonomous follow-through are not established.
Capturing a meeting is only the start. Preserve the reason for a decision, its owner and the next check.
Aker BP builds an AI-first future with Copilot and agents
Microsoft customer story / Aker BP · 23 June 2025 · Driving a company-wide digital transformation; AI-powered technology yields results; Pioneering AI
Vendor-authored customer account reviewed 9 October 2026; not an independent performance audit.
Copilot / Virtual Management Meeting agent
Oil & gas operator / Customer-reported released agents
INPEX has a named agent that helps people prepare for management discussions.
Microsoft describes a Virtual Management Meeting agent that reviews uploaded materials, suggests questions and surfaces overlooked perspectives. It also names shared regulatory agents and a Copilot environment spanning SharePoint, OneDrive, Office and Teams. Separately, its SAP Cloud ERP began operation in Japan in January 2026, following a process-standardization program.
Further process integration and automation remain an ambition.
This is meeting preparation, not evidence of continuous executive monitoring or automated decisions. Complete communications capture, a decision ledger and verified follow-through are undisclosed. The source’s broader AI value estimate is not a return for this agent. That ERP disclosure does not establish integration with these agents or their access and action permissions.
A management-paper review is a concrete first test: can it catch material omissions before the meeting?
INPEX transforms business processes with Copilot and AI agents
Microsoft customer story / INPEX · 1 September 2026 · Copilot’s impact as a research tool; Building a foundation for change with AI agents, especially Virtual Management Meeting and approval passages
Vendor-authored customer account reviewed 9 October 2026. Meeting preparation is distinct from executive authority. Reopened 14 October 2026; now linked to register profile AI-294.
INPEX commences operations of SAP Cloud ERP
INPEX / SAP Japan / ABeam Consulting · 24 June 2026 · PDF pages 1–2; Japan go-live and process-standardization approach on page 2
Official PDF read and page2 rendered17October2026. Enterprise-system context; no AI implementation or ERP-to-Copilot integration established.
Digital Work Environment / Land AI remit
Oil & gas operator / Digital-work foundation; AI executive layer unverified
EQT already describes the communication and shared-work foundation a company brain would need.
EQT’s 2025 10-K describes DWE as its primary communication and collaboration platform, with critical work processes and shared operational data. The filing does not identify its vendor.
A current Land data-science job brief names Salesforce, Databricks, AI pipelines and a mandate to turn Claude Enterprise into measurable Land outcomes. This is evidence of intended work, not an integrated live deployment.
Neither every communication occurring in Salesforce nor an AI executive assistant is established. The Land hiring remit does not prove delivery, usage or outcomes. April 2025 exploratory-AI remarks are historical.
An existing shared workspace may make the next step easier. Verify which decisions and commitments it actually records.
EQT 2025 Form 10-K
EQT / SEC filing · 18 February 2026; reporting year 2025 · Human Capital Resources, printed p. 32: primary collaboration platform and shared operational data
Operator filing reviewed; primary does not mean exclusive or exhaustive. Vendor and AI executive layer are not identified.
Life at EQT — Remote Work
EQT · Undated; checked 9 October 2026 · Remote Work and Current Openings sections
Current operator description reviewed. It does not identify an AI assistant.
Talking AI, climate goals with EQT
S&P Global interview / Courtney Loper, EQT · 18 April 2025 · Transcript: Salesforce, digital workspace and exploratory AI remarks
Direct executive interview reviewed; the statement is historical.
Land Data Scientist II/III/Senior
EQT job posting · Undated; checked 9 October 2026 · Responsibilities: STAQ team, Salesforce, Databricks, AI pipelines and Claude Enterprise; preferred experience
Hiring mandate reviewed; this is not proof of completed integration or adoption.
OneDEC / Snowflake / ServiceNow
Oil & gas operator / Company-described governed data and agents
OneDEC is a useful example of linking company data to work people can approve and inspect.
Diversified describes operational and financial data in Snowflake, ServiceNow playbooks and approvals, and agents using governed data with human oversight. Prompts, actions and outputs are monitored and versioned.
The report describes continued integration and scaling through its AI teams.
It does not establish communications capture or executive decision memory. Individual workflow models, deployment denominators, accuracy and returns are undisclosed. An approval workflow is not AI approval authority.
Useful intelligence needs a route into accountable work. Keep the evidence and approval record beside the proposed action.
Diversified 2025 Sustainability Report
Diversified Energy · 2026 publication; 2025 reporting period · PDF pp. 63–64: Information Systems / Artificial Intelligence, data, workflows and human oversight
Issuer report reviewed 9 October 2026. Company architecture claims have no independent performance validation.
ENERGYai
Oil & gas operator / Reported trial and subsequent deployment narrative
ENERGYai connects proprietary energy knowledge to specialist upstream agents, rather than a disclosed communications memory.
January 2025’s release reports a completed 90-day trial using proprietary data and an LLM for seismic and reservoir work. ADNOC’s 2025 review says deployment began across the value chain.
Broader enterprise and boardroom use is the stated direction; the initial five-agent rollout was a plan in the January release.
Email, meeting capture, decision memory and executive authority are not established. Trial accuracy claims lack a disclosed benchmark. Early rollout plans should not be read as a measured company-wide result.
Technical agents can provide useful inputs to leadership. They still need a way to connect those inputs to decisions and outcomes.
ADNOC and AIQ complete ENERGYai trial
ADNOC / AIQ · 16 January 2025 · Completed trial, natural-language engineer interaction and planned operational version
Issuer announcement reviewed; trial findings are separate from rollout targets.
ADNOC 2025 — A year in review
ADNOC · 2025 reporting period; checked 9 October 2026 · ENERGYai panel: deployment started across the value chain during 2025
Operator year-in-review reviewed; deployment narrative does not establish complete communications coverage.
AI for Energy
ADNOC · Undated; checked 9 October 2026 · ENERGYai and AI strategy sections
Operator strategy page reviewed; broad ambitions are not deployment denominators.
Glean / Wrap-Up Agent / governed Salesforce actions
Named customer outside energy / Vendor-reported customer use; story undated
Storable’s case is close to the daily briefing and follow-up part of your idea.
Glean describes a Wrap-Up Agent reading Gmail and Slack to summarize developments and open to-dos. A separate agent writes to Salesforce through role-controlled actions; search spans company tools and work ownership.
The vendor describes further agent use for product development and operations.
The story is undated; an in-window rollout date is unverified. It does not prove CEO usage, complete communications capture or automatic task closure. Promotional time-saving figures lack a disclosed method.
Start with unresolved commitments in a daily brief. Grant write access separately from permission to read and summarize.
How Storable uses Glean
Glean customer story / Storable · Undated; checked 9 October 2026 · Search as a single pane of glass; Wrap-Up Agent; Governed Salesforce Agent
Vendor-reported customer deployment reviewed; no dated rollout or independently validated outcome.
AI-supported chiefs-of-staff workspace
Internal practice outside energy / Promoted internal practice; demonstration listed
Notion describes its own chiefs of staff using AI in the executive operating cadence.
A dated webinar listing names People and GTM chiefs of staff and promotes a demonstration of AI-assisted communication plans, weekly summaries and coordination of priorities.
The listing describes practice and a planned demonstration, rather than a technical roadmap.
The reviewed source is promotional webinar copy, not a transcript or deployment audit. Connectors, persistent decision memory, action permissions, usage and measured outcomes are unspecified.
Design around the human chief of staff’s weekly work: decisions, owners, exceptions and follow-up. Test whether AI makes that record more useful.
How Notion’s Chiefs of Staff use Notion
Notion · 27 August 2025 · Webinar description, AI executive-rhythm bullets and named speakers
Public webinar listing reviewed; claimed practice is not independently measured.
Dash / Agentic Work Management
Software product offer / Phased commercial rollout; customer activation unverified
Asana’s product description directly targets the AI-chief-of-staff idea. It is a product offer, not an operator deployment case.
June materials advertise meeting, Slack and email follow-ups linked to structured work. September management remarks describe mid-September AWM packaging for new logos, self-service and renewals. The current Dash page says users retain approvals and final decisions.
Rollout was phased. Specific customer activation, enabled connectors and coverage still need verification.
No named Dash deployment or validated Dash-specific outcome was established in this review. FedEx and COS results in the launch refer to other Asana products. An advertised chief of staff has no demonstrated CEO authority.
The design connects conversation to work. Evaluate actual feature access, source coverage and completed actions before treating it as a company brain.
Asana unveils operating system for human-agent teams
Asana · 4 June 2026 · Dash description and Availability; named customer examples concern other products
Issuer announcement reviewed; product capabilities and customer results kept separate.
Asana Q2 FY2027 earnings call
Asana investor relations / earnings transcript · 3 September 2026 · PDF pp. 5–6 and 11: Dash, mid-September packaging and AI teammate routing
Management describes phased commercialization, not a named customer outcome.
Asana Dash product description
Asana · Undated; checked 9 October 2026 · Decision mapping, morning brief, teammate routing, approvals and final decisions; FAQ
Product claims reviewed. Users retain approvals; advertised capabilities do not establish a named customer deployment.
SANAD drilling knowledge assistant
Oil & gas operator / Pilot and validation; later field test reported
SANAD gives drilling engineers a way to find guidance in selected procedures and well records, with people checking the answers.
The pilot used 14 North Kuwait procedures, safety material and well records. KOC says its standards team checked every pilot output. OpenWELL context was tested; wider integration is still planned. The engineer retains the decision. SLB supported model training and data processing. A later attributed interview describes one Lower Burgan field test.
CAPEX, stage-gate, production support and performance modules are roadmap items. Pilot OpenWELL context does not establish full company-wide integration.
No active-user count, output error rate, full cost or matched baseline is disclosed. Potential benefits across 47 wells are an extrapolation, not an achieved rollout. CAPEX, stage-gate and production modules remain roadmap items. This is separate from G Agent scheduling in AI-085. The reported 13-day and KD 97,000 figures do not isolate AI. No comprehensive communications capture, decision-outcome memory or autonomous executive is established.
A narrow knowledge system can be useful without becoming a company brain. Start with the procedures engineers actually need and test whether the answers remain correct and traceable after review.
SANAD AI: The Generative AI Drilling Assistant
Kuwait Oil Company · 2026-Q2 · PDF pages 28–31 / printed 26–29; 14 SOPs on PDF 30; validation, ownership and roadmap on PDF 31
Editor read the original 52-page PDF and rendered pages 30–31. Issue period is April–June 2026; exact posting date unknown.
KOC’s Sanad uses AI to unlock drilling knowledge
KUNA interview, republished by Kuwait Times · 2026-09-29 · Timestamp; lower-Burgan field test, SOP-related savings, RAG description and human responsibility
Editor read the attributed interview. No original KUNA page located in the bounded search. Savings are engineer-reported, not independently validated.
X.brain enterprise AI engine
Oil & gas operator / Engine launch described; eight-agent rollout unverified
PTTEP says it launched X.brain, an internal AI engine grounded in its own operating knowledge and data.
The operator’s November 2025 feature describes a launched internal engine grounded in company knowledge and data, for analysis and recommendations. Active use and the actual agent population are not disclosed.
Eight specialist agents across subsurface, drilling, production, supply chain and corporate work; all-eight rollout is unverified.
No proof of comprehensive communications capture, decision-outcome memory or autonomous executive authority. DORA combines machine learning and physics-based simulation; it is not shown to run on X.brain. No asset scope, test population or full costs support the DORA/LogSight comparisons. Separate DORA/LogSight metrics are not X.brain results. No active-user population, evaluation method or full-cost return is given.
Ask to see the agents people use and the approvals behind their recommendations. An internal engine can support better decisions without having executive authority.
PTTEP’s AI Vision for a Smarter, Sustainable Future of Energy
PTTEP · 2025-11-03 · Page date; X.brain introduction; eight-agent plan; DORA and LogSight examples
Full original operator page read by editor; Thai date 3 November 2568 converts to 2025-11-03. Company claims, not independent validation.
ChatDVN internal company-data assistant
Oil & gas operator / Operating internal workflow; issuer-described daily use
Devon’s CEO says its firewalled ChatDVN tool was a standard part of daily work by May 2026, after years spent organizing internal data.
The issuer-hosted May 2026 call confirms a firewalled internal tool in daily work, after company data cleanup. Coding and operations-data examples are described alongside it, without a claim that every example runs inside ChatDVN.
Management describes a next phase of AI-centered process redesign without specifying completed processes or agent deployment counts.
Daily use supports an operating workflow, not verified enterprise-wide active adoption. No comprehensive communications capture, decision-outcome memory, execution permissions or AI executive authority is established. Training and productivity claims from the earlier trade interview retain their separate, unverified status. No tool-level cash benefit, full cost or task-quality comparison is supplied. Autonomous lift operation is a separate workflow, not ChatDVN execution authority.
Daily use is a better starting point than a launch announcement. Ask which jobs are completed and how much review is still needed.
Devon Q1 2026 earnings call — internal ChatDVN workflow
Devon Energy / FactSet CallStreet · 2026-05-06 · PDF and printed p. 4; Clay Gaspar prepared remarks, three waves of AI
Issuer-hosted corrected 20-page transcript retrieved HTTP200. Editor read relevant p.4 and visually inspected it; management claims, not independent performance validation.
Our proposal is a supervised executive brief that remembers the decision, the evidence behind it, the owner and the next check. It would bring together the company’s existing records rather than depend on moving every conversation into one application.
Start with one asset’s weekly operating review. Replay previous reviews using only the information available then; compare the brief with the actual record before a supervised live trial.
Check material omissions, unsupported claims, correct owner/action links, missed commitments, source freshness and reviewer effort. Measure completed follow-through, rather than the number of answers produced.
The asset owner and executive team approve decisions. Tool permissions and operating authority stay explicit. Stop if the brief obscures uncertainty, exposes restricted information or creates more coordination work.
Reviewed 9 October 2026 broad review; selected additions and source checks through 21 October. Three Luna researchers examined energy deployments, outside-industry analogs and evidence boundaries. Sources include issuer reports, direct executive interviews, vendor customer stories and a webinar listing. Dated disclosures fall within 9 October 2024–9 October 2026; undated pages are labeled and do not establish an in-window rollout. Customer or issuer descriptions, advertised product capabilities, plans and unestablished capabilities remain separate. No independently verified net return or complete autonomous executive was established. Targeted addition on 18 October: KOC SANAD procedures and engineer validation; earlier cases retain their source-review dates. Targeted addition on 21 October: PTTEP X.brain launch description and unverified specialist-agent scope; operating-tool examples remain separate from engine results. Additional 21 October review: Devon daily-use evidence from an issuer-hosted transcript; earlier interview metrics remain separate. Targeted evening review on 21 October: Quantum Lakebase data foundation; categorical capability assessments remain unchanged.