Kindred replaces months of manual model-building with an AI that reads your data, asks the right questions, and generates a governed planning model — automatically.
The Problem
Anaplan. Oracle EPM. Workday Adaptive. Every major planning platform shares the same flaw: a human expert must manually define every dimension, hierarchy, formula, and driver. That takes months to implement, costs a fortune to maintain, and breaks whenever the business changes.
The result is that FP&A teams spend 80% of their time managing the model and 20% actually planning. Kindred inverts that ratio.
The Solution
Kindred connects to your existing data sources, automatically interprets field names and data structures, constructs a governed semantic layer, and runs a guided interview to resolve any ambiguity. The result is a production-ready planning model in days, not months.
Because the model is AI-native, it continuously updates itself based on actual versus forecast variances — getting more accurate with every planning cycle without any manual intervention.
Capabilities
Seven interconnected AI modules that replace the entire traditional planning stack.
Connects to any source system and maps fields to a canonical planning model automatically. No manual field matching or ETL scripting required.
When the AI detects ambiguous data, it asks targeted MECE questions to resolve conflicts — capturing institutional knowledge that never makes it into the model otherwise.
Automatically identifies which variables materially impact business performance and focuses modeling depth where it matters most.
Converts planning assumptions into a live causal graph linking financial and operational variables — Revenue = Price × Volume, built and maintained automatically.
Generates multi-dimensional rolling forecasts with scenario comparison, cell-level overrides, and variance analysis — all at the grain you define.
Ingests actuals automatically, decomposes variances by driver, and updates model assumptions to improve forecast accuracy over every cycle.
A growing library of codified use-case models — Revenue Planning, Workforce Planning, and more — each encoding the financial rules that FP&A teams rely on.
Automatically generates planning tasks, approval routes, and deadlines based on the driver model — adapting instantly when assumptions change.
Every mapping decision, interview answer, and model change is logged with full provenance. Enterprise-grade data residency and access control built in.
How It Works
Link your ERP, CRM, HCM, or upload flat files. Kindred supports SQL, REST API, Databricks, and flat-file ingestion.
AI profiles your data, maps fields to the canonical model, and asks targeted questions to close any gaps.
Configure your planning grain and horizon. Kindred generates a driver-based rolling forecast with scenario comparison.
Actuals flow in automatically. The model updates its assumptions and improves accuracy with every cycle.
The IP
Codified financial logic for every enterprise planning use case — the rules that LLMs alone cannot reliably enforce.
Customer, product, region, channel dimensions. Recognizes additive vs. non-additive measures. Links CRM pipeline to GL actuals automatically.
Available NowHeadcount (semi-additive, point-in-time), comp spend, benefit load. Filters GL to Employee Detail rows. Joins Workday/BambooHR on employee ID.
Available NowCampaign, channel, segment dimensions. CAC as a non-additive ratio measure. Pipeline contribution and ROI tracking.
Coming SoonSKU, location, supplier dimensions. Inventory as semi-additive. Lead time as non-additive. Integrates ERP and WMS sources.
Coming SoonARR, MRR, churn, NRR metrics. Cohort and customer tier dimensions. Integrates Stripe, Zuora, Salesforce automatically.
Coming SoonOEE, yield, throughput, scrap rate. Machine hours × output logic. BOM cost structure interpretation.
RoadmapWhy Kindred
Compare what matters for enterprise planning teams.
Join the private beta. Bring your ERP, CRM, or HCM files and watch Kindred map them in minutes.
The Platform
Every other AI planning tool wraps an LLM around a traditional model builder. Kindred is different: a deterministic rules engine — the Ontology Library — does the financial logic. The LLM handles what it's actually good at: interpreting language, resolving ambiguity, and talking to humans.
Architecture
The LLM layer handles column name interpretation, free-form user instructions, and natural-language explanation. It proposes — in structured JSON — what it believes the mapping should be.
The Rules Engine validates every proposal against the Ontology Library. It enforces filter rules, aggregation types, join topology, and gap detection. It decides. The numbers are always deterministic, auditable, and correct.
Integrations
Pre-built connectors for the most common enterprise source systems — with a generic adapter for anything else.
NetSuite, SAP, Oracle Financials, Microsoft Dynamics. Handles cryptic field names, chart of accounts variance, multi-entity consolidation.
Salesforce, HubSpot, Dynamics 365. Maps pipeline stages to revenue timing, links opportunity IDs to GL reference numbers.
Workday, BambooHR, ADP. Handles snapshot vs. event records, headcount deduplication, benefit load calculation.
CSV, Excel, PDF financial reports, EPM exports. Handles wide vs. long format, merged headers, multi-sheet workbooks.
Snowflake, Databricks, BigQuery, Redshift, Synapse. Direct SQL query with incremental load and partition-aware ingestion — any warehouse, any cloud.
Runs on AWS, Azure, or GCP. Deploys entirely within your cloud tenant for data residency and security compliance. No vendor lock-in, ever.
The live demo walks through data mapping, the interview loop, and a rolling forecast — end to end.
Walk through a complete planning cycle — from raw data to governed forecast — in under 10 minutes.
The full 10-step Kindred workflow — from use case selection through to a finalized, shareable rolling forecast. Click Load Sample Data in Step 2 to run the complete demo instantly, or upload your own CSV files.
Try It Yourself
The demo accepts any CSV. Try your own ERP export, Workday snapshot, or Salesforce opportunity report and watch Kindred map it in real time.
Select Revenue Planning, Workforce Planning, or enter a custom use case. This sets the canonical ontology the AI maps toward.
Upload 1-3 CSV files from different source systems, or load the built-in sample datasets (CRM, GL, EPM).
Review AI-generated mappings, complete the MECE interview, and refine via the free-form chat interface.
Confirm your gold dataset, select planning dimensions, configure the rolling forecast, override cells, and finalize.
Join the private beta and get a guided session with your own source systems.
Get In Touch
We're working with a small group of design partners to refine the platform against real enterprise data. If you're an FP&A leader tired of managing your planning model instead of running it, we'd like to hear from you.
We're at an early stage and deliberately working with a small number of design partners. Every conversation helps us make the platform better for the use cases that actually matter in enterprise FP&A.
rchawla80@gmail.com
Issaquah, WA · Remote-first
U.S. Provisional Patent Filed · Assignee: Kindred LLC
We respond within 1 business day.
Thank you for reaching out. We'll be in touch within one business day.