IBM Planning Analytics Review
IBM Planning Analytics is an enterprise planning platform on the TM1 in-memory engine, used by large finance teams for budgets, forecasts and workforce plans in Excel or on the web, sold as SaaS on IBM Cloud, AWS or Azure, or installed on premises.
Independently scored, one dimension at a time. Every dimension is sourced, and we say whether a claim is published, reported, or estimated.
TM1 cubes with rules carry driver-based models, what-if scenarios and plan versions, and IBM publishes use cases for FP&A, workforce, sales, supply chain and IT planning (IBM product pages, checked 30 September 2026). AI forecasting (baseline, on-demand, spreading) starts at Standard, and PAfE gives Excel users write-back into the same model. Statutory consolidation is a separate product, IBM Controller.
TM1 keeps cubes in memory and consolidates on demand without precalculation (Wikipedia, checked 30 September 2026), which is why large enterprises run big models on it. The SaaS plans are sized at 16, 32 and 64 GB of RAM with 5, 10 and 20 users, with larger estates by quote (IBM pricing page). What breaks first is memory, and memory is also the lever IBM prices on.
IBM documents an in-product MCP server (release notes 3.1.4, 27 January 2026, and 3.1.9, 25 June 2026), and IBM's 26 April 2026 blog says it now reaches Local, on Cloud and Certified Containers as well as the SaaS. IBM's workshop repo (updated 15 September 2026) lists read tools for cubes and MDX plus write tools that create cubes and TurboIntegrator processes. It needs the PA Agent add-on entitlement and admin-issued credentials, and is not in the public MCP registry. In-product outcome claims for the PA Agent are vendor-claimed.
IBM's pricing table (checked 30 September 2026) puts auditing and logging, high availability, and development and production environments with source control in Standard and Premium only. PAfE applies role-based access so users see only what they are authorized to see (IBM Excel page). IBM's 13 December 2024 deployment guide says its IBM Cloud option (Planning Analytics on Cloud) suits regulatory needs 'such as SOC 1 type 2 & ISO 27K', and an IBM community answer (5 June 2025) says the Planning Analytics on Cloud SOC 1 Type 2 report comes through a support ticket. No public SOC 1, SOC 2 or ISO 27001 statement for the AWS and Azure plans was found.
IBM's own store lists Essentials at $10,500 for a 12-month term, or $875 a month, which is the 'Starting at $875.00 USD' figure on ibm.com (checked 1 October 2026). IBM's AWS Marketplace listing charges $9,900 for Essentials and $19,800 for Standard for 12 months (checked 30 September 2026). Premium, on-premises licenses, the PA Agent add-on and the SAP connector are quote only. AWS orders are non-cancellable and non-refundable.
SaaS customers open tickets through IBM Support (AWS Marketplace support tab), and most also lean on a partner. PeerSpot (checked 30 September 2026) shows 4.2 out of 5 from 28 reviews but summarizes support as mixed, with users escalating to IBM or partners. AWS Marketplace's AI-generated sentiment summary also rates customer service as mixed (checked 1 October 2026).
The one documented native ERP connector is the Planning Analytics Connector for SAP, which reads and writes SAP BW, S/4HANA and HANA Cloud over OData and is licensed per SAP instance (IBM docs). Other sources run through TurboIntegrator over ODBC or REST; on the SaaS, which runs TM1 12, IBM's docs say the database no longer calls ODBC drivers, so relational sources such as Snowflake go through the ODBC Data Connector that the customer hosts. No native NetSuite, Sage Intacct, QuickBooks Online, Xero, Business Central or Workday connector was found, and the August 2025 App Connect connector belongs to a separate IBM product.
Builds need TM1 skills (rules, feeders, TurboIntegrator and MDX), usually supplied by partners such as Cubewise, Octane, Aramar or QueBIT. PeerSpot's summary puts deployments from one month to a year. IBM offers a 30-day trial with prebuilt samples and the SaaS removes server work, but relational sources such as Snowflake still need IT to install and secure IBM's ODBC Data Connector on a customer machine (IBM docs).
Fast, deep TM1 models and a working MCP server, but you need a TM1 partner and your own data pipes.
Written by the Audit Friendly research team. No vendor edits, no sponsored placement.
You are buying TM1, a multidimensional in-memory engine that has run enterprise budgets since 1983, with a web client (Planning Analytics Workspace) and an Excel add-in (PAfE) that writes back to the same cubes. Rules and TurboIntegrator scripts let a model builder encode almost any driver logic, and results recalculate on demand. The SaaS is sized by RAM and named users: Essentials (16 GB, 5 users) lists at $10,500 a year on IBM's own store and $9,900 a year on AWS Marketplace, and Standard (32 GB, 10 users) at $19,800 a year on AWS Marketplace (published, checked 30 September and 1 October 2026), with Premium (64 GB, 20 users) by quote. On-premises licenses, subscription or perpetual, are also quote only.
The catch is that TM1 is a toolkit. The model is only as good as the developer who writes the rules, feeders and TurboIntegrator processes, and most customers hire a partner to do it; PeerSpot's review summary puts deployments anywhere from one month to a year. Native data connectivity is thin: SAP is the one documented native connector and it is licensed separately per SAP instance, while relational sources such as Snowflake reach the SaaS through IBM's ODBC Data Connector, which your IT team installs and runs inside your own network. Essentials also leaves out auditing and logging, high availability and separate development and production environments.
Before signing, get in writing the RAM and user counts your model needs over the term, the price of the PA Agent add-on that the MCP server requires, the SAP connector license if you run SAP, and a cap on renewal uplifts (a Vendr community buyer could not remove a 5 percent uplift when renewing through a reseller). Ask which SOC report IBM will provide for your deployment, and get the partner's build estimate as a fixed fee.
- Enterprises with large multidimensional driver models (many products, cost centers or SKUs) that have outgrown Excel and need on-demand recalculation
- Finance teams that want to keep Excel as the front end through PAfE while one governed model holds the data and security
- SAP shops and existing TM1 customers who want governed AI agents working on the same cubes
- Mid-market teams on NetSuite, Sage Intacct or QuickBooks Online that want click-to-connect actuals with no scripting
- Teams with no budget for a TM1 partner or an in-house modeler who knows rules, feeders and TurboIntegrator
- Buyers who want statutory consolidation in the same product; IBM sells that separately as IBM Controller
The jobs a finance team actually hires IBM Planning Analytics for
Jobs, mapped to how finance teams actually buy.
Run budgets and forecasts in one live model
Contributors enter numbers in Excel or the web, and TM1 consolidates them on demand so there is no spreadsheet roll-up. IBM's FP&A page quotes a Salesforce engineering lead saying cycles went from days or weeks to a few hours (vendor-claimed, checked 30 September 2026).
Plan headcount and compensation with HR
IBM's workforce page covers headcount, compensation, location and attrition scenarios tied to the budget. It quotes Volkswagen Group Services' HR controller saying report generation fell from 7 days to a few seconds; that figure is vendor-claimed.
Generate baseline forecasts from actuals
AI forecasting runs scheduled baseline and on-demand time-series forecasts with outlier correction, then spreads results down the hierarchy; it is included from the Standard plan. IBM credits Idaho Forest Group with cutting forecasting effort from over 80 hours to under 15 (vendor-claimed).
Plan in TM1 and write the numbers back to SAP
The Planning Analytics Connector for SAP pulls actuals over SAP OData gateways and sends planned values back after budgeting, without persisting data in the connector (IBM docs). It is licensed per SAP instance and is not supported on Planning Analytics Advanced Certified Containers.
Let an agent chase budget variances
IBM's workshop repo (updated 15 September 2026) wires the Planning Analytics MCP server into watsonx Orchestrate agents that flag variances over $100K, pull CRM and ERP context, and draft a CFO-ready explanation. It is a lab built on demo data, so treat it as a pattern to adapt.
Watch the same task in each tool
Forget the 40-row feature grid. Pick a finance workflow and see how this tool and its closest peers actually do it, with our verdict on who wins for whom.
Compare on what matters
What it actually costs
Here's what it really costs: published rates where a vendor lists them, aggregated quote data where they don't, plus a path to your own number.
5 named users, 16 GB RAM, 512 GB storage, prebuilt application, core model building, self-service data import, and either the Excel or the web interface. IBM's store bills the 12-month term at $10,500 up front or $875 a month; IBM's AWS Marketplace listing charges $9,900 for 12 months.
10 users, 32 GB RAM, both Excel and web, self-service backup, dev and production environments with source control, high availability, auditing and logging, AI forecasting.
20 users and 64 GB RAM with everything in Standard plus large-scale AI-based forecasting; aimed at large enterprises.
On-premises on Windows or Linux, or hybrid on IBM Cloud Pak for Data; subscription or perpetual licensing.
AI agent add-on, formerly the PA Assistant; IBM's MCP catalog says the MCP server requires it. The only public price is an IBM community blog of 31 July 2025, so confirm the current price in the quote.
Read and write connector for SAP BW, S/4HANA and HANA Cloud, licensed for a set number of SAP instances via Passport Advantage.
Trial of the SaaS with prebuilt planning and financial samples across use cases.
What it actually takes to go live
IBM's pricing table (checked 30 September 2026) marks auditing and logging, high availability, development and production environments with source control, self-service backup and AI-based forecasting as Standard and Premium features only. Essentials also limits you to one interface, Excel or web. A controller who needs an audit trail on model changes should price Standard at $19,800 a year (AWS Marketplace) as the real floor.
PeerSpot's review summary (checked 30 September 2026) says setup ranges from straightforward to months of work with skilled consultants, and puts deployments from one month to a year. One IBM partner's UK government rates run 1,200 to 1,600 GBP a day for architect roles (G-Cloud 15). Budget for ongoing developer time as models change, on top of the first build.
The SAP connector reads and writes BW, S/4HANA and HANA Cloud, but it is licensed per SAP instance (IBM docs). On the SaaS, which runs TM1 12, ODBC sources such as Snowflake go through IBM's ODBC Data Connector, which your IT team installs on a machine in your own network and exposes to the SaaS over HTTPS with a valid certificate (IBM docs and IBM onboarding guide, checked 1 October 2026). No native NetSuite, Sage Intacct, QuickBooks Online, Xero, Business Central or Workday connector was found on 30 September 2026; partner tools such as Cubewise Atmosphere and IBM App Connect fill the gap at extra cost.
IBM extended 2.0.9 general support to 31 October 2025 and sold extended support to 31 October 2026 (IBM community post, 6 December 2024). The 2.1 release retires TM1 Architect, Perspectives and Performance Modeler, according to partner Converge (30 April 2025), so old client tools have to be replaced. IBM's December 2024 post said 2.1.x would be supported to 30 April 2027, while IBM's lifecycle page, last updated 12 June 2024, still shows no end date for 2.1 (checked 1 October 2026).
IBM's GitHub MCP catalog says the server requires the PA Agent add-on entitlement, and IBM's 26 April 2026 blog ties all MCP tool access to that license. The tools launched on the SaaS in 3.1.4 (27 January 2026) and reached Local and on Cloud with Workspace 2.1.20, so older on-premises installs must upgrade first. Release 3.1.9 (25 June 2026) replaced the separate cube and analysis endpoints with /ibm-pa-tools and said the old ones are no longer available. Pin the endpoint and key rotation in your runbook before you build workflows on it.
A Vendr community buyer (3 January 2024) renewing through the reseller Cubewise could not get a 5 percent uplift removed. IBM's AWS Marketplace listing says all orders are non-cancellable and all fees non-refundable (checked 30 September 2026). Negotiate a renewal cap and a right to step down a plan before the first term ends.
Scored on what an outside agent can actually reach: the documented API, how it authenticates, and whether the vendor ships an MCP server.
Exists: yes. IBM documents an in-product MCP server for Planning Analytics. Custom agents over MCP arrived for Planning Analytics as a Service in release 3.1.4 (27 January 2026), and IBM's 26 April 2026 blog says releases 3.1.7 and 2.1.20 extended the tools to Planning Analytics Local, on Cloud and Certified Containers. Release 3.1.9 (25 June 2026) merged the separate cube and analysis endpoints into one /ibm-pa-tools endpoint, retiring the old ones; Local and on Cloud reach it from Workspace 2.1.22. IBM's GitHub MCP catalog lists it and notes it requires the PA Agent feature add-on entitlement. It was not in the public MCP registry on 1 October 2026. A community server, flameY3T1/tm1-mcp-server, wraps the TM1 REST API with 113 tools; it is not an IBM product.
Scope: read and write. IBM's workshop repo (IBM/pa-bob-agents, updated 15 September 2026) lists read tools such as get_cube_dimensions, execute_mdx_and_get_view, get_data_from_data_explorer, perform_outlier_detection and perform_impact_analysis, and write tools: save_mdx_view, create_tm1_cube, create_tm1_process, update_tm1_process, delete_tm1_process and execute_tm1_processes_asynchronously. IBM's docs add AI-generated MDX and scenario management inside MCP workflows. Data changes go through TurboIntegrator processes. The workshop's tool list has nothing for plan approvals or security changes, but IBM's 26 April 2026 blog announces Workflow Agent tools for contribution, review and approval of plans, so confirm which of those have shipped.
Turnkey: no. An administrator enables the agent features and hands each client an endpoint plus credentials: https://PAW_HOST/api/TENANT_ID/v0/agentic-ai/ibm-pa-tools/mcp on the SaaS, or https://PAW_HOST/api/v0/agentic-ai/ibm-pa-tools/mcp on Local and on Cloud (IBM blog, 26 June 2026). IBM's 3.1.4 note gives a username (a watsonx Orchestrate API key) and a password (an MCP token), and IBM's workshop sends a username and password as HTTP Basic auth, so both describe Basic credentials; IBM added OAuth with releases 3.1.7 and 2.1.20 (IBM blog, 26 April 2026). Local and on Cloud customers need Workspace 2.1.20 or later and the PA Agent license.
Inside the product, the Planning Analytics Agent (renamed from the AI assistant) answers natural-language questions, explains drivers and variances, and builds exploration reports; IBM says it follows existing permissions. AI-based forecasting is included from Standard, with large-scale forecasting in Premium. IBM's case studies, such as Idaho Forest Group cutting forecasting effort from over 80 hours to under 15, are vendor-claimed. Since 3.1.11 (17 September 2026) admins can choose Amazon Bedrock as the model provider for AI insights.
Build a driver-based budget and reforecast
Who wins for whom: Teams with a TM1 developer who want rule-based driver models and Excel write-back: IBM Planning Analytics. Enterprises planning across finance, workforce, sales and supply chain with prebuilt apps: Anaplan. Teams that want prebuilt Financials, Workforce and Capital modules linking the three statements: Oracle Cloud EPM.
IBM Planning Analytics
TM1 rules carry the drivers. TM1 cubes with rules carry driver-based models, what-if scenarios and plan versions, AI forecasting with baseline and spreading starts at the Standard plan, and PAfE gives Excel users write-back into the same model.
Where it loses: Rules, feeders and TurboIntegrator are code that most customers hire a partner to write, and PeerSpot's summary puts deployments anywhere from one month to a year.
- Modeling depth: 88/100
- Time to live: 1 month to 1 year (reported)
Source: AF score rationale: Modeling depth 88; IBM Planning Analytics watch-outs.
Watch: IBM Planning Analytics overview 4:34 · Consultant walkthrough · Octane Software Solutions on YouTube
Anaplan
Three statements in one model. Anaplan plans revenue, opex and capex across all three statements in its Integrated Financial Planning app, and the same platform carries workforce, consolidation, sales and supply chain models with scenario modeling and Forecaster ML forecasts.
Where it loses: Builds are mostly partner-led, Vendr says services often cost 50 to 150 percent of first-year subscription, and a workspace's Polaris or Classic engine cannot be changed later.
- Modeling depth: 91/100
- Time to live: Few months to 1 year+ (Vendr)
Source: AF score rationale: Modeling depth 91; Anaplan use cases, implementation note and watch-outs.
Watch: Anaplan Planning 101 How to Plan, Budget & Forecast (+Demo) 1:00:15 · Official demo · Business Solution on YouTube
Oracle Cloud EPM
Prebuilt modules link three statements. EPM Standard ships prebuilt Financials, Workforce, Capital, Projects and Strategic Modeling modules, and Oracle says income statement changes roll to the balance sheet and cash flow; Enterprise adds custom cubes, Groovy and Monte Carlo simulation.
Where it loses: EPM Standard limits Planning to one input and one reporting cube, so custom cubes, Groovy logic and most AI features need EPM Enterprise at $500 per user a month with a 25-user minimum.
- Modeling depth: 87/100
- Time to live: 6 to 26 weeks (partner)
Source: AF score rationale: Modeling depth 87; Oracle Cloud EPM use cases, pricing tiers and watch-outs.
Watch: Overview: Welcome to Financial Planning 3:27 · Official demo · Oracle EPM Tutorials on YouTube (official)
Pull ERP actuals into the plan
Who wins for whom: SAP BW or S/4HANA shops that want actuals read and plans written back by one connector: IBM Planning Analytics. Enterprises pulling from SAP, Salesforce, Snowflake and Databricks through Data Orchestrator: Anaplan. Oracle Cloud ERP, EBS or PeopleSoft shops that want budgets and journals written back: Oracle Cloud EPM.
IBM Planning Analytics
Native connector stops at SAP. The Planning Analytics Connector for SAP pulls actuals from BW, S/4HANA and HANA Cloud over OData and writes planned values back after budgeting, while other sources load through TurboIntegrator over ODBC or REST.
Where it loses: No native NetSuite, Sage Intacct, QuickBooks Online or Workday connector was found, the SAP connector is licensed per SAP instance, and SaaS ODBC sources need a connector your IT team hosts.
- ERP & data connectivity: 62/100
- Time to live: 1 month to 1 year (reported)
Source: AF score rationale: ERP & data connectivity 62; IBM Planning Analytics use cases and watch-outs.
Watch: What is IBM Planning Analytics? #paw #analytics #ibm #business 0:44 · Consultant walkthrough · Cubewise on YouTube
Anaplan
SAP and warehouse connectors built in. Anaplan's Data Orchestrator lists 14 import connectors, including SAP HANA, BW and ECC over OData v4, Salesforce, Snowflake, BigQuery, Databricks and Oracle Database, with writeback pipelines to Snowflake, Databricks, Oracle and cloud storage.
Where it loses: There is no native NetSuite, Workday HCM or Sage Intacct connector, so those systems arrive through files, CloudWorks or an integration platform, and Data Orchestrator is sold in capacity packages.
- ERP & data connectivity: 76/100
- Time to live: Few months to 1 year+ (Vendr)
Source: AF score rationale: ERP & data connectivity 76; Anaplan use cases and watch-outs.
Watch: Introduction to Anaplan 17:16 · Consultant walkthrough · Source Corp on YouTube
Oracle Cloud EPM
Fusion actuals in, budgets back. Data Integration has prebuilt sources for Oracle Cloud ERP, which loads actuals and takes budgets and journals back, plus Oracle HCM, NetSuite through Oracle SuiteApps, and EBS and PeopleSoft through the EPM Integration Agent.
Where it loses: SAP data arrives by file, agent JDBC, scripting or the Oracle Integration Cloud SAP adapter, and no native connector is listed for Workday, Salesforce, Snowflake, Sage Intacct or Xero.
- ERP & data connectivity: 72/100
- Time to live: 6 to 26 weeks (partner)
Source: AF score rationale: ERP & data connectivity 72; Oracle Cloud EPM comparison row.
Watch: Planning AI Agent in Oracle Cloud EPM: Demo 6:01 · Official demo · Oracle on YouTube (official)
IBM Planning Analytics, kept current
Release 3.1.11 added Workspace REST endpoints to list, start, stop and restart TM1 databases and to create, delete and back up TM1 12 databases. Admins can also pick Amazon Bedrock as the AI model provider.
Release 3.1.9 consolidated all MCP tools behind /ibm-pa-tools and retired the earlier endpoints, so existing integrations had to be repointed. The same release made watsonx Orchestrate the in-product chat assistant.
Release 3.1.4 let customers build their own agents against Planning Analytics as a Service using MCP tools that list TM1 objects, run MDX, return views and run TurboIntegrator scripts.
IBM ended general support for on-premises 2.0.9 on 31 October 2025, with paid extended support to 31 October 2026, pushing customers to 2.1.x.
IBM App Connect Enterprise added a Planning Analytics connector to update models from sales or operational data and trigger workflows on forecast thresholds. App Connect is licensed separately.
IBM said the July 2025 Workspace release let the assistant run either IBM-managed or on a customer's own watsonx Orchestrate and watsonx.ai, opening it to custom agentic workflows. The post put the license at 30 dollars per user per month (reported, July 2025).
IBM Planning Analytics FAQ
IBM's own store lists Essentials at $10,500 a year for 5 users and 16 GB of RAM, payable as $875 a month over the 12-month term (checked 1 October 2026). On AWS Marketplace, IBM sells Essentials at $9,900 a year and Standard at $19,800 a year for 10 users and 32 GB (checked 30 September 2026). Premium, at 20 users and 64 GB, is quote only, as is any larger size.
Most enterprise FP&A teams need Standard or above, because Essentials lacks auditing and logging, high availability and separate development and production environments. Add the PA Agent add-on if you want the MCP server (reported at 30 dollars per user per month in an IBM community blog of 31 July 2025; confirm the current price), the SAP connector if you run SAP, and partner build costs. Vendr's public IBM page shows only an all-IBM median of $97,952 a year across 96 purchases, which is not a Planning Analytics figure.
The AWS Marketplace plans run on 12-month contracts, and IBM's listing says orders are non-cancellable and fees non-refundable. Direct IBM and reseller deals are negotiated; a Vendr community buyer reported a 5 percent renewal uplift they could not remove.
Data and model objects are reachable through the TM1 REST API, which IBM describes as an OData Version 4 interface to query and update TM1 models, and TurboIntegrator processes can export cubes and dimensions to files. Release 3.1.11 added REST endpoints to manage TM1 12 database backups. Ask IBM in writing how long it keeps data after termination; the AWS listing does not say.
IBM does not publish a standard timeline. PeerSpot's review summary reports deployments from one month to a year, depending on model complexity and data integration (checked 30 September 2026).
Most builds are partner-led by IBM business partners such as Cubewise, Octane, Aramar, QueBIT, PMsquare or Lodestar, because the work is TM1 rules, feeders, TurboIntegrator and MDX. The SaaS removes server installs, and IBM offers a 30-day trial with prebuilt samples, but your FP&A lead still owns the model design and IT still owns source extracts.
Yes, with the PA Agent add-on. IBM ships an in-product MCP server at an /ibm-pa-tools endpoint that any MCP client can call with admin-issued credentials, on the SaaS and, according to IBM's 26 April 2026 blog, on Local, on Cloud and Certified Containers since releases 3.1.7 and 2.1.20. IBM's workshop repo lists tools to explore cubes, run MDX, detect outliers, explain drivers, save views, create cubes and create, update and run TurboIntegrator processes.
It requires the PA Agent add-on entitlement, and Local or on Cloud customers need Workspace 2.1.20 or later (2.1.22 or later for the /ibm-pa-tools endpoint). Older Local installs can still reach the TM1 REST API through a community server such as flameY3T1/tm1-mcp-server, which is not an IBM product. Treat any write tool as production access and scope the credentials accordingly.
IBM Planning Analytics alternatives
Ask us anything
Answered only from our own published research on this tool, never from general internet noise. If we cannot answer it well, our research agents will dig in and publish a sourced answer.