Cue workspace context graph. 4 hubs.
BaseThread is the shared context layer for every AI tool your team uses. Below are example questions answered from a sample company's shared workspace, the same way your own team's AI would answer from yours.
A lot, most of it pointed at the Copilot GA launch. AI Engineering got citation quality v2 past the eval bar (the last hard gate before GA), Frontend shipped local-timezone rendering across the app, and Sales logged a mid-market win where Copilot was the deciding factor. Elena Vasquez also walked the readiness checklist with every team lead and got sign-off pending the final quality gate.
Elena Vasquez (CPO) decided Copilot ships free for a 60-day GA window, then converts to paid, to drive adoption and proof during the company's bet to make Copilot the reason teams pick Cue. This decision superseded an earlier plan to launch Copilot as a paid add-on on day one. The older decision is kept and linked, marked superseded, so the history of why is intact.
A few things, and they are the gate. Arjun Mehta's launch-readiness checklist is in progress: citation v2, the SLO load test, and the rollback rehearsal are done, but the support runbook sign-off and Sales enablement are still open. Naomi Tran's team has design QA on the GA experience in progress, and the enterprise Copilot pricing decision is still pending, so reps cannot quote a number yet.
Close. The hard quality gate is green: citation quality v2 passed the eval bar, the endpoints held the 99.9% SLO under launch load, and the rollback was rehearsed. Elena Vasquez has sign-off from Product, Engineering, Design, PMM, Sales, and CX. The one open risk is pricing after the free window, the enterprise model is actively contested and deliberately held until it resolves.
It is genuinely contested between two people. Bianca Rossi (Enterprise AE) logged a decision to commit to flat per-seat pricing for enterprise Copilot, because procurement distrusts consumption pricing. Daniel Osei (Senior PMM) logged the opposite in the pricing revamp: move enterprise Copilot to usage-based, to align price to value. Both are flagged as a conflict and left unresolved for humans to settle, with Robert Klein (CFO) on the numbers.
Grounding is the whole product, so Arjun Mehta set a hard rule: Copilot does not go to GA until citation quality v2 passes the eval bar and the hallucination rate is below threshold, with a documented eval run. Hana Sato even expanded the benchmark with 40 real answers that had exposed weak citations, so the harness catches them before customers do. As of this week, the gate is green.
It is the gate Copilot cannot skip, owned by Arjun Mehta and signed off by every team lead. Done: citation quality v2, the 99.9% SLO under load, and a rehearsed rollback. Still open: the support runbook sign-off (CX) and Sales enablement, with enterprise pricing explicitly marked pending so no rep quotes a number.
It is cross-functional by design. Elena Vasquez (CPO) is accountable for scope and the GA call and owns onboarding. Arjun Mehta owns citation quality and launch stability. Owen Mitchell owns the GA experience, Daniel Osei owns messaging and win-story capture, Bianca Rossi owns design-partner-to-reference conversion, and Joy Abara owns CX enablement and the support runbook.
Yes. Omar Farouk closed a mid-market deal and tagged the win note that the buyer chose Cue specifically for grounded Copilot answers. Bianca Rossi converted Titan Fitness into a named reference on the time-savings story, where prep dropped from a day to under an hour. The GA goal is at least five named win stories where Copilot was the deciding factor.
Copilot converts to paid, but the exact model after the window is the open question. The pricing revamp is deciding it, and the enterprise piece specifically is stuck on the flat-per-seat versus usage-based conflict. The GA launch deliberately leads on value and grounding and holds on enterprise price until that conflict resolves, so Sales marks enterprise pricing as pending rather than quoting a number.
Two teams reached opposite conclusions about enterprise Copilot pricing. Sales (Bianca Rossi) wants flat per-seat for predictability and procurement comfort. PMM (Daniel Osei) wants usage-based to align price to value and match the Pulse model. BaseThread detected that the two logged decisions contradict, flagged them as a conflict, linked them, and surfaced it on both teams and the pricing project, then left it for humans to resolve.
Bianca Rossi's case: enterprise buyers want predictable budgets and procurement distrusts consumption pricing, so flat per-seat is easier to get through a deal. She also has live enterprise deals asking for a number now, which is what forced the question into the open.
Daniel Osei's case: usage-based aligns price to value, matches Pulse's proven consumption model, and protects margin as Copilot usage grows. He proposes a committed-use floor so procurement still gets predictability. Data Science is modeling revenue under each option to inform the call.
No, and that is on purpose. The conflict is flagged and unresolved, and all external pricing comms are on hold until it settles. Grant Webb (CRO) flagged to leads that reps cannot quote enterprise Copilot pricing until the flat-per-seat versus usage-based question resolves in the pricing revamp.
This is the part most tools miss. Bianca logged her decision from Sales and Daniel logged his from PMM, days apart, in different parts of the workspace. Neither saw the other. BaseThread noticed the two settled decisions contradict, linked them as a conflict, surfaced it on both teams and the pricing project, and left it for humans, so the disagreement is visible instead of buried in two separate threads.
Cue's pricing grew product by product and no longer tells one story: Signal is seat plus source, Pulse is usage-based, Copilot is undecided. Daniel Osei owns a revamp to unify the suite onto Team, Business, and Enterprise tiers so a rep can explain it in two minutes, with a decided model for Copilot including enterprise. The enterprise Copilot model is the one piece still blocked by the conflict.
Data Science and Finance. Pierre Moreau delivered revenue and margin models for flat per-seat versus usage-based to the pricing revamp, and Sophie Tremblay (FP&A) modeled margin under the proposed packaging and the committed-use Copilot floor. Nothing goes external until the CFO signs off the migration math.
Signal's navigation was built when workspaces had dozens of themes. At 200-plus themes it strains, so power users build workarounds and new users get lost, which caps how far an account can grow. Owen Mitchell owns a redesign that makes the path from a theme to its underlying feedback the fastest action in the product and holds up cleanly at 500-plus themes.
Aisha Bello ran twelve sessions plus a diary study. Three findings drove the redesign: scanning breaks down past roughly 100 themes, revenue weighting is loved but buried, and the path from a theme to its quotes is the magic moment but two clicks too many. The team has a rule that no study runs without a decision attached, so each finding maps to a design change.
In phases with a way back, no big-bang cutover on a daily-use tool. Internal and design partners first, then opt-in beta with a one-click switch back, then default for new workspaces, then migrate existing accounts in waves with the largest and most vocal last. The team advances only when the new view matches or beats the old on trust and time-to-first-insight.
More than you'd expect for an in-flight redesign. Tom Whitfield built the new theme explorer behind a flag, Priya Nair profiled it at 500 themes and landed scroll and search within target, and the team opened it to internal dogfooding. Hassan Qureshi also closed a keyboard-navigation accessibility gap ahead of an enterprise review.
Whether theme merge and split should be manual, Copilot-assisted, or both. Research recommends testing a Copilot-assisted version once GA stabilizes, and Anika Gupta already prototyped Copilot suggesting theme merges, measured against the benchmark. It is logged as an open question rather than quietly decided.
Almost. Alan Pierce (CISO) owns a program to clear the security reviews that keep stalling enterprise deals. SSO, SAML, and SCIM are live in production, and a signable DPA with data-residency terms is done. The long pole is SOC 2 Type II, which is in progress and expected by end of year.
Yes. Yara Saleh shipped SSO, SAML, and SCIM to production, which both unblocks enterprise login requirements and captures the access-control evidence the SOC 2 program needs. Infrastructure also began region-pinning groundwork so EU customer data can be pinned for the data-residency option.
Diana Foster (General Counsel) finalized a standard, signable DPA with data-residency terms, which removes a frequent blocker in enterprise legal review. The DPA gates several enterprise deals independent of the SOC 2 audit, so getting it done was high leverage.
An honest narrative, reviewed by Legal and Security: data is encrypted in transit and at rest, access is role-based, Cue never uses customer data to train models, SSO and SCIM are available, and a DPA with EU residency is offered. On SOC 2, reps say it is in progress, expected by end of year, and never imply it is already held. That last point is both a values rule and a legal one.
Not yet, and the company is careful to say so. The CEO committed to SOC 2 Type II by end of year, and the rule everywhere is to state it as in progress, never as held, until the audit completes. Evidence collection across the observation window is the long pole, since Type II requires showing the controls operated over time.
The SOC 2 evidence-collection window is the main one. Martin Reier's task to start continuous evidence capture is in progress and due imminently, with access-control evidence done but change-management evidence and an incident-response rehearsal still to log. Full on-prem and air-gapped mode are explicitly out of scope for this phase, logged as future, not now.
Mostly hardening for GA and Signal 2.0. AI Engineering passed citation v2, tuned retrieval to respect viewer permissions, and added grounding monitoring in production. Frontend shipped local-timezone rendering and built the new theme explorer behind a flag, and Infrastructure load-tested Copilot to the 99.9% SLO and rehearsed the rollback.
The motion is leading with Copilot. Omar Farouk logged a Copilot-deciding-factor win, Derek Cho advanced Polar Bank to a Copilot proof, and Aaliyah Johnson booked four qualified enterprise meetings this week. Helena Schmidt also drove a mid-market POC to first insight in nine days, well under the 14-day bar.
Owen Mitchell locked the Signal 2.0 navigation direction in critique, Isabel Romero designed the dense-data and empty states (the real test at 200-plus themes), and Kenji Watanabe promoted a theme-detail panel and a weighting badge into the shared library so 2.0 ships without one-offs. Naomi Tran ran design QA on the Copilot GA experience before opt-in beta.
Joy Abara shipped the Copilot support runbook and enabled the CX team before launch traffic, and refreshed the health-scoring model with Data Science toward weekly active insight-makers. Bruno Alves onboarded a new account to first insight in 11 days, and Ethan Clarke ran a QBR with Northwind Logistics off their churn-reduction story.
Anika Gupta raised theme-labeling precision above the 94 percent baseline and prototyped Copilot-assisted theme merge for Signal 2.0. Pierre Moreau modeled revenue under each Copilot pricing option for the conflict, and Samuel Eze instrumented the Signal 2.0 events so the team can compare time-to-first-insight in beta. Every metric is single-sourced in one catalog.
Yara Saleh shipped SSO/SAML/SCIM and cleared all high-severity vulnerabilities within SLA. Keiko Mori ran an incident-response tabletop and published the customer-data handling policy, and Martin Reier kicked off the SOC 2 evidence-collection window. Access reviews now run quarterly and on every role change, which doubles as SOC 2 evidence.
A handful of launch-critical items are due within days: Arjun Mehta's launch-readiness checklist, Naomi Tran's design QA on the GA experience, Aaliyah Johnson's four enterprise meetings, and Martin Reier's SOC 2 evidence kickoff. The enterprise pricing decision is also due soon and blocking, since it gates what reps can quote.
Mara Devlin (CEO) has the quarterly board pack in progress, due in about a week. Financials and progress on the three bets are done; the risks-and-watch-items and the asks for the board are still open. She also asked Alan Pierce to sign off the enterprise-readiness go/no-go.
Bianca Rossi (Enterprise AE) has the Polar Bank security questionnaire prep accepted and due in a few days, and converting two design partners (Titan Fitness and Polar Bank) to named Copilot references in progress. She is also the one pushing hardest for an enterprise pricing answer, since two live deals need a number.
Arjun Mehta (AI Engineering Lead) owns the launch-readiness checklist, which is in progress and due in days. He already cleared citation quality v2 (the hard gate), and his team expanded the eval benchmark and added production grounding monitoring. He is effectively the person standing between Copilot and GA.
Across the four big projects: Copilot GA launch has the readiness checklist and pricing decision open; Enterprise readiness has the SOC 2 evidence window and trust narrative in flight; the Pricing revamp is blocked on the enterprise model conflict; and Signal 2.0 has the theme explorer build and dense-state design in progress. Most open items trace back to two things, the GA launch and the pricing conflict.
The headline ones: Copilot ships free for a 60-day GA window then converts to paid (superseding the paid-add-on plan), Copilot GA is gated on citation quality v2, and the suite is unifying onto Team, Business, and Enterprise tiers. The enterprise Copilot pricing model is the notable decision that is still unresolved, flagged as a conflict.
Yes, and BaseThread merged them instead of leaving three duplicates. Grace Kim (Frontend), Owen Mitchell (Design), and Maya Iqbal (Support) each independently logged that all customer-facing dates and times should render in the customer's local timezone. Harmonization linked the three into one canonical decision with a corroboration count of three, which is a strong signal the whole company agrees.
It did. Grace Kim shipped local-timezone rendering across the app, implementing the standardization, with UTC kept only at the API boundary. Jordan Lee on Support then verified the fix with a confused customer and closed the issue. So you can trace it from three independent decisions, to one harmonized call, to shipped code, to a resolved ticket.
Each member's AI logs decisions as work happens, then BaseThread harmonizes them across the company. It catches three patterns automatically: a conflict (two decisions contradict, like the enterprise pricing one, flagged for humans), a supersession (a newer decision replaces an older one, like Copilot going free for 60 days), and a corroboration (several people log the same call, like the timezone fix, merged into one with a count). That is how a 100-person company keeps one honest source of truth.
Elena Vasquez is Cue's Chief Product Officer and the owner of the Copilot GA launch. She made the call to ship Copilot free for a 60-day window, confirmed the GA decision with every team lead, and is accountable for scope and onboarding. Most of the launch threads run through her.
Arjun Mehta is the AI Engineering Lead and effectively the gatekeeper for Copilot GA. He set the rule that Copilot does not launch until citation quality v2 passes, owns the launch-readiness checklist, and keeps the team model-agnostic via a provider abstraction so the cheapest model that clears the eval bar wins.
Bianca Rossi is an Enterprise Account Executive and the lead on the Sales side of enterprise Copilot. She logged the flat-per-seat pricing position, converted Titan Fitness to a named reference, and is pushing for a pricing answer because two live enterprise deals need a number. She is one half of the pricing conflict.
Daniel Osei is a Senior Product Marketing Manager and owns the pricing revamp. He logged the usage-based position for enterprise Copilot (the other half of the conflict), set grounding and citations as the lead Copilot message, and wrote the customer-facing trust narrative with Legal and Security.
Grace Kim is a Frontend Engineer. She authored the canonical decision to standardize customer-facing times to the customer's local timezone (independently corroborated by Design and Support), and then shipped that change across the app. A clean example of one person's call becoming a company-wide, harmonized decision and then real code.
Three company bets, set by CEO Mara Devlin: make Cue Copilot the reason teams pick Cue, win the mid-market upmarket motion, and be enterprise-ready by year end. Every team goal is supposed to ladder to one of these, and progress is reviewed monthly. The Copilot GA launch and the enterprise readiness program are these bets made concrete.
Cue is an AI-native customer insight platform. It pulls scattered customer signal (support tickets, sales calls, reviews, surveys, churn interviews, community threads) into one place and turns it into product decisions. The suite is Cue Signal (the flagship insight hub), Cue Pulse (in-app surveys), and Cue Copilot (the AI layer that answers questions across all the feedback with citations).
Five values, and they show up in the decisions. Customer truth over opinion, ship to learn, earn trust with security, default to clarity, and own the outcome. The security one is why the company refuses to claim SOC 2 before it is held and never trains models on customer data.
The flagship one is Northwind Logistics, which cut churn-related tickets 22 percent in one quarter after acting on a Signal theme. Ethan Clarke ran a QBR reviewing that result and opened an expansion conversation. It is the proof point Sales and Marketing lead with.
Because every answer comes from Cue's BaseThread workspace, the shared context every team writes to as work happens. Decisions, activity, and tasks across the company, products, teams, and projects all live in one place that every member's AI reads automatically. That is what you are exploring on the right. With BaseThread, your team's AI tools read your real context the same way, instead of starting from a blank slate.
They are the three streams BaseThread keeps for a team. Activity is what happened (the work journal). Decisions are what got settled and why (with the reasoning attached). Tasks are what needs doing, by whom, by when. Each member's AI writes to them as work happens, and they cross-reference each other and the people and files behind them, which is what makes the graph on the right feel alive.
It is Cue's context graph: a live map of the workspace where company, products, teams, projects, people, files, and the activity, decision, and task streams all connect. Click any node to open it, and hover to preview. It is the same view a teammate's AI traverses to answer a question, which is why it can follow a thread from a decision to the work and the people behind it.