What leaders and teams rolling out ChatGPT report, with every finding tied to its evidence.
Read the findingsPractitioner sentiment, leaders and teams
+14 Mixed from 38 sources that take a side
How this figure is madeFrom 36 authors in 37 discussions. Run a pilot on your team's recurring tasks to measure how often answers are used without rework.
Jan 2025 to Oct 2026.
86 sources from practitioners, by month and sentiment.
Team and organization accounts lean positive. Campus-wide rollouts paired with training lead the positive accounts, and agents and connectors draw more criticism than praise. The admin complaints are lockouts: a workspace deactivated for five days, and another locked when its only admin left. On cost, teams report surprise bills past included credit. 12 of 41 accounts keep a second assistant.
Positive, mixed and negative sources
The score counts them as 15 positive, 11 mixed and 10 negative votes, one per author. See How the numbers work.
See the evidenceBelow: 1 praise theme, 3 concerns, 1 point of agreement and 2 disagreements.
Counted for Leaders and teams rolling it out, from 41 sources. Change group
Every capability practitioners describe positively, ranked by how many sources say so. Open one to read who said it.
Counted for Leaders and teams rolling it out.
Positive, 3
Positive
cloudking, Hacker News commenter
Comment on: Google is making AI in Gmail and Docs free, but raising the price of WorkspaceHacker News comment
Their company uses ChatGPT Team every day across product and engineering, valuing project folders, custom GPTs and pulling files in from Google Drive.
Positive
BBVA, quoting Guillermo Vieira, Head of Training and Engagement on the AI Global Adoption team
BBVA is now using ChatGPT to streamline legal queries and marketing processesCompany news article, BBVA
BBVA's nine-lawyer advisory team answers more than 40,000 legal queries a year with a custom GPT, and a person reviews every answer before it goes out.
Positive
BBVA, quoting Elena Alfaro, Global Head of AI Adoption
BBVA expands its agreement with OpenAI to 11,000 ChatGPT licences for the bank's employeesCompany news article, BBVA
BBVA grew from 3,300 to 11,000 ChatGPT Enterprise licenses; 83% of licensed users use it daily and those automating tasks report saving 2.8 hours a week.
Mixed or negative, 2
Negative
Peter, ChatGPT Business workspace owner
Migrated GPT is private in ChatGPT Business, and plugin sharing is disabled for the workspace ownerOpenAI Community forum post
The workspace owner migrated an HR policy GPT employees rely on, but the resulting plugin is private and sharing is disabled, taking the tool out of use.
Negative
Sebastian Fruchtzweig, AI consultancy owner
Frustration with OpenAI Workspace Changes - AI Professionals Penalized for Building Custom GPTsOpenAI Community forum post
After paying for a Team workspace to host client and teaching GPTs, the admin can no longer create custom GPTs there and must build them in a personal account.
Every concern practitioners raise, most raised first.
Counted for Leaders and teams rolling it out.
Raised as a concern, 7
Negative
Peter, ChatGPT Business workspace owner
Migrated GPT is private in ChatGPT Business, and plugin sharing is disabled for the workspace ownerOpenAI Community forum post
The workspace owner migrated an HR policy GPT employees rely on, but the resulting plugin is private and sharing is disabled, taking the tool out of use.
Negative
walnutcoding, workspace member posting for the company
ChatGPT Business Workspace Deactivated for 5+ Days - ~600 Users AffectedOpenAI Community forum post
The company's Business workspace was deactivated without warning for five days, locking out about 600 members, with billing pages unreachable before renewal.
Negative
Joseanperez123, replacement workspace administrator
Lost access to ChatGPT Teams workspace after an unpaid extra seat, need help restoring admin rightsOpenAI Community forum post
When the only admin left, adding a seat to take over locked the company's Team workspace over the unpaid seat, cutting off every member.
Negative
RootAccess, Business workspace admin
All 4 of my paid Business Standard seats missed the Oct 2 global reset, while Pro 500 was fixedOpenAI Community forum post
None of the organization's three Business Standard seats received the October 2 usage reset, and support looped them through checklists for over 40 hours.
Negative
mrngld, Hacker News commenter
Comment on: Understanding ChatGPT WorkHacker News comment
Avoids sharing personal matters with ChatGPT, and worries that using the employer's corporate seat lets HR read every chat through compliance tools.
Mixed
submeta, Hacker News commenter
Comment on: AI subscriptions are a ticking time bomb for enterpriseHacker News comment
Two months after the company rolled out ChatGPT, IT operations told the commenter their usage had produced a bill of at least a thousand euros beyond included credit.
Negative
Sebastian Fruchtzweig, AI consultancy owner
Frustration with OpenAI Workspace Changes - AI Professionals Penalized for Building Custom GPTsOpenAI Community forum post
After paying for a Team workspace to host client and teaching GPTs, the admin can no longer create custom GPTs there and must build them in a personal account.
Discussed without concern, 3
Mixed
mentorcliq, Hacker News commenter
Comment on: Ask HN: AI rollout for 100 employees, mostly non-technical - what worked?Hacker News comment
Their organization rolled out ChatGPT Enterprise and Team plans over about two years and is now moving toward Claude, keeping admin oversight of usage logs.
Positive
Dr. Leslie Kennedy, Assistant Vice Chancellor for Academic Technology Services, California State University
What Higher Ed Leaders Can Learn from CSU's ChatGPT Edu Rollout (Changing Higher Ed podcast 315)Podcast episode, Changing Higher Ed podcast
CSU set governance first, then deployed ChatGPT Edu to about 225,000 accounts; 97% of faculty say outputs need verification, and communication was the hardest part.
Positive
Andrew T. Guzman, Provost and Senior Vice President for Academic Affairs, University of Southern California
Faculty Guidance: Preparing for the ChatGPT Edu Rollout in Spring 2026Provost's letter, USC Office of the Provost
USC opened ChatGPT Edu campus-wide and left AI rules in each course to faculty; prompts stay out of model training, and faculty cannot see student chats.
GILD's fact checks, which don't count as sources
Checked Oct 12, 2026.SCIM, role-based access controls, Enterprise Key Management, IP allowlisting, and the Compliance API logs platform are listed for Enterprise only. Business includes SAML SSO, MFA, and an admin console. ChatGPT pricing
Open them in the recordCounted for Leaders and teams rolling it out.
Raised as a concern, 6
Negative
RootAccess, Business workspace admin
All 4 of my paid Business Standard seats missed the Oct 2 global reset, while Pro 500 was fixedOpenAI Community forum post
None of the organization's three Business Standard seats received the October 2 usage reset, and support looped them through checklists for over 40 hours.
Negative
galleywest200, Hacker News commenter
Comment on: Companies Are Throttling Employees' AI Use Because It's Too ExpensiveHacker News comment
After meetings urging staff to use AI, the company discontinued ChatGPT Enterprise accounts for anyone not using it often enough, to save money.
Negative
z2, Hacker News commenter
Comment on: Outsourcing plus local AI will soon become more economical vs. frontier labsHacker News comment
The commenter's company saw ChatGPT credit usage double with stable use, and found Claude Enterprise users exhausted their limits within hours.
Mixed
submeta, Hacker News commenter
Comment on: AI subscriptions are a ticking time bomb for enterpriseHacker News comment
Two months after the company rolled out ChatGPT, IT operations told the commenter their usage had produced a bill of at least a thousand euros beyond included credit.
Mixed
toephu2, Hacker News commenter
Comment on: Boom, bubble, bust, boom. Why should AI be different?Hacker News comment
Gets rate-limited every day even though the company holds enterprise subscriptions to Gemini, ChatGPT and others.
Negative
dpiers, Hacker News commenter
Comment on: Y Combinator says Google is a monopolist, no comment about its OpenAI tiesHacker News comment
The commenter's company is abandoning ChatGPT Enterprise because Google now bundles Gemini into Google Workspace at no extra cost.
Discussed without concern, 3
Positive
Duke University Office of Information Technology
Expanded ChatGPT Edu now available at Duke at no costUniversity IT news post, Duke OIT
Duke moved ChatGPT Edu from consumption and credit-based billing to a no-cost combined plan for eligible users and forgave month-to-date Premium charges.
Negative
Sebastian Fruchtzweig, AI consultancy owner
Frustration with OpenAI Workspace Changes - AI Professionals Penalized for Building Custom GPTsOpenAI Community forum post
After paying for a Team workspace to host client and teaching GPTs, the admin can no longer create custom GPTs there and must build them in a personal account.
Mixed
IMTDb, Hacker News commenter promoting AI use at their company
Comment on: OpenAI O3-MiniHacker News comment
Promoting AI across a 150-person company with ChatGPT Plus for all employees, the commenter struggles to explain which of several models staff should pick.
GILD's fact checks, which don't count as sources
Checked Oct 12, 2026.Business offers credit usage limits by seat type and per user. They are unset by default. Automatic reload has no monthly cap unless one is entered. A report of an unexpected bill does not establish that limits were unavailable. Managing credits and spend controls in ChatGPT Business
Open them in the recordCounted for Leaders and teams rolling it out.
Raised as a concern, 3
Negative
mrngld, Hacker News commenter
Comment on: Understanding ChatGPT WorkHacker News comment
Avoids sharing personal matters with ChatGPT, and worries that using the employer's corporate seat lets HR read every chat through compliance tools.
Negative
Lionel Vest, ChatGPT user at an organization
Request for native WebDAV support in ChatGPT Knowledge BaseOpenAI Community forum post
Company knowledge only reads uploads or Google and Microsoft connectors, which forces data duplication and raises sovereignty and GDPR concerns.
Mixed
CJ Loosbrock, senior director and chief technology officer, University of Minnesota Office of Information Technology
Gemini: A powerful AI partnerUniversity news feature, Brief (University of Minnesota)
With more than 13,000 people on free ChatGPT accounts in March 2025, Minnesota's CTO chose Gemini, already part of its Google license, as the more secure campus option at no added cost.
Discussed without concern, 5
Positive
Dr. Leslie Kennedy, Assistant Vice Chancellor for Academic Technology Services, California State University
What Higher Ed Leaders Can Learn from CSU's ChatGPT Edu Rollout (Changing Higher Ed podcast 315)Podcast episode, Changing Higher Ed podcast
CSU set governance first, then deployed ChatGPT Edu to about 225,000 accounts; 97% of faculty say outputs need verification, and communication was the hardest part.
Positive
Andrew T. Guzman, Provost and Senior Vice President for Academic Affairs, University of Southern California
Faculty Guidance: Preparing for the ChatGPT Edu Rollout in Spring 2026Provost's letter, USC Office of the Provost
USC opened ChatGPT Edu campus-wide and left AI rules in each course to faculty; prompts stay out of model training, and faculty cannot see student chats.
Positive
Manish Parashar, Chief AI Officer, University of Utah
U launches ChatGPT Edu, a university-centered generative AI tool for campus useUniversity news post, @theU
Utah's Chief AI Officer calls ChatGPT Edu a major leap for responsible AI on campus. The launch keeps university data out of training and adds OpenAI training for more than 100 campus session leaders.
Positive
rawgabbit, Hacker News commenter
Comment on: Microsoft lets bosses spot teams that are dodging CopilotHacker News comment
Because staff were already working in personal ChatGPT accounts, the company bought ChatGPT Enterprise for compliance; Copilot feels far more limited.
Positive
Edward Aractingi, Chief Information Officer, William & Mary
William & Mary launches ChatGPT Edu pilotUniversity news article, William & Mary
William & Mary CIO Edward Aractingi describes ChatGPT as offering a strong blend of features and the enterprise license as making advanced AI available through a secure, centrally managed platform. The pilot targets selected faculty and staff and evaluates responsible use in teaching, research and operations.
GILD's fact checks, which don't count as sources
Checked Oct 12, 2026.Business, Enterprise, and Edu content is not used for training by default. Personal plans train on new conversations unless Improve the model for everyone is turned off. Business admins can view, export, and delete members' conversations. Enterprise privacy at OpenAI
Open them in the recordStatements this group's sources agree on, then the ones close to agreement.
People and organizations using ChatGPT also pay for or use Claude, Gemini, or Copilot, and pick per task or per team.
12 supporting sourcesClose to agreement: each needs 9 supporting sources
Training, usage guidelines, champions, and shared examples decide how much of a licensed workforce actually uses ChatGPT.
7 supporting sourcesQuestions the sources split on, with the count on each side.
Evidence remains mixed
Give everyone access, 8 sources
Some license the whole organization and build adoption with training and shared GPTs.
Start targeted, 4 sources
Others run a measured pilot first, or remove seats from people who do not use them.
Evidence remains mixed
ChatGPT stays primary, 5 sources
Some prefer ChatGPT over Copilot, Gemini, or Claude as the everyday assistant for their team or their own work.
Another assistant leads, 4 sources
Others moved their main use to Claude or Gemini, often for quality, limits, or because a suite they already pay for bundles it.
The three groups, then every practitioner by the context they work in, and how emphasis changes with scale.
Three groups drawn from the same sources. Choose a group to select whose experience counts. Within that group, each author counts once in the practitioner sentiment, with their latest view. Signals open on Leaders and teams rolling it out.
41 sources. Team and organization accounts focus on rollout decisions and adoption, including admin controls, spending, and data handling.
45 sources. Individual users report on ChatGPT research and writing, along with memory and usage limits.
86 sources. These accounts mix individual ChatGPT use with team and company rollouts, whose requirements and conclusions differ.
86 sources from practitioners, by the context they work in. Filled rows count for this group.
Emphasizes Research, writing, memory, and switching assistants
Emphasizes Workspace admin friction, seat usage, and sharing GPTs
Emphasizes Connectors, data handling, and which assistant teams prefer
Emphasizes Pilots, enablement, and governance before scale
Emphasizes Adoption, training, credits, and cost at scale
What changes with scale. Individuals focus on research and writing, along with memory. Teams focus on workspace administration and cost, including credits, seats, and training. More team accounts report a second assistant alongside ChatGPT.
Where this group's sources were published.
Counted for Leaders and teams rolling it out, from 41 sources. Change group
57 sources for this group, 16 of them documentation
Older evidence is retained when it still applies. Sources that depend on a retired model, a renamed plan, or a feature that has changed stay in the history, marked Needs a recheck.
See sources by originWho each source is, how it knows, whether it still applies and what it counts toward.
Newest first, on its own page. Every count above opens its sources there.
Read all 102 sourcesAbout this read
A running review of what people using ChatGPT at work report, from individual users to company-wide rollouts. Each finding links to its sources, and each conflict shows both sides.
Counts cover only the sources GILD reviewed.
Choose a group to select whose experience counts. Within that group, each author counts once in the practitioner sentiment, with their latest view. GILD files each source in a group from how the account describes its author. A role or employer marked self-described is what the source states about themselves. Documentation checks facts for every group and never counts toward sentiment.
A source's date and whether it still applies are separate: a recent post can describe an old version, and an old one can still hold.
Flagged for a recheck, of the sources that take a side: leaders and teams 2 of 38, individual users 3 of 30, all users 5 of 68.
Each row opens its sources for the group you are reading.
| Finding | Leaders and teams rolling it out | Individual users | All users | Sources |
|---|---|---|---|---|
| Current readCurrent sentiment | 17 positive, 11 mixed, 10 negative | 14 positive, 7 mixed, 9 negative | 31 positive, 18 mixed, 19 negative | Sources |
| PraiseWriting & Drafting | 2 of 2 positive (under the minimum) | 7 of 7 positive | 9 of 9 positive | Sources |
| PraiseResearch & Search | 1 of 1 positive (under the minimum) | 7 of 7 positive | 8 of 8 positive | Sources |
| PraiseFiles & Data Analysis | 2 of 2 positive (under the minimum) | 0 of 2 positive (under the minimum) | 2 of 4 positive (under the minimum) | Sources |
| PraiseCustom GPTs, Projects & Memory | 3 of 5 positive | 4 of 9 positive | 7 of 14 positive | Sources |
| PraiseAgents & Connectors | 2 of 4 positive (under the minimum) | 4 of 8 positive | 6 of 12 positive | Sources |
| ConcernAccuracy | 3 of 3 raise it (under the minimum) | 4 of 7 raise it | 7 of 10 raise it | Sources |
| ConcernUsage Limits & Cost | 6 of 9 raise it | 2 of 4 raise it (under the minimum) | 8 of 13 raise it | Sources |
| ConcernData Privacy & Retention | 3 of 8 raise it | 2 of 2 raise it (under the minimum) | 5 of 10 raise it | Sources |
| ConcernModel & Feature Changes | 2 of 3 raise it (under the minimum) | 2 of 2 raise it (under the minimum) | 4 of 5 raise it | Sources |
| ConcernWorkspace Administration | 7 of 10 raise it | 0 of 0 raise it (under the minimum) | 7 of 10 raise it | Sources |
| ConsensusTeams keep a second assistant | 12 supporting | 9 supporting | 21 supporting | Sources |
| ConsensusEnablement drives adoption | 7 supporting (below threshold) | 0 supporting (under the minimum) | 7 supporting (below threshold) | Sources |
| ConsensusA person checks the output | 3 supporting (under the minimum) | 4 supporting (under the minimum) | 7 supporting (below threshold) | Sources |
| ConsensusIt saves time on routine work | 2 supporting (under the minimum) | 3 supporting (under the minimum) | 5 supporting (below threshold) | Sources |
| DisagreementShould an organization give everyone ChatGPT, or start with a targeted group? | Give everyone access 8, Start targeted 4 | Give everyone access 0, Start targeted 0 (under the minimum) | Give everyone access 8, Start targeted 4 | Sources |
| DisagreementIs ChatGPT the assistant teams and users settle on? | ChatGPT stays primary 5, Another assistant leads 4 | ChatGPT stays primary 2, Another assistant leads 2 (under the minimum) | ChatGPT stays primary 7, Another assistant leads 6 | Sources |
| DisagreementDoes ChatGPT save time once checking and rework are counted? | Net time saved 3, Gains absorbed 1 (under the minimum) | Net time saved 1, Gains absorbed 1 (under the minimum) | Net time saved 4, Gains absorbed 2 | Sources |
| Current readEvidence mix | 57 sources | 61 sources | 102 sources | Sources |
GILD's checks of claims against official documentation, kept apart from the source counts.
