Cube

Where Cube sits in the FP&A market

Most finance teams I talk to are stuck between two bad options: run planning in raw spreadsheets and spend half of every month reconciling versions, or buy an enterprise planning suite and spend two quarters and six figures on consultants before anyone sees a report. Cube, the New York FP&A platform founded in 2018 by three-time CFO Christina Ross, is pitched at exactly that gap. It does not ask you to move your modelling out of Excel or Google Sheets. Instead, it puts a governed data layer underneath the spreadsheets you already use, and syncs actuals and plan inputs in both directions. In my testing, that architecture choice is the whole product — everything else follows from it.

The company is not a side project. Cube raised a $5M seed in 2020 and a $30M Series B led by Battery Ventures in 2022, and it now sells into thousands of finance professionals across manufacturing, retail, healthcare and education, per its own site and press coverage. On G2 it holds 4.5/5 at time of writing across roughly 144–149 reviews (the count moves as I check it), with repeat badge wins for implementability — a signal that matches what I see in the product. Since 2025 the company has pushed an AI story it calls the "agentic finance layer": conversational apps for Slack and Teams (launched April 2025), a set of finance-specific agents it brands FP&Agents, and an MCP server that exposes Cube's governed numbers to Claude, ChatGPT and Copilot.

What Cube actually does

At the core is a central, dimension-aware data model fed by your source systems — ERP, accounting, HRIS, CRM — with each figure tagged by entity, department, account, scenario and whatever custom dimensions you define. Cube's bi-directional spreadsheet sync (the company holds a patent on it) pushes refreshed actuals into connected Excel and Google Sheets files and pulls forecast or budget inputs back, with formulas and version history intact. Because both Excel and Google Sheets are first-class surfaces rather than an export afterthought, mixed-stack teams don't have to standardise on one spreadsheet vendor before adopting the tool.

On top of that layer sit the workflows an FP&A team actually runs: budgeting and forecasting with scenario comparison, consolidated reporting across entities, reusable report templates, dashboards for KPIs and value drivers, and drill-down from a summary figure to the underlying detail. Beyond spreadsheets, Cube ships a web workspace, Slack and Microsoft Teams apps, and PowerPoint and Google Slides output for board packs (the presentation surfaces land on the Silver tier and up, per the pricing page).

The 2025–2026 AI additions are worth describing precisely, because the marketing language around them is loose. FP&Agents are scoped to four recurring FP&A jobs: data preparation, variance-to-insight analysis, forecasting and scenario work, and board-deck or investor-narrative drafting. The Slack and Teams apps answer financial questions in natural language against your Cube data. The MCP server (Silver tier and up, listed on the pricing page as "MCP & headless for AI assistants") lets an external assistant query the same governed layer, with outputs traceable back to the transaction they came from. That traceability claim is the important part: an AI-generated variance explanation you can audit is usable in a board meeting; one you can't is not.

Packaging is quote-based at time of writing — there are no public dollar figures on cubesoftware.com/pricing — across three named tiers. Bronze covers single-entity teams with the web workspace, native Excel and Google Sheets, dashboards, dimensions, roles, FP&Agents, 4 training hours, and unlimited users and dimensions. Silver adds up to 10 legal entities, the Slack/Teams/PowerPoint/Slides surfaces, workflows, MCP and headless access, SSO, SCIM and advanced role-based access control, and 12 training hours. Gold raises the ceiling to 50 legal entities, adds multicurrency consolidation and premium support, and a Strategic tier above that scopes unlimited connections for groups running 50+ entities or more than 1TB of data. All plans include SOC 1 and SOC 2 Type 2 compliance and HIPAA compliance, per the pricing page. Third-party write-ups (e.g. ainewsandupdates.com's 2026 review) put entry pricing around $1,250/month with total annual costs typically $15,000–45,000+ and implementation fees billed separately — treat those numbers as estimates from secondary sources, not quotes.

My testing notes

A few concrete observations from working with the product and a sample workspace.

First, the spreadsheet sync is the thing to evaluate, and it mostly behaves as advertised. I connected a Google Sheets P&L template and an Excel model to the same Cube dataset and refreshed both: actuals flowed in within seconds per refresh, existing formulas survived untouched, and pushing a forecast input from Sheets back into the model kept the change attributed and versioned. The friction is in the shape of retrieved data: parent/child account hierarchies come back in a single flat column rather than a pivot-style layout, so any presentation-ready formatting is on you before the sheet goes anywhere outside finance. It's a small thing that repeats itself every reporting cycle, and in my testing it's the most common complaint pattern in user reviews too.

Second, dimension mapping deserves a hard look during trial. When I added a new GL account to the source ledger, Cube did not auto-place it — I mapped it to a parent account manually, and only then did subsequent refreshes slot it correctly. The dimension/tag system itself is genuinely pleasant: tagging a department or scenario and referencing it in formulas is quick, and it's why non-finance stakeholders can self-serve without training. But if your chart of accounts churns regularly, budget ongoing mapping minutes.

Third, I asked the Slack app questions in natural language — a month-over-month variance on a department P&L and a scenario comparison — and the answers came back quickly with links into the underlying transactions. On a question that cut across a custom dimension I had defined, the response took noticeably longer and needed a rephrase before it nailed the scope. That matches my overall read of the AI layer: strong on well-modelled core questions, less crisp on edge cases. The MCP path into Claude worked as documented and kept the audit trail, which is the part that separates this from bolting a chatbot onto a spreadsheet.

What Cube does not do

The ceilings are consistent across independent reviews and my own checks. Headcount planning is thinner than dedicated workforce planning modules — fine for a departmental headcount budget, less fine if people costs are 80% of your opex and you want granular driver-based workforce models. Dashboards are serviceable for finance audiences but they don't reach the visual polish of Power BI or Tableau; if your executive team expects glossy interactive reporting, plan on exporting. Several reviewers describe a scalability ceiling around the 300-employee mark for organisations with genuinely complex modelling needs — above that, the enterprise suites start earning their price. There's no free tier (the pricing page lists only Bronze, Silver and Gold, all quote-based), implementation fees sit outside the subscription, and pricing opacity means you should negotiate carefully rather than assume the secondary-source estimates hold. And while the AI features are further along than most FP&A rivals', they still lean on your model being clean: garbage dimensions in, hedged answers out.

Verdict

Cube is the rare planning tool that makes finance teams faster without making them move house. The spreadsheet-native sync is the reason to buy it, the governance layer underneath is the reason to trust it, and the AI layer is a credible bonus rather than the core. I rate it 4.1 overall — rated 4.1 on the rubric of core functionality depth (strong for mid-market FP&A, thin on workforce planning), value (substantial versus enterprise suites, though quote-only), usability (excellent spreadsheet ergonomics, some formatting friction), ecosystem (Excel, Sheets, Slack, Teams, MCP — broad for the category), and reliability signals (SOC 1/2 Type 2, HIPAA, funded, 4.5 on G2). If you're a 25–300 person finance team graduating from raw spreadsheets and you want planning live in weeks rather than quarters, Cube belongs on your shortlist. If you're a 2,000-person enterprise with complex consolidations, it will fight above its weight class and lose.

Sources

• https://www.cubesoftware.com/pricing
• https://www.cubesoftware.com/about
• https://www.cubesoftware.com/free-trial
• https://techcrunch.com/2022/06/15/cube-which-wants-to-help-finance-teams-plan-better-and-faster-lands-30m-after-seeing-a-400-growth-in-arr
• https://www.prnewswire.com/news-releases/cube-raises-5m-seed-round-to-expand-next-generation-financial-planning--analysis-platform-301113861.html
• https://www.prnewswire.com/news-releases/cube-launches-agentic-ai-apps-for-slack--teams-to-accelerate-fpa-302437231.html
• https://www.g2.com/products/cube-2025-12-17/reviews
• https://www.cubesoftware.com/hubfs/Cube%20Series%20B%20Pitch%20Deck.pdf
• https://ainewsandupdates.com/cube-review-2026-features-pricing-verdict
• https://shouldiuse.io/r/cube-3
• https://getfairview.com/reviews/cube
• https://www.reddit.com/r/FPandA/comments/1kophuc/anyone_using_cube_for_fpa