Team
10 tracked changes
ChatGPT Work creates a distinct outcome-agent surface between Chat and Codex.
Buying-impact feed
Recent tracked changes that could move a recommendation
10 meaningful changes across 9 tools in the last 31 days.
Changes tracked
10 - Meaningful product, pricing, and access changes still inside the last 31-day review window.
Tools touched
9 - Covered products with at least one change that can move shortlist language or cost expectations.
Comparisons flagged
19 - Direct compare pages linked to those changes through recentDelta or verdict review.
Urgent reviews
0 - Changes whose pricing or packaging delta is strong enough to demand immediate editorial follow-through.
Original research
Recheck pricing changes against official sources and cost assumptions
Compare price sources and change evidence across 20 tools, with 1-, 5-, and 25-seat scenarios for three core tools.
My stack impact
See changes that actually touch saved stacks
This panel compares locally saved watchlist stacks against the change feed so returning readers can start with what affects their shortlist.
Checking saved stacks in this browser.
Change timeline
Review buyer impact by week
The key metrics stay compressed above while the full evidence remains in the weekly timeline.
Segments and follow-throughExpand impacted segments and comparison follow-throughOperational evidence stays rendered below the summary without crowding the timeline.
Which buyer segments moved the most
This view infers where recommendation language is most exposed by combining change type, severity, and the tool's strongest fit contexts.
Enterprise
8 tracked changes
Claude Cowork expands Claude into a scoped desktop outcome agent for paid plans.
Individual
2 tracked changes
ChatGPT Work creates a distinct outcome-agent surface between Chat and Codex.
Comparison follow-through
Each affected comparison should reflect the change in recentDelta or verdict language.
- ChatGPT vs ClaudeReflected
Budget Work as an agentic consumption surface and validate workspace eligibility, roles, enabled apps, allowed actions, and cloud-versus-local access before rollout. Do not assume an included chat seat grants every file-creation or desktop-control workflow.
- ChatGPT vs GrokReflected
Business becomes a lower fixed-seat baseline, but coding-heavy teams must separately forecast token-credit consumption. Do not use the Business seat price as the complete Codex budget.
- ChatGPT vs PerplexityReflected
Business becomes a lower fixed-seat baseline, but coding-heavy teams must separately forecast token-credit consumption. Do not use the Business seat price as the complete Codex budget.
Business becomes a lower fixed-seat baseline, but coding-heavy teams must separately forecast token-credit consumption. Do not use the Business seat price as the complete Codex budget.
- Notion AI vs ChatGPTReflected
Business becomes a lower fixed-seat baseline, but coding-heavy teams must separately forecast token-credit consumption. Do not use the Business seat price as the complete Codex budget.
Weekly grouping
Week of Aug 10, 2026
8 meaningful changes landed in this weekly review bucket.
- ChatGPTBest AI meeting assistants by suite and follow-throughBest AI research assistants for sourced decision-makingChange historyAug 14, 2026HighUpdate
ChatGPT Work creates a distinct outcome-agent surface between Chat and Codex.
ChatGPT Work handles longer multi-step deliverables across enabled apps and files, including documents, spreadsheets, presentations, reports, and Sites. Cloud Work is available through web and mobile, while desktop access to local files and apps needs permission. Its usage follows the same structure as Codex rather than an unlimited-chat model.
TeamIndividualBuying impact
Budget Work as an agentic consumption surface and validate workspace eligibility, roles, enabled apps, allowed actions, and cloud-versus-local access before rollout. Do not assume an included chat seat grants every file-creation or desktop-control workflow.
- ClaudeBest AI research assistants for sourced decision-makingBest AI writing tools for real team workflowsChange historyAug 14, 2026HighUpdate
Claude Cowork expands Claude into a scoped desktop outcome agent for paid plans.
Cowork performs multi-step work across local files and apps from the Claude desktop app. Anthropic describes containment around the user-selected workspace and a VM, while mount modes and approved tools or connectors determine the practical access boundary. It is available on paid plans, but consequential output still requires human review.
TeamEnterpriseBuying impact
Treat the selected workspace, mount policy, and enabled connector or MCP set as the governance unit—not only the Claude seat. Pilot on approved folders with a review gate, separate usage or spend cap, and a clear escalation path for consequential actions.
- ChatGPTBest AI meeting assistants by suite and follow-throughBest AI research assistants for sourced decision-makingChange historyAug 13, 2026HighUpdate
ChatGPT Business annual pricing is now $20 per user per month, while Codex usage moves to a token ratecard.
OpenAI reduced annual Business pricing from $25 to $20 per user per month while monthly pricing remains $30. Codex usage and purchased credits now follow the token ratecard, and new Codex-only pay-as-you-go seats are no longer available for Business workspaces.
TeamEnterpriseBuying impact
Business becomes a lower fixed-seat baseline, but coding-heavy teams must separately forecast token-credit consumption. Do not use the Business seat price as the complete Codex budget.
- Notion AIBest AI meeting assistants by suite and follow-throughBest AI writing tools for real team workflowsChange historyAug 13, 2026HighUpdate
Notion autonomous work is now a shared-credit operating model, not simply included workspace AI.
Custom Agents and Workers consume a shared Notion-credit pool. Workers began requiring credits on August 11, 2026; allocations and limits can pause credited work when the cap is reached.
TeamEnterpriseBuying impact
Keep bundled core AI and autonomous execution in separate budget lines. Before scheduling work, set the spend owner, allocation, alert threshold, hard limit, and human fallback for a paused workflow.
- Aug 13, 2026HighUpdate
Grok now spans automations, background workflows, app building, and a metered Voice Agent Builder beta.
xAI added scheduled and email-triggered automations, background Build workflows, Build Mode, and the Voice Agent Builder beta with telephony, retrieval, tools, guardrails, MCP, observability, and human transfer. Voice audio is priced at $0.05 per agent minute; a provisioned free number adds $0.01 per minute.
EnterpriseTeamBuying impact
Evaluate Grok by job surface, not as one undifferentiated chat seat. A voice-bot pilot should have an approved knowledge set, tool and transfer boundaries, recorded-call review, retention decisions, and a hard spend cap; it is not included in a Business seat.
- Aug 13, 2026MediumUpdate
Cursor's Auto router adds team-governed quality modes with routed-model billing.
Auto can select Intelligence, Balance, or Cost. Balance and Intelligence bill at the routed-model rate, while admins can manage modes, models, and providers by team or group.
EnterpriseTeamBuying impact
The seat price is not the whole budget for agent-heavy teams. Start Cost mode as the baseline, then enable higher modes for a controlled group only after their output lift justifies routed-model spend.
- Aug 13, 2026MediumUpdate
Replit makes active background-task capacity and effort-based Agent usage a planning constraint.
Replit documents one active background task for Core and ten for Pro. Agent usage is effort-priced, and some third-party API usage passes through at the public rate.
TeamIndividualBuying impact
Choose Core or Pro based on intended concurrent background work, then model Agent effort and API pass-through separately from included credits. A plan price is not a complete all-in cost ceiling.
- GeminiBest AI meeting assistants by suite and follow-throughBest AI research assistants for sourced decision-makingChange historyAug 13, 2026MediumUpdate
Google Workspace's new-customer promotion lowers the first three months, not the long-term Gemini baseline.
Eligible new customers can receive three-month promotional pricing for Workspace Business Starter and Standard through October 7, 2026. The public post-promotion rate remains the appropriate run-rate for longer planning.
TeamEnterpriseBuying impact
Use the incentive to reduce onboarding cost, but do not claim annual savings from the promotional rate. Compare the steady-state Workspace price with any separate assistant or Microsoft 365 base-plan-plus-Copilot alternative.
Weekly grouping
Week of Aug 3, 2026
2 meaningful changes landed in this weekly review bucket.
- Aug 9, 2026MediumUpdate
Synthesia now publishes annual Starter and Creator prices as $264 and $804.
Synthesia's current pricing FAQ lists Starter at $29 per month or $264 per year and Creator at $89 per month or $804 per year. The tracked records previously used lower annualized figures.
TeamEnterpriseBuying impact
Annual buyers should recalculate the paid baseline before comparing Synthesia with a social-video tool. The updated annual figures are still not a like-for-like comparison with a vendor that meters output differently.
- Aug 9, 2026MediumUpdate
Zebracat's public pages now describe paid capacity in different units across product surfaces.
The pricing page describes Cat Mode as 15 videos and 350 generative credits per month, while the public product page presents a separate credits-and-approximate-minutes view. The same public pages list Cat Mode at $39 per month and an annual view at $19 per month.
TeamEnterpriseBuying impact
A team should not forecast Zebracat spend from the headline price alone. Confirm the current checkout allowance, its unit, and whether revisions or generative work consume the same pool before approving a publishing-volume plan.
Watchlist
Track changes for this shortlist
Save the stack, monitor buying-impact changes, and turn the result into a decision memo.