--- title: "ChatGPT vs Copilot Memory in Microsoft 365." description: "Compare ChatGPT vs Copilot memory on one Microsoft 365 renewal brief. Test source authority, permissions, citations, corrections, export, and handoff." canonical: "https://scalewithsearch.com/articles/chatgpt-vs-copilot-memory-microsoft-365" date: "2026-08-24" modified: "2026-10-09" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # ChatGPT vs Microsoft Copilot Memory: Which One Should Prepare the Next Microsoft 365 Job. A manager asks for a client renewal brief. Microsoft 365 Copilot can see an approved Outlook email, a Teams decision, and two Word files. ChatGPT remembers how the manager likes briefs formatted. One Word file contains an obsolete renewal term. Neither product label proves which source will control the answer. The buyer does not need a broad feature contest. The buyer needs one renewal job tested against the same known facts, conflict, permissions, correction, and stopping point. Copilot's access to Microsoft 365 work data and ChatGPT's memory features solve different parts of that job. Source authority still belongs in a record the business can inspect. ## Copilot memory and work grounding are different inputs Microsoft says Microsoft 365 Copilot can personalize chat with saved memories, inferences from chat history, and custom instructions. Temporary chat does not access or add personalized information. [Microsoft: Personalize what Microsoft 365 Copilot remembers](https://support.microsoft.com/en-US/Microsoft-365-Copilot/personalize-what-microsoft-365-copilot-remembers) The guide to [make Microsoft 365 Copilot remember your work](/articles/how-to-make-copilot-remember) sets those inputs in order and tests one saved preference. That personal layer is separate from work grounding. Microsoft says Copilot Chat can draw from work data such as emails, Teams messages, and files. Users can control available sources and inspect inline citations and a source list. [Microsoft: Control and review Copilot Chat sources](https://support.microsoft.com/en-us/Microsoft-365-Copilot/control-review-sources-copilot-chat) Microsoft's Copilot APIs use Entra ID and trim results to content the signed-in user can access. Sensitivity labels and existing permissions remain relevant to returned content. [Microsoft: Copilot API security and authentication](https://learn.microsoft.com/en-us/microsoft-365/copilot/extensibility/copilot-apis-security-authentication) Access answers "what may this user retrieve?" Memory answers "what preference or prior context may shape the response?" Neither answers "which renewal term is approved?" ## ChatGPT brings another context arrangement ChatGPT can organize ongoing work in Projects with project instructions, files, chats, and project memory. OpenAI says project instructions override global custom instructions. Project-only memory can keep context inside the project, with behavior that varies by plan and workspace type. [OpenAI: Projects in ChatGPT](https://help.openai.com/en/articles/10169521-projects-in-chatgpt) ChatGPT may also use connected apps when the workspace enables them. Connection does not turn every accessible record into approved truth. The project still needs a source rule, and the test still needs citations that a reviewer can open. Use the same account class intended for production. A personal subscription, business workspace, and enterprise workspace can have different controls, retention, sharing, and connectors. The correct comparison is not "which chatbot remembers me better?" It is "which configured workspace prepares this renewal brief with the right sources, evidence, and stop?" ## Read the January 2026 feature snapshot A January 2026 comparison recorded what each product remembered. Treat it as dated context for the trial below, and recheck the provider pages before you rely on a row. ChatGPT kept two layers. Saved memories were short statements about the user, such as "product manager who prefers concise summaries" or "company uses Python for backend work." Chat-history reference, available since April 2025, let ChatGPT draw on past conversations searchable back about a year. In January 2026, OpenAI upgraded that layer for Plus and Pro users: a past chat used in an answer appeared as a source link. ChatGPT held preferences, work context, writing style, past conversations, and recurring topics. It did not hold files outside the chat where you uploaded them, or work documents you did not paste in. It also did not hold complex project state across sessions or anything from Temporary Chat. Users could view, edit, and delete memories, turn memory off, or use Temporary Chat. Copilot also kept two layers. Copilot Memory picked up details from chats, such as "I prefer Python for data science" or "I'm working on Project Alpha." Custom instructions set standing rules, such as "keep my emails concise." A "memory updated" notice appeared when Copilot saved something, and users could view, edit, delete, or turn off memory. The second layer, which Microsoft calls Work IQ, drew on the user's role, company structure, project histories, and work patterns. It covered Word, Teams, Outlook, Excel, and SharePoint. With it, Copilot could read emails, documents, meetings and transcripts, SharePoint files, and the organization chart without an upload. A request such as "summarize the feedback from last week's customer calls" searched Teams transcripts and email. A question about last quarter's sales found the relevant Excel file. Copilot did not reach files outside Microsoft 365, conversations from before memory was on, or a running narrative of project state across months. The data stayed inside Microsoft's infrastructure. | Feature | ChatGPT | Microsoft 365 Copilot | |---|---|---| | Conversation memory | Saved memories and chat-history reference | Copilot Memory and custom instructions | | Work files | Uploaded per chat or Project | Automatic access to Microsoft 365 documents, email, and meetings | | Integration | Standalone, with optional connected apps | Built into Word, Outlook, Teams, Excel, and SharePoint | | Data location | OpenAI servers | Microsoft 365 infrastructure | | User control | View, edit, and delete memories; turn memory off | The same, plus control over Microsoft 365 source access | | Price position | Entry plan cost less per user | A per-user add-on on top of a Microsoft 365 license | ChatGPT's strength was range. It works for writing, code, research, and brainstorms with any data you bring. It is not tied to one suite, and its memory does not depend on Microsoft. Copilot's strength was depth inside Microsoft 365. It reads the work context without an upload, sits inside the apps where the work happens, and knows team structure and relationships without an explanation. Both shared one gap. They remembered facts about the work but kept no maintained record of project state. ChatGPT could recall a conversation from three months earlier without a current list of decisions and next steps. Copilot could open the files without loading a project-level context at the start of each session. The guide on [why Copilot file access is not memory](/articles/why-does-copilot-forget) routes each kind of context to its layer. That record belongs in a maintained file. One example is a `CLAUDE.md` that Claude Code reads at session start; another is the source packet in the test below. You version it with Git and update it as the work changes. ## Define one renewal-brief job Choose a synthetic or de-identified account. Prepare a fixed source packet: - one approved Outlook email that states the current renewal date; - one Teams decision that names the approved discount; - one current Word policy with an effective date; - one superseded Word policy with a conflicting term; - one external market source that must be cited; - one manager preference about layout; - one prohibited record the test user cannot access. Write the expected output before running either product. Require the renewal date, approved discount, policy citation, open question, and draft-only state. This turns a vague memory comparison into a scored business job. ## Inspect `m365-memory-job-test.csv` ```csv source,permission,status,required_citation,prohibited,output_field,retention_check renewal-approved.eml,allowed,current,yes,no,renewal_date,no teams-decision.txt,allowed,current,yes,no,discount,no policy-2026.docx,allowed,current,yes,no,cancellation_term,no policy-2024.docx,allowed,superseded,yes,no,none,no executive-private.docx,denied,current,no,yes,none,no manager-layout.md,allowed,preference,no,no,format,yes ``` Add a known-answer key outside both tools. Record the account, license, workspace setting, connectors, source selections, prompt, output, citations, and test date. The [small-business memory comparison](/articles/best-ai-small-business-memory) explains why office-stack fit matters. This fixture supplies the evidence for one office rather than assuming the suite with the most integrations wins. ## Run the same job in both products Start clean. Do not let one product see the other product's output. In Copilot, select the intended work sources. In ChatGPT, configure the intended Project and approved connection or upload route. Give both the same task brief and output requirements. Then run these tests: 1. Produce the renewal brief from the current source packet. 2. Open every citation and verify the claimed line. 3. Ask which source controlled each output field. 4. Confirm the superseded policy did not supply a current term. 5. Confirm the prohibited record was not retrieved or inferred. 6. Correct the discount in the authoritative source. 7. Start a fresh session and run the job again. 8. Remove one required source and confirm the tool stops or flags the gap. Score verified fields, incorrect fields, unsupported fields, accessible citations, prohibited-source exposure, and reviewer time. Fluency is not a substitute for a correct renewal term. ## Test deletion, export, and replacement Memory has an operating lifecycle. Before the trial, inspect the supported memory-removal, project-deletion, export, and retention controls in each intended account. Record what the current plan and administrator permit; do not infer identical controls across the two products. Test the exact account and plan. Remove the synthetic manager preference. Start a new session. Confirm it no longer appears. Delete the test project or memory according to policy. Record what the administrator and user can still access. Then perform a replacement test. Give the task brief, source packet, and known-answer key to a fresh operator. The job should not depend on the first manager's private chat history. The [memory ownership test](/articles/do-you-own-ai-memory) is practical here: can the company reconstruct the job, sources, correction path, and acceptance test outside the vendor interface? ## Include administrator and reviewer effort The employee running the prompt is not the only operator. Microsoft 365 permissions, sensitivity labels, licenses, connectors, and source controls may require an administrator. ChatGPT workspace apps, project sharing, retention, and account roles may require another administrator. Record setup time, permission-remediation time, and citation-review time. A tool can finish the visible prompt faster while creating more work for the people who must govern it. Also test an ordinary staff account. An administrator's successful retrieval does not prove the intended user has the same access. Run the prohibited-record case with both accounts and save the permission result. The operating choice should state who maintains the source packet, who resolves conflicts, who audits access, and who reruns the fixture after a workspace change. Without those owners, the comparison expires as soon as the office configuration changes. ## Choose by the operating job Copilot is the stronger candidate when the job depends heavily on permission-trimmed Microsoft 365 work data. Reviewers must also be able to inspect its cited sources in their existing environment. ChatGPT is the stronger candidate when the team already has a governed ChatGPT workspace and a controlled Project. The Project also needs reliable access to the required source packet through files or approved apps. Either product fails the job when it selects an obsolete term, hides source provenance, or relies on one person's memory. It also fails when nobody can reconstruct the job after a staff or vendor change. The choice matters less in two cases. If AI does no business-critical work for you, pick the interface you prefer. If a company contract already standardizes on one platform, improve memory inside that platform before you switch. The [ChatGPT and Claude memory comparison](/articles/chatgpt-vs-claude-memory-business) applies the same principle across another vendor pair: test continuity on a named output, not on marketing language. ## Approval and stopping boundary This comparison may use synthetic records, create test projects, select approved sources, produce drafts, and delete test objects. It may inspect permissions and save local scorecards. It stops before it uploads real employee or client data, enables organization-wide connectors, or changes retention settings. It also stops before it grants access to restricted files or sends a renewal message. Data owners, Microsoft 365 administrators, ChatGPT workspace administrators, and the account owner approve those actions. If the current renewal term has no authoritative record, neither tool should resolve the conflict. Stop and ask the business owner to approve the source. The same owned-record method runs each client's content library with SEO and signal desk: [How the build works: Owned files, checks, handoff.](/how-it-works) ## Sources - [Microsoft: Personalize what Microsoft 365 Copilot remembers](https://support.microsoft.com/en-US/Microsoft-365-Copilot/personalize-what-microsoft-365-copilot-remembers) - [Microsoft: Control and review Copilot Chat sources](https://support.microsoft.com/en-us/Microsoft-365-Copilot/control-review-sources-copilot-chat) - [Microsoft: Copilot API security and authentication](https://learn.microsoft.com/en-us/microsoft-365/copilot/extensibility/copilot-apis-security-authentication) - [OpenAI: Projects in ChatGPT](https://help.openai.com/en/articles/10169521-projects-in-chatgpt) - [Microsoft: Work IQ overview](https://learn.microsoft.com/en-us/microsoft-365/copilot/extensibility/work-iq/) ## Questions about ChatGPT vs Microsoft Copilot Memory: Which One Should Prepare the Next Microsoft 365 Job? ### Should I switch from ChatGPT to Microsoft 365 Copilot for business work? Choose from one defined Microsoft 365 job, not a general preference. Run the same renewal brief in both products and score source authority, permission handling, citations, correction behavior, reviewer effort, deletion, export, and replacement. ### What is the difference between Copilot memory and Microsoft 365 work grounding? Memory is one context route, while work grounding retrieves permitted organizational content such as files, mail, meetings, and other Microsoft 365 records. Test them separately so a recalled preference is not confused with a current, permissioned business source. ### Why can Microsoft 365 Copilot forget information from the same chat? A product's visible integration does not prove complete conversational recall or source use. Reproduce the bounded job, record what sources were available, require citations, and preserve the accepted brief and corrections outside the chat so the work can continue or move tools. ## Related: Owned Memory - [Test what Microsoft Copilot remembers and keep the rest in a file you own](/articles/copilot-memory-problems) - [Make Microsoft Copilot remember your work with its memory settings and a context file](/articles/how-to-make-copilot-remember) - [Route Microsoft 365 and GitHub Copilot context to the right memory layer](/articles/why-does-copilot-forget) ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```