ChatGPT Consulting and Business Workspace Setup
Buying ChatGPT seats takes ten minutes. Setting up a workspace a business can defend takes more: the right plan, a verified domain and single sign-on, owners and admins named, the day-one defaults reviewed (apps on, workspace discovery on, no spend limits), rules for who may publish GPTs and agents, and a plan for the personal accounts staff already use. NetSys does that setup for ChatGPT Business (formerly ChatGPT Team) and ChatGPT Enterprise, then hands it over documented or keeps running it.
The short answer
ChatGPT consulting for a company is the work of choosing between ChatGPT Business and ChatGPT Enterprise, setting up the workspace so that access, data and spending are under the company's control, and getting staff onto it with rules they understand. NetSys delivers it as a fixed-scope setup: plan choice and approved use cases, workspace ownership and domain verification, single sign-on and roles, data controls, connected apps and GPT governance, a pilot group and a documented handoff to your team or to our managed AI services.
Closest related page: AI services for business. That page is the hub for everything NetSys does with AI across Copilot, ChatGPT and Claude; this page is the specific offer to set up and hand over a ChatGPT company workspace.
Owners and office managers whose staff already use personal ChatGPT accounts on company work. Firms that bought a handful of Business seats, invited people and never opened the settings. Companies large or regulated enough to need Enterprise controls (SCIM provisioning, role-based access, the Compliance Platform, data residency) and a sales conversation with OpenAI they would like an engineer beside them for.
The two plans are different products to administer. ChatGPT Business is OpenAI's self-serve plan for a shared workspace of 2 to 200 paid seats, managed inside ChatGPT's workspace settings. ChatGPT Enterprise is the sales-managed plan, administered through OpenAI's Admin Console with SCIM, role-based access, the Compliance Platform and custom retention. Much of the setup work is deciding which one your controls require, then configuring what the plan offers rather than what a brochure implied.
Plan, identity, controls, pilot, handoff
We start with the use cases and the data they touch, because that decides the plan and the settings. Then the workspace is created or taken over under a company-owned account, the domain is verified and single sign-on is set up against Microsoft Entra ID or Google Workspace, owners and admins are named, and the day-one defaults are reviewed against your policy. Connected apps, GPT and agent permissions and spend controls are set before the pilot group is invited, using the exact emails on any personal accounts staff want to merge. The pilot runs with the safe-use briefing, the settings are adjusted, and the rest of the company is invited. You get the configuration record, the policy settings and the offboarding runbook at handoff.
What ChatGPT workspace setup includes
Plan choice and approved use cases
Business or Enterprise, decided by the controls you need.
- Use cases and data types listed first: drafting, analysis, research, client documents, code
- ChatGPT Business (2 to 200 seats, self-serve, SSO and domain verification included) compared with Enterprise (SCIM, role-based access, the Compliance Platform, data residency, custom retention, a BAA by agreement)
- Standard versus Premium seats mapped to the people who will use reasoning models, ChatGPT Work and Codex heavily
- What the plan cannot do written down before purchase, so nobody expects an Enterprise control on a Business workspace
Workspace ownership, domain and single sign-on
The company owns the workspace, not whoever signed up first.
- Workspace created or transferred under a company-controlled owner account, with a second owner named
- Domain verification by DNS record, so workspace discovery and automatic account creation work for your staff and nobody else
- SAML or OIDC single sign-on against Microsoft Entra ID or Google Workspace, with MFA enforced in the identity provider, since ChatGPT cannot enforce it itself
- On Enterprise, SCIM provisioning from your directory so joiners and leavers follow the HR system
Roles, seats and member settings
Owners, admins, analytics viewers and members, named in writing.
- Owner and admin roles limited to named people, and an analytics viewer for whoever reads the reports
- Default seat type set, members invited by CSV with the right role and seat, invitation and join-request settings reviewed
- Workspace discovery and external-domain invites decided deliberately rather than left on their defaults
- A note on what Business cannot restrict (members inviting other members) and how a monthly review covers it
Data controls, retention and memory
What the vendor keeps, who can see it and for how long.
- Training-data position confirmed and recorded: Business and Enterprise workspace data is not used to train OpenAI's models by default
- Retention policy set on Enterprise; on Business, what happens to a removed member's chats, Projects and GPTs written into the offboarding runbook
- Workspace memory and shared-link settings set to match the policy
- Admin visibility explained plainly to staff: what owners can export and what they cannot see day to day
Connected apps and permissions
Apps are on by default in Business; we decide which ones stay on.
- The connected apps (SharePoint, Outlook, Teams, Google Drive, Slack and others) reviewed against the data each one can reach
- An approval policy per app: always ask, read only or low-risk actions, with write actions kept behind a prompt
- Microsoft Graph permissions granted deliberately by your Entra administrator for the Microsoft apps, with the consent recorded
- Custom MCP servers and Company Knowledge enabled only when a use case needs them, each with an owner
Custom GPT, agent and plugin governance
Who may build, who may publish and where it may be shared.
- GPT settings in the workspace: creation, sharing levels and third-party GPT access set to the policy
- Workspace agents and plugins: building and publishing limited to named people, with write actions set to always ask
- An owner recorded for every published GPT, agent or plugin, and a plan for OpenAI's scheduled retirement of custom GPTs on December 11, 2026
- Spend controls: monthly credit limits per seat type, usage alerts and auto-reload decided rather than left unset
Pilot, migration and handoff
A small group first, then everyone, with the personal accounts handled.
- A pilot group invited with the exact emails on their existing personal accounts, so history does not appear to vanish
- Personal-to-business merges explained (permanent, no undo, export first) and run with each person who wants one
- The safe-use briefing delivered with the pilot, and the settings adjusted from what the pilot shows
- A handoff pack: configuration record, policy settings, offboarding runbook and open decisions, or a transition to managed AI services
Why businesses bring NetSys in for ChatGPT setup
Tell us how many people would use ChatGPT, what they would use it for, whether staff already have personal accounts and whether you run Microsoft 365 or Google Workspace. A NetSys engineer will say which plan the controls point to and what the setup would include, then quote it after the scoping call.
Who does the work: meet the NetSys team on our About page.
- The identity and security work is the same work we do in Microsoft 365 every week: domain records, Entra ID single sign-on, admin consent, access policies and offboarding
- Vendor-neutral: we also set up Claude Team and Enterprise and Microsoft 365 Copilot, and will say when one of those fits the use case better
- Published AI rollouts: a Copilot rollout with custom agents for an 18-person organization, and a Copilot and Claude rollout for 25 employees with usage policies, connector access controls and staff training
- Settings explained against OpenAI's own defaults, with the configuration recorded so the next audit or insurer questionnaire has an answer
- Fixed scope for the setup, month to month for anything ongoing; you buy seats from OpenAI under your own account
- In business since 1998, from one office in Brooklyn, on site across the New York area and remote everywhere else
ChatGPT Business Setup in practice
- Importers and distributors
Regional food importing company
Microsoft Copilot and Claude rollout for 25 employees, AI usage policies, connector access controls, and staff training.
Microsoft Copilot ConsultingIT for Importers and Distributors
Client names are withheld. Each card is the scope of a real NetSys engagement, as delivered.
ChatGPT Business vs ChatGPT Enterprise for a company
Read from OpenAI's plan comparison and help center on October 17, 2026. Prices and features change, so we recheck them at every scoping call:
| ChatGPT Business | ChatGPT Enterprise | |
|---|---|---|
| How it is bought | Self-serve, by card, 2 to 200 paid seats | Through OpenAI sales, custom pricing, invoicing and net terms available |
| Seat prices (as of October 2026) | Standard $25 per user per month, or $20 billed annually; Premium $125, or $100 billed annually | Custom pricing on credit-based or token-based agreements |
| Administration | Workspace settings inside ChatGPT: members, billing, identity, GPTs, apps, usage | The Admin Console, with groups, custom roles and a global view across workspaces |
| Identity | SAML or OIDC single sign-on, domain verification, automatic account creation | Adds SCIM provisioning and sign-in policies per product |
| Access control | Owner, admin, analytics viewer and member roles | Role-based access control per feature, group permissions, granular GPT controls |
| Compliance tooling | SOC 2 Type 2 audit; owners can export a member's data | Compliance Platform for eDiscovery, DLP and SIEM tools, IP allowlisting, enterprise key management, ISO 27001 family certification |
| Retention | Deleted conversations removed within 30 days; a removed member's chats kept indefinitely | Custom retention policies set by the admins |
| Data residency | Listed as not included; a help article says it is rolling out gradually at checkout | Ten regions, included |
| HIPAA | No BAA offered | A BAA by agreement on eligible offerings such as ChatGPT for Healthcare and Enterprise with Regulated Workspace |
| Training on your data | Not by default | Not by default |
Comparing plans? ChatGPT Business vs Enterprise: which fits your company · ChatGPT Business vs Plus and Pro for a company.
More than 200 paid seats, invoice billing, zero data retention or a BAA all point to Enterprise. Our comparison posts go deeper on when a company outgrows Business and when a personal plan was never the right tool.
Day-one settings we review in a ChatGPT Business workspace
Each default below is documented in OpenAI's help center as of October 2026. Enterprise has more controls and a different starting position (new apps and plugins disabled until enabled), which the Enterprise setup covers:
| Setting | Official default | What we usually set |
|---|---|---|
| Workspace discovery | On: anyone with a verified-domain email can find and request to join | On with auto-accept once the domain and SSO are in place; off until then |
| Single sign-on | Off until an owner sets it up | SAML or OIDC against your identity provider, with MFA enforced there |
| Members inviting members | Allowed, and cannot be restricted on Business | Documented, with a monthly member review under managed AI services |
| Connected apps | Enabled by default | Only the apps a use case needs, with approval set to ask before write actions |
| Spend controls | No limits specified for any seat or user | Monthly credit limits per seat type, usage alerts and a decision on auto-reload |
| Workspace memory | Available; owners can turn it off for the whole workspace | Set to the policy before rollout, because turning it off later deletes members' saved memories |
| Default seat type | Chosen by the owner | Standard, with Premium only for roles whose usage needs it |
| GPT settings | Every role can create a GPT; owners manage the settings | Creation and sharing limited per policy, with an owner for every published GPT |
How a ChatGPT workspace setup runs
The order is the same for a new workspace and for one we take over:
- Scoping: use cases, data types, headcount, identity platform, existing personal accounts and any workspace that already exists.
- Plan decision: Business or Enterprise, Standard or Premium seats, recorded with the reasons and the controls each choice gives up.
- Ownership: the workspace is created or taken over under a company-owned account, with two owners and the billing contact set.
- Identity: the domain is verified by DNS record, single sign-on is configured and tested with one user, and automatic account creation and discovery are decided.
- Controls: roles, seat defaults, apps and their approval policy, GPT and agent permissions, memory, sharing and spend limits are set and recorded.
- Pilot: a small group is invited with their existing emails, personal workspaces are merged where the person wants it, and the safe-use briefing is delivered.
- Adjust and roll out: settings are corrected from the pilot, and the rest of the company is invited by CSV with role and seat type.
- Handoff: the configuration record, offboarding runbook and open decisions go to your admin, or carry into managed AI services.
Dates depend on how quickly DNS and identity changes can be made in your tenant and, for Enterprise, on OpenAI's own contract steps, so they go in the proposal rather than on this page.
Deliverables, owners and boundaries
Each deliverable has an owner on both sides and a boundary written into the scope:
| Deliverable | NetSys | Your company | Boundary |
|---|---|---|---|
| Plan recommendation | Writes it, with the controls compared | Signs the OpenAI agreement and pays for seats | Seats are bought from OpenAI under your own account and billing |
| Workspace and identity configuration | Configures and tests it | Owns the accounts, the domain and the identity provider | Owner credentials stay with your named people |
| Settings record | Produces it against the official defaults | Approves each deviation from the policy | Recorded as of the handoff date; OpenAI changes defaults |
| Connected app approvals | Reviews scopes and proposes the list | Your Entra or Google administrator grants consent | Only apps a use case needs; write actions behind a prompt |
| Personal account merges | Explains the merge and assists each person | Each employee decides, and exports first | Merges are permanent; NetSys never runs one without the person |
| Offboarding runbook | Writes it | Runs it, or hands it to managed AI services | Business keeps a removed member's chats; the runbook says who owns what next |
What is included, and what is scoped separately
Training, custom builds and ongoing administration each have their own page and are quoted separately:
| Area | Included in workspace setup | Scoped separately |
|---|---|---|
| Plan and purchase | The recommendation and help through self-serve checkout or the Enterprise sales process | OpenAI's seat fees, credits and contract terms |
| Identity | Domain verification, single sign-on and, on Enterprise, SCIM against your existing directory | Building or cleaning up the identity platform itself (Microsoft 365 or Google Workspace work) |
| Settings | Roles, seats, apps, GPT and agent permissions, memory, sharing and spend controls | Connecting the Enterprise Compliance Platform to a DLP or SIEM tool, quoted as a project |
| Staff | A pilot group, personal-account merges and the safe-use briefing | Role-based training for the whole company |
| Custom tools | Governance settings for GPTs, agents and plugins | Building GPTs, workspace agents, plugins or integrations |
| After handoff | The configuration record and runbook | Monthly administration, reviews and reporting |
The services on the right-hand side: Corporate AI and ChatGPT training for employees · Custom GPT and workspace agent development · OpenAI and ChatGPT integrations · Managed AI services.
Rollout acceptance checklist
The setup is accepted when every line is true. The full step-by-step version is in our setup checklist post:
- The workspace is owned by a company account with two named owners and a billing contact.
- The domain is verified, single sign-on is tested, MFA is enforced in the identity provider, and discovery and automatic account creation are set deliberately.
- Every member is on the right role and seat type, and nobody holds admin rights without a reason.
- Connected apps are reduced to the approved list, write actions are set to ask, and Microsoft Graph consent is recorded.
- GPT, agent and plugin creation and sharing follow the policy, with an owner for every published item.
- Spend limits, usage alerts and auto-reload are decided; memory and shared-link settings match the policy.
- Pilot members' personal accounts are merged or kept separate by their own choice, after an export.
- The configuration record, offboarding runbook and safe-use briefing have been delivered.
Step by step: ChatGPT Business setup checklist for a company.
What moves the quote for ChatGPT setup
NetSys publishes no price list. The setup is quoted as a fixed scope after the scoping call, and these are the inputs that move it:
| Factor | Why it moves the price |
|---|---|
| Plan | Enterprise adds SCIM, role design, Compliance Platform decisions and the sales process |
| Headcount and roles | More members mean more seat and role decisions, merges and briefing sessions |
| Identity platform state | A clean Entra ID or Google Workspace tenant is quick; one with shared logins and no MFA needs work first |
| Connected apps | Each Microsoft or Google app needs consent, a scope review and an approval policy |
| Existing workspace | Taking over a workspace someone set up personally adds an ownership and settings audit |
| Regulated data | HIPAA or client-confidentiality rules can rule out Business and add retention and residency decisions |
On the vendor costs: ChatGPT Business pricing: licenses, setup and ongoing costs.
Managed AI services afterward run month to month. Our pricing post explains the vendor-side costs: seats, Premium usage and credits.
Staff already on personal ChatGPT accounts?
That is the usual starting point. Tell us the headcount and whether you run Microsoft 365 or Google Workspace, and we will outline the workspace, the identity setup and how the merges work.
ChatGPT Business Setup FAQs
What is ChatGPT consulting?
ChatGPT consulting is help deciding how a company should use ChatGPT and then setting it up properly: which plan fits (ChatGPT Business or ChatGPT Enterprise), which use cases and data types are approved, how the workspace is owned and secured, which apps and custom tools are allowed, and how staff move off personal accounts. NetSys delivers it as a fixed-scope setup with a configuration record at the end, rather than as advice alone. Our AI consulting page covers the wider question of where AI fits in the business at all.
What is the difference between ChatGPT Team and ChatGPT Business?
They are the same plan. OpenAI renamed ChatGPT Team to ChatGPT Business on August 29, 2025; existing Team workspaces became Business workspaces with no migration, and invoices began saying ChatGPT Business. Pricing and seat options have changed since: Premium seats were added on August 24, 2026, and the cap is now 200 paid seats per subscription. Anything you read about Team limits is out of date. Source: OpenAI help center, What is ChatGPT Business and the ChatGPT Business general FAQ, read October 17, 2026.
Should a small company choose ChatGPT Business or ChatGPT Enterprise?
Start from the controls you must have. Business covers most companies of 2 to 200 people: SAML or OIDC single sign-on, domain verification, an admin console, usage analytics, shared projects and no training on your data by default. Move to Enterprise when you need SCIM provisioning from your directory, role-based access per feature, the Compliance Platform for eDiscovery or DLP, custom retention, IP allowlisting or a BAA, none of which Business offers as of October 2026, or data residency, which OpenAI's pricing comparison lists for Enterprise only while a help article says it is rolling out gradually to Business at checkout. More than 200 paid seats also requires Enterprise. Source: OpenAI's business pricing comparison, read October 17, 2026.
Can we require staff to use the company workspace instead of personal ChatGPT?
Partly. With a verified domain, workspace discovery lets anyone with a company email find and request to join the workspace (or join at once where owners auto-accept), and single sign-on ties access to your identity provider. On Business, admins cannot force anyone to merge or delete a personal workspace; on Enterprise, onboarding requirements can require the merge, and Enterprise can request a restriction so that company-domain accounts connect apps only from authorized workspaces. The rest is policy plus controls on managed devices, which our AI governance service covers. Sources: OpenAI help center articles on single sign-on for ChatGPT Business and on connected apps, read October 17, 2026.
Does OpenAI train its models on our company's ChatGPT data?
Not by default. OpenAI's enterprise privacy page states that data from ChatGPT Business, ChatGPT Enterprise and the API is not used to train its models unless you explicitly opt in, and that you own your inputs and outputs. The Business help center adds that this covers memories, data accessed through apps, GPT conversations and projects. Personal plans are different: training is on unless the user turns off the improve-the-model setting, and feedback can send a whole conversation for training. That gap is one reason to move company work into the workspace. Sources: openai.com/enterprise-privacy and the ChatGPT Business help article, read October 17, 2026.
Can our admins read employees' ChatGPT conversations?
OpenAI's two official pages describe it differently, so we quote both. The enterprise privacy page says Business workspace admins can view, access, export and delete end-user conversations. The Business help center says each member has their own history, that admins and owners cannot see private member chats by default, that usage analytics show no transcripts, and that owners can export a member's data from the Members page where that option is available. In practice there is no live admin view of chats, but an owner can export them. We put that sentence in the safe-use briefing so staff know where they stand. Sources read October 17, 2026.
How is ChatGPT implementation priced, and do you resell licenses?
The setup is quoted as a fixed scope after a scoping call, based on the plan, headcount, the state of your identity platform, the apps to approve and whether an existing workspace must be taken over. NetSys publishes no price list and resells no licenses: you buy seats from OpenAI under your own account and billing. As of October 2026, Business Standard seats are $25 per user per month, or $20 billed annually, and Premium seats $125, or $100 annually; Enterprise is custom pricing. Scoping and pilot sessions run on site across New York City, Long Island, Westchester, the Hudson Valley, North Jersey and Fairfield County, CT, and remotely elsewhere. Source: OpenAI help center, What is ChatGPT Business, read October 17, 2026.
How do consulting firms themselves use ChatGPT safely?
Firms whose product is advice (management consultants, accountants, lawyers) use ChatGPT for first drafts, summaries of interviews and meeting notes, search across their own past work through Projects or Company Knowledge, and a first pass at analysis. The rules matter more than the features: client material goes only into the company workspace, on a plan whose data terms the firm has read; a consultant reviews every client-facing output; and the engagement letter says what the firm's AI use is. That is the setup this page delivers, with the safe-use briefing covering the review step.
Sources and technical references
- OpenAI help center: What is ChatGPT Business (the Team rename, seat prices and minimums, sales-led options)
- OpenAI help center: ChatGPT Business general FAQ
- OpenAI: business pricing and the Business versus Enterprise feature comparison
- OpenAI: enterprise privacy (training, ownership, retention and admin access)
- OpenAI help center: managing data sharing and privacy in ChatGPT Business
- OpenAI help center: setting up single sign-on for ChatGPT Business (SAML and OIDC, no SCIM, personal workspaces)
- OpenAI help center: members, seat types and roles in ChatGPT Business (workspace discovery default, member invites, GPT settings)
- OpenAI help center: billing and seats in ChatGPT Business (the 200-seat limit and seat reductions)
- OpenAI help center: administering connected apps and plugins (defaults, approval settings, Microsoft Graph consent)
- OpenAI help center: spend controls in ChatGPT Business
- OpenAI help center: memory controls in ChatGPT Business
- OpenAI help center: multi-factor authentication (admins cannot enforce it at workspace level)
- OpenAI help center: merging a personal workspace into ChatGPT Business
- OpenAI help center: onboarding employees who already have ChatGPT accounts
- OpenAI help center: business associate agreements (none for ChatGPT Business)
- OpenAI help center: the scheduled retirement of custom GPTs and migration to plugins
Guides on this topic
- ChatGPT Business setup checklist for a company
- ChatGPT Business pricing: licenses, setup and ongoing costs
- ChatGPT Business vs Enterprise: which fits your company
- ChatGPT Business vs Plus and Pro for a company
- ChatGPT Business security settings and data privacy checklist
- ChatGPT single sign-on setup for Business and Enterprise
- Moving personal ChatGPT use into a company workspace
- Corporate AI and ChatGPT training for employees
- Managed AI services: administration after setup
- AI governance consulting and acceptable use policies
- Claude consulting and business implementation
- Microsoft 365 Copilot consulting and rollout
- Single sign-on with Microsoft Entra ID
- Free AI readiness assessment
- Case study: Copilot rollout and custom agents for 18 people
- AI services overview
Related services
Put the company, not an individual, in charge of the workspace.
Tell us the headcount, the use cases and your identity platform. We will say whether Business or Enterprise fits, what the setup includes and what would be scoped separately, then quote it after a short scoping call.