Review of Pigment, Enterprise Planning Software Vendor

By Léon Levinas-Ménard
Last updated: December, 2025

Go back to Market Research

Pigment is a Paris-founded, venture-backed planning and performance management (EPM) SaaS positioned as a single modeling layer for finance and operational planning: users build multidimensional models (dimensions/metrics/“blocks”), write formulas that automatically recompute dependent blocks, define access controls, and publish reports and scenarios to stakeholders. Beyond FP&A and consolidation-style use cases, Pigment markets templated applications for S&OP and demand & inventory planning, and provides forecasting capabilities via built-in statistical functions (e.g., ETS-style time-series forecasting) plus a separate “Predictions” feature described as machine-learning-based forecasting with support for external factors. Pigment also ships AI-facing features (e.g., an “Analyst Agent” and related AI pages), but public technical detail varies by feature: some core modeling mechanics are documented precisely (dependency checks, circular-reference mitigation via PREVIOUS/PREVIOUSBASE, etc.), while ML/AI components are partly explained at a systems level (e.g., a horizontally scalable Predictions architecture using Dask) without consistently disclosing model classes, training/evaluation procedures, or reproducible benchmarks.

Pigment overview

Pigment sells a browser-based planning platform centered on a shared semantic model for planning, simulation, and reporting across multiple functions (finance, sales, HR, supply chain). The vendor emphasizes replacing spreadsheet-centric planning with a centralized model that supports collaboration, permissions, auditability, and scenario analysis.12 In supply-chain-adjacent positioning, Pigment publishes dedicated pages for supply chain planning, S&OP, and demand & inventory planning, framing the product as a way to connect demand, inventory, and capacity plans with financial outcomes inside one planning workspace.345

From the available documentation, Pigment’s “modeling engine” is formula-driven: when formulas change, dependent blocks are recalculated, and the platform actively prevents or manages circular dependencies using specific functions (e.g., PREVIOUS for within-block iteration; PREVIOUSBASE for multi-block iterative configurations), with explicit constraints such as limits on “allowed metrics” and documented performance implications.6 This is concrete evidence that Pigment is not “just CRUD + dashboards”: the core system is a dependency-tracked computation engine for multidimensional planning models.

Pigment vs Lokad

Pigment and Lokad address “planning” from fundamentally different starting points. Pigment is primarily an enterprise planning and performance management platform: its core deliverable is an interactive, organization-wide modeling layer (dimensions/metrics, formulas, workflows, permissions, reports) that enables humans to build plans, run scenarios, and align stakeholders—optionally augmented by forecasting tools (statistical functions and ML-oriented “Predictions”) and AI-driven querying/assistance.3467 Lokad, by contrast, is built as a quantitative supply chain optimization platform: it centers on producing decision recommendations (e.g., purchasing, inventory allocation, production planning) under uncertainty using probabilistic forecasting and explicit optimization objectives, implemented programmatically through its Envision DSL and executed as repeatable computational pipelines.8910

This difference matters technically. Pigment’s public materials emphasize model flexibility, collaboration, and speed of scenario iteration, with some forecasting and AI assistance, but provide limited public evidence that the platform routinely solves large constrained combinatorial optimization problems (e.g., mixed-integer optimization for supply chain decisions) as a first-class product outcome; Pigment’s supply chain pages are framed around connecting plans and improving responsiveness rather than proving an optimizer’s architecture, objective functions, and constraint handling at scale.345 Lokad’s public technical narrative, by comparison, foregrounds “probabilistic forecasts → optimized decisions” as the product’s core output, including published descriptions of probabilistic/quantile forecasting and optimization-oriented methods, and evidence of forecasting-oriented technical work (e.g., participation in the M5 forecasting competition).118

Commercially, Pigment resembles a high-growth EPM/IBP vendor: multiple large VC rounds and a broad cross-functional planning positioning.1213 Lokad resembles a specialized supply chain optimization vendor that emphasizes bespoke, code-defined optimization apps rather than a generalized enterprise planning layer.89

Company history, funding, and milestones

Pigment was founded in 2019 (Paris) by Éléonore Crespo and Romain Niccoli, positioning early on as a modern alternative to legacy EPM/planning tooling.214 Public reporting documents a sequence of venture rounds: a Series A reported in 2020, a Series C reported in 2023, and a large Series D reported in 2024, indicating sustained investor appetite and material commercial scaling.14151213

No acquisition activity (Pigment acquiring others or being acquired) was identified in the public sources reviewed for this report. This absence should be treated cautiously: it may reflect “no acquisitions,” or it may reflect a lack of publicly indexed disclosures.

Product and technical capabilities

Planning model: multidimensional blocks, formulas, and recalculation

Pigment’s documentation provides unusually specific detail on how its modeling engine behaves when formulas create dependency cycles. The platform recalculates dependent blocks when formulas change and performs checks to detect circular dependencies; it provides mechanisms to express iterative computations (time-shifted dependencies) without creating infinite loops, such as PREVIOUSBASE for multi-block iterative configurations.6 The same documentation describes how Pigment consolidates formulas from a dependency cycle into a “single base formula” for computation, and warns about performance impact and debugging failure modes when the merged base formula becomes invalid.6

This is strong evidence that Pigment’s core value is a computation engine coupled to a multidimensional planning model, not merely a workflow wrapper around spreadsheets.

Supply chain-oriented templates and use cases

Pigment maintains dedicated supply-chain pages describing S&OP and demand/inventory planning use cases (including “resilient plans,” scenario simulation, and aligning supply chain with finance).345 A concrete customer story claims Danone used Pigment for S&OP and long-range demand planning, with an “implementation time” stated as three months, and describes replacement of Excel-heavy processes with a customized Pigment model built with an implementation partner.7 While still marketing material, this is at least a named, attributable reference with implementation details and declared scope.

Forecasting, ML, and “Predictions”

Pigment exposes statistical forecasting functions in documentation/community materials (e.g., ETS forecasting functions).1617 Separately, Pigment positions “Predictions” as an ML-oriented feature for forecasting, including “Choose prediction model” documentation that describes configuration options but does not, in the public page, specify exact algorithm families, training methodology, validation metrics, or reproducible benchmarks.18

Pigment’s engineering blog provides the most substantive technical evidence for the Predictions subsystem. In a post on scaling predictions, Pigment describes an architecture using a Dask cluster to run many forecasts in parallel and states that “each time series is forecasted independently,” enabling horizontal scaling.19 This supports the claim that Pigment’s forecasting is implemented as a genuine compute pipeline (not a superficial UI label), but it still leaves key ML questions unanswered publicly: which models are used, how feature engineering is handled (beyond “external factors” claims), how drift is monitored, and what accuracy/cost tradeoffs are achieved in production.

AI assistants and “agentic” components

Pigment markets AI-facing capabilities (e.g., “Pigment AI,” “Analyst Agent Overview”) and has engineering content describing an “Insights Assistant” and “agentic AI,” including references to common LLM-agent frameworks (e.g., LangGraph) and system design considerations.2021 These sources substantiate that Pigment is implementing AI assistant functionality as product features. However, from a skeptical technical standpoint, these materials are better interpreted as UX-layer augmentation (querying, summarization, guided actions) rather than evidence that core planning decisions are optimized end-to-end by AI; public documentation does not provide reproducible demonstrations that the AI layer reliably produces planning decisions under hard constraints.

Architecture and technology stack evidence

Pigment does not publicly publish a single, canonical architecture whitepaper in the sources reviewed here. However, multiple independent signals converge on a plausible stack:

  • A “tech” profile on Welcome to the Jungle lists core technologies including PostgreSQL, .NET (C#), React, D3.js, TypeScript, Google Cloud Platform, Kubernetes (GKE), and CircleCI.22
  • Pigment’s engineering blog indicates use of Python ecosystem tooling (e.g., Dask) to scale forecasting workloads.19
  • Pigment’s product documentation demonstrates a custom modeling engine capable of dependency analysis and iterative computation management (PREVIOUS/PREVIOUSBASE), implying a non-trivial computation runtime integrated with the planning model.6

Taken together, the public evidence supports that Pigment is a cloud-native SaaS with a modern web frontend, a .NET-based backend core, and specialized compute subsystems for forecasting.

Deployment, rollout, and operational assurances

Pigment’s case study content provides an example rollout claim: Danone’s implementation is described as going from research/build/testing/rollout in “just three months,” involving an implementation partner and repeated model iteration (multiple versions tested in a day).7 This suggests a deployment pattern closer to “solution modeling + change management” than “install software and go,” even though Pigment is delivered as SaaS.

On security and compliance, Pigment’s security page claims SOC 2 Type 2 and SOC 1 Type 2 compliance, plus GDPR/CCPA references; it also describes enterprise identity integration (SAMLv2 SSO, SCIM provisioning, MFA), encryption in transit (TLS 1.3, HSTS) and at rest (AES-256), plus stated RTO/RPO targets and a security assurance program (penetration tests, bug bounty, audits, vulnerability scanning).23 These are specific operational claims, though independent certification details should be verified via Pigment’s referenced “trust report” if used for procurement.23

Clients, references, and case studies

Publicly named customer references exist and are not purely anonymous. Pigment publishes named customer stories (e.g., Danone), including stated use cases and an implementation timeframe.7 Third-party reporting also references Pigment’s traction and customer adoption in general terms alongside funding announcements.1213

Still, readers should distinguish:

  • Named, attributable references (e.g., a dedicated customer story page naming the company and describing scope/timeline).7
  • Logo walls / generalized adoption claims that may not specify which modules, geographies, or maturity of rollout.1213

Skeptical technical assessment

What Pigment delivers (precisely): a cloud planning computation environment where organizations define a multidimensional planning model (data structures + formulas), run scenario simulations, collaborate on plan inputs with permissions and auditability, and produce reports; optionally, they can add forecasting via statistical functions and the “Predictions” pipeline, and use AI-assisted querying/insights features.3461923

How it achieves those outcomes: evidence supports (1) a dependency-tracked formula engine with mechanisms to resolve iterative/circular calculations,6 (2) a cloud SaaS architecture with common enterprise identity and encryption controls,23 and (3) a scalable forecasting subsystem using distributed compute (Dask) for parallel time-series forecasting.19 The publicly verifiable architecture details are strongest for the modeling engine mechanics and weaker for the exact ML/AI algorithms used.

State-of-the-art evaluation: Pigment appears technically modern in platform engineering terms (cloud-native stack, strong security posture claims, distributed compute for forecasting).192223 However, many “AI” aspects remain hard to validate externally because Pigment does not consistently publish model classes, evaluation protocols, or reproducible benchmarks for Predictions/AI features. From a skeptical standpoint, the substantiated innovation is the planning computation layer and operational productization; the least substantiated claims are those implying advanced AI-driven decision-making beyond forecasting and assistance.

Commercial maturity: Pigment’s multiple late-stage venture rounds (including a large Series D) and named enterprise customer stories indicate a commercially established scale-up rather than an early-stage product experiment.71213

Conclusion

Pigment is best evidenced as a cross-functional enterprise planning platform with a real computation core: the documentation around dependency management, iterative calculations, and formula recomputation provides concrete proof of a purpose-built modeling engine rather than a thin UI around spreadsheets.6 Pigment’s supply chain positioning (S&OP, demand & inventory planning) is supported by dedicated product pages and at least one named story (Danone) that includes scope and a stated implementation timeline.3457 Machine learning and AI claims are partially substantiated—most strongly by engineering disclosure of a distributed forecasting architecture (Dask-based parallelism), and by engineering discussion of AI assistant design—but public documentation remains insufficient to independently verify algorithmic specifics, accuracy claims, or the reliability of AI-driven decisioning in constrained planning contexts.192021

Against Lokad, Pigment reads as an enterprise planning/modeling layer with forecasting and AI assistance, whereas Lokad reads as a supply chain optimization platform whose core deliverable is prescriptive decision optimization under uncertainty through programmatic modeling.891011

Sources

  1. Pigment — Supply Chain Planning (SCP) Software — retrieved Dec 17, 2025 ↩︎

  2. TechCrunch — Pigment raises $25.9M to take on spreadsheets in business planning — Dec 2, 2020 ↩︎ ↩︎

  3. Pigment — Sales & Operations Planning (S&OP) Software — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  4. Pigment — Demand Planning & Inventory Planning Software — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  5. Pigment — Supply chain teams (solutions navigation) — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎

  6. Pigment Knowledge Base — Iterative Calculations across multiple Blocks using PREVIOUSBASE — updated Dec 2, 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  7. Pigment Customer Story — Danone chooses Pigment to nurture demand planning maturity — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  8. Lokad — Quantitative Supply Chain (overview) — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎

  9. Lokad Docs — Envision language (DSL) — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎

  10. Lokad — Stochastic Discrete Descent (SDD) — retrieved Dec 17, 2025 ↩︎ ↩︎

  11. Lokad Blog — Ranked 6th out of 909 teams in the M5 forecasting competition — Jul 2, 2020 ↩︎ ↩︎

  12. Tech.eu — Pigment raises €133M ($145M) Series D — Apr 23, 2024 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  13. TechCrunch — Pigment raises $145M Series D — Apr 23, 2024 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  14. TechCrunch — Pigment raises $25.9M (founders/background) — Dec 2, 2020 ↩︎ ↩︎

  15. TechCrunch — Pigment raises $88M Series C (planning software) — Apr 26, 2023 ↩︎

  16. Pigment Community — FORECAST.ETS function (ETS forecasting) — retrieved Dec 17, 2025 ↩︎

  17. Pigment Community — FORECAST.LINEAR function — retrieved Dec 17, 2025 ↩︎

  18. Pigment Knowledge Base — Choose prediction model — retrieved Dec 17, 2025 ↩︎

  19. Pigment Engineering — Scaling predictions using Dask — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎ ↩︎

  20. Pigment Engineering — The road to agentic AI: building an Insights Assistant — retrieved Dec 17, 2025 ↩︎ ↩︎

  21. Pigment — Analyst Agent Overview — retrieved Dec 17, 2025 ↩︎ ↩︎

  22. Welcome to the Jungle — Pigment tech stack (PostgreSQL, .NET, React, GCP, Kubernetes, etc.) — retrieved Dec 17, 2025 ↩︎ ↩︎

  23. Pigment — Security (SOC2 Type 2, encryption, SSO/SCIM, RTO/RPO) — retrieved Dec 17, 2025 ↩︎ ↩︎ ↩︎ ↩︎ ↩︎