Review of KetteQ, Supply Chain Planning Software Vendor

By Léon Levinas-Ménard

Last updated: December, 2025

KetteQ is a US-based supply chain planning software vendor founded in 2018 that positions itself as an “adaptive” and “AI-first” alternative to legacy APS systems, built natively on Salesforce for the UX and on AWS for the compute-heavy back end.¹²³⁴ The company’s main commercial proposition is its PolymatiQ™ solver, described as a patent-pending “agentic AI engine” that runs thousands of planning scenarios across demand, inventory, production and service, automatically tuning parameters and continuously learning from market shifts to recommend more resilient plans.⁵¹⁶⁷ KetteQ targets mid-market and large enterprises across manufacturing, distribution and service, and cites brands such as Coca-Cola, Carrier, Johnson Controls, NCR Voyix and Alliance Consumer Group as customers.⁷⁸⁹ Since 2021 the company has raised roughly USD 30.9M in venture funding, culminating in a USD 20M Series B round in August 2025 led by Vocap Partners to accelerate global expansion and R&D on agentic AI.¹⁰¹¹¹²¹³¹⁴ Architecturally, KetteQ offers a multi-tenant SaaS platform that uses Salesforce as the transactional and collaboration layer while off-loading data- and compute-intensive work to AWS, with embedded workflow, notifications and analytics built primarily on SQL, Python and JSON rather than a proprietary DSL.¹²⁶⁴¹⁵ Technically, public information suggests a solver that orchestrates multi-pass simulation and scenario generation with machine learning-driven forecasting and parameter tuning, but without detailed disclosure of model classes, objective functions or optimization algorithms, and with no independent benchmarking comparable to forecasting competitions.⁵¹⁶³¹⁰ Commercially, the company is still relatively young: it has real customers, a modern stack and non-trivial funding, but it remains in the scale-up phase rather than a fully mature, globally entrenched planning incumbent.³¹⁰⁷⁸⁹ The sections below unpack KetteQ’s history, product and technology in more detail, then contrast its approach with Lokad’s quantitative supply chain platform using explicitly sourced material on both sides.¹⁶¹²¹⁷¹³¹⁸¹⁹²⁰

KetteQ overview

KetteQ is an Atlanta-based software company offering an “adaptive” supply chain planning and execution platform with a strong emphasis on AI-driven scenario planning and Salesforce integration.⁵¹²⁴ According to Sourcing Innovation, an independent industry blog, KetteQ was founded in 2018 and consciously built “in the modern age on a fully modern multi-tenant cloud-native SaaS stack,” learning from more than 100 previous supply chain planning implementations and two earlier supply chain companies created by its founders.³¹¹ KetteQ’s own materials present it as a cloud-native platform “built on the trusted foundation of Salesforce and AWS,” with architecture that combines Salesforce for user experience and collaboration with AWS-hosted analytics and solver components.¹²⁴¹⁵

The flagship capability is the PolymatiQ™ solver, described as a patent-pending, agentic AI engine that automatically runs and evaluates thousands of scenarios, tunes planning parameters and delivers “real-time adaptability” across demand, inventory, production and service.⁵¹⁶⁷⁸ The product portfolio is structured into supply chain planning (demand, inventory, supply planning) and supply chain execution (control tower, fulfillment, work orders, service parts), each presented as powered by PolymatiQ™ scenario analysis and AI/ML forecasting.⁵¹⁷²¹ KetteQ positions itself explicitly against “static” legacy planning systems built before 2010, arguing that older APS tools lack horizontal scalability, real-time integration and dynamic adaptability; the Sourcing Innovation review broadly supports the characterization of KetteQ’s platform as modern, multi-tenant and cloud-native.⁶³²²

On the commercial side, KetteQ lists reference customers and logos, including Coca-Cola, Carrier, Johnson Controls, NCR Voyix and others, and external funding-related press coverage echoes these names and cites metrics such as ~170% average annual CARR growth and a 100% implementation success rate.⁷⁸¹³¹⁴⁹ The most recent Series B round (USD 20M in August 2025) brings total disclosed funding to roughly USD 30.9M over three early-stage rounds, underscoring that KetteQ is past the seed stage but still scaling, not yet a long-established incumbent.¹⁰¹¹¹²²³¹³¹⁴

KetteQ vs Lokad

From a supply-chain-science perspective, KetteQ and Lokad occupy overlapping problem space (demand forecasting, inventory and supply planning, and broader supply chain decision-making) but with markedly different philosophies and architectures.

Programming model vs. configurable app. KetteQ is presented as a configurable, cloud-native application suite where planners interact primarily through Salesforce-based UIs, standard data models and scenario workbenches; extensibility relies on mainstream technologies (SQL, Python, JSON) rather than a domain-specific language.¹²⁶ Lokad, by contrast, is fundamentally a programmable platform built around its Envision DSL, a domain-specific language for supply chain predictive optimization that expresses all data transformations, probabilistic forecasts and optimization logic as code.¹⁷¹³¹⁸ Envision is tightly integrated with Lokad’s execution engine and columnar storage, and is explicitly designed to support probabilistic modeling and optimization over large supply chain datasets.¹⁷¹³ This means KetteQ leans toward a “configurable packaged app” model, whereas Lokad exposes a language-centric environment where bespoke decision logic is implemented as scripts.

Treatment of uncertainty. KetteQ’s materials emphasize “multi-pass probabilistic approaches” and thousands of scenarios but give limited detail on how probability distributions are represented or how uncertainty is mathematically propagated through the planning pipeline.⁵¹⁶³ By comparison, Lokad publicly documents a multi-generation evolution from classic point forecasts (2008) through quantile forecasts (2012), quantile grids (2015) and probabilistic forecasting (2016) to differentiable-programming-based forecasting and optimization (2019 onwards), explicitly centered on full probability distributions over demand and other uncertain variables.¹²²⁴²²²³¹⁵¹⁹ Lokad’s documentation describes an algebra of random variables embedded in Envision and probabilistic models that feed directly into its decision optimization algorithms.¹⁷¹³¹⁸¹⁹

Optimization & solver transparency. KetteQ’s PolymatiQ™ is positioned as a patent-pending agentic AI solver that runs thousands of scenarios, tunes parameters, and returns resilient plans, but public materials do not specify whether the underlying optimization is mathematical programming, metaheuristics, reinforcement learning or a hybrid, nor how objective functions and constraints are formally expressed.⁵¹⁶³¹⁰ Lokad, on the other hand, provides reasonably detailed descriptions of its optimization paradigms: stochastic discrete descent for inventory decisions under uncertainty, differentiable programming for joint forecasting/optimization, and domain-specific heuristics for combinatorial planning, all orchestrated through Envision and documented in both general and technical references.¹²¹⁷¹³¹⁸¹⁹²⁵ Lokad also references external validation of its forecasting and optimization stack via the M5 competition, where it ranked 6th overall out of 909 teams and 1st at SKU aggregation level, using probabilistic and differentiable-programming-based models.¹⁹¹⁴⁹²⁶

Decision outputs. KetteQ’s messaging stresses real-time planning, control tower visibility, and “agentic AI” running “thousands of scenarios to plan for every possibility,” but public examples focus mainly on scenario generation, dashboards and improved forecast accuracy, rather than on financially-ranked action lists with explicit economic drivers.⁵¹⁶⁷²¹⁸ Lokad’s technology pages and case studies emphasize monetized objective functions (economic drivers such as holding cost, stockout penalty, obsolescence, basket effects) and ranked decision lists (investment/divestment recommendations, order lines, transfers) ordered by expected ROI.¹³¹⁸²⁶²⁰²⁷²⁵ The Air France Industries MRO case, for example, documents prioritized lists of parts to invest in or divest from, reflecting explicit trade-offs between service levels and working-capital tied up in inventory.²⁰²⁷²⁵

Architecture and cloud posture. Both vendors are multi-tenant SaaS and cloud-native, but implement this differently. KetteQ splits responsibilities between Salesforce (UX, collaboration, security, data sharing) and AWS (solver, analytics, data management) and highlights openness via standard tools like SQL and Python rather than proprietary stacks.¹²⁶⁴¹⁵ Lokad operates as a single-stack SaaS on Microsoft Azure, with a custom event-sourced data store, content-addressable blob storage, and a distributed execution engine for Envision scripts; third-party dependencies are intentionally minimized in favor of a tightly integrated in-house stack, including its own forecasting and optimization libraries.¹⁷¹³¹⁸

Evidence & maturity of “AI” claims. KetteQ’s AI language is ambitious (agentic AI, world’s most adaptive planning, real-time adaptability, multi-pass probabilistic planning), but public documentation remains high-level and largely marketing-oriented; there are no detailed algorithmic whitepapers or external benchmark results beyond customer quotes and analyst/blog coverage.⁵¹⁶³¹⁰⁸¹³ Lokad’s AI narrative is grounded in specific algorithmic families (probabilistic forecasting, deep learning, differentiable programming) explained in public documentation and videos, and tied to external evidence such as the M5 competition and ten-year longitudinal case studies like Air France Industries.¹²²³¹⁷¹³¹⁸¹⁹¹⁴⁹²⁰²⁵

In short, KetteQ appears as a modern, Salesforce-centric, scenario-driven planning suite that uses AI/ML primarily to power forecasts and scenario scoring, while Lokad is a language-driven platform that integrates probabilistic forecasting and economic optimization into a programmable environment, with more explicit technical disclosure and external validation. For a buyer, this translates into a choice between a packaged, Salesforce-native application with strong IT alignment (KetteQ) and a programmable quantitative supply chain “engine” that demands more modeling work but offers deeper control over how uncertainty and economics are modeled (Lokad).

Corporate history, funding and positioning

Founding and leadership. Sourcing Innovation dates KetteQ’s founding to 2018, emphasizing that it was designed “from the ground up” to embody lessons learned from many prior supply chain planning implementations and from two earlier supply chain software ventures by the founders.³¹¹ KetteQ’s own “About” page describes the company as built by industry veterans with decades of experience in technology, data management and supply chain processes, and as aiming to “redefine the way businesses approach supply chain planning and execution” on Salesforce and AWS.² In 2021, Mike Landry, a long-time supply chain software executive (ex-Servigistics, ex-Genpact), was appointed CEO; independent coverage notes that he took over from founding CEO Cy Smith and was tasked with scaling the platform.¹⁵⁴²¹

Funding rounds. Public funding disclosures and secondary coverage are reasonably consistent:

Pre-Series A: ~USD 1.9M in 2021 (reported in later summaries, though not heavily documented in primary press at the time).³¹²
Series A: USD 9M in 2023 (mentioned in funding summaries and Series B-related coverage).³¹¹¹²
Series B: USD 20M announced on 5 August 2025, led by Vocap Partners with participation from existing investor Circadian Ventures, bringing total disclosed funding to USD 30.9M.¹⁰¹¹⁷¹²¹³¹⁴

The Series B press releases and news articles consistently position the funding as intended to accelerate global expansion, expand the agentic AI roadmap (PolymatiQ™ and Agentforce) and expand delivery capacity.¹⁰¹¹¹³¹⁴⁹

Taken together, KetteQ presents as an early-growth, VC-backed platform vendor: funded and commercially active, but still in a scale-up phase, not yet a decades-old incumbent.

Market positioning. KetteQ’s messaging consistently positions the company as:

An “adaptive,” “AI-first” supply chain planning platform designed for volatility and uncertainty.⁵¹⁶⁴¹⁰
A modern, multi-tenant, cloud-native alternative to legacy APS systems built pre-2010.⁶³²²¹⁹
The “only” supply chain planning solution that can be deployed natively on Salesforce, offering a 360° view by blending supply chain and commercial data.¹⁶⁴

Independent sources (Sourcing Innovation, third-party funding coverage) reinforce the modern, cloud-native characterization, but do not independently validate the uniqueness claims (“only solution,” “world’s most adaptive”) which should be treated as marketing language rather than verified fact.³¹¹⁸¹⁹⁹

Product and architecture

Product surface

KetteQ’s product portfolio is organized into planning and execution:

Supply chain planning: demand planning, inventory planning, supply planning, MRP and MEIO (multi-echelon/multi-item optimization).⁵¹²¹
Supply chain execution: control tower, fulfillment and allocation, asset management, work order management, and service parts planning.⁷²¹

Each module section on the website emphasizes:

AI-driven statistical / machine-learning-based forecasting.
Automated scenario analysis via PolymatiQ™.
Multi-echelon and multi-item optimization (for inventory), considering budgetary and service-level constraints.
Support for complex product structures (multi-level BOMs) and constraints (capacity, lead times, yield, vendor constraints).⁵¹⁷²¹

Concrete examples of functions include:

Safety stock optimization and order policy optimization for inventory planning.²¹
Supply planning under production capacity and lead-time constraints.¹
Service parts planning and FSL/truck-stock optimization.²¹
Real-time monitoring and adjustment via control-tower-style dashboards.⁷

The product messaging is consistent with a mid-to-upper-tier APS replacement or augmentation, spanning tactical planning and some operational execution visibility.

Architecture

KetteQ’s platform page and related blogs outline a two-tier architecture that:

Deploys the user experience and collaboration layer on Salesforce, taking advantage of Salesforce’s data model, security, workflow and ecosystem.¹²⁴¹⁵
Hosts the data- and compute-heavy solver and analytics components on AWS, using cloud-native services for elasticity and performance.¹²⁴

Key architectural characteristics highlighted by KetteQ and echoed by Sourcing Innovation:

Multi-tenant cloud-native SaaS, designed for horizontal scalability and real-time integration.¹⁶³²²
Use of mainstream technologies (SQL, Python, JSON) for data handling and extensibility, rather than proprietary databases or expression languages.¹⁶
Tight Salesforce integration (including Salesforce Manufacturing Cloud) so that planning results and data are directly visible to commercial and finance users without complex data replication projects.¹⁴²⁴¹⁸

The Sourcing Innovation review adds independent confirmation that the platform is indeed multi-tenant and cloud-native, and that it leverages modern web technologies rather than retrofitted legacy stacks.³¹¹ However, neither KetteQ’s own pages nor third-party write-ups provide low-level diagrams or specifics on data models (e.g., event sourcing vs. relational schemas), concurrency controls or failure modes.

Deployment model

Deployment is SaaS-based and cloud-hosted. Public materials emphasize:

Faster deployments via Salesforce-native UX and existing IT support; reuse of Salesforce security and integration patterns.¹²⁴¹⁵
Incremental roll-out by domain (e.g., start with demand planning, extend to inventory and supply planning later).⁵¹
A conversational Gen-AI interface for planners to ask questions, run scenarios and access data using natural language.⁵¹

However, there is little public information on typical implementation timelines, project staffing, or reference implementation patterns (e.g., whether KetteQ provides its own delivery teams vs. partner system integrators). Customer quotes mention “dramatically improved efficiency and precision” and 2–3x improvements in planning accuracy or alignment, but these are high-level and not backed by detailed time-series data or before/after KPIs.⁵¹⁸

AI, ML and optimization claims

Forecasting and AI

Across its planning pages, KetteQ asserts:

Use of “AI-driven insights, advanced machine learning for demand forecasting, and automated scenario analysis.”⁵¹²¹
AI/ML-enhanced MRP that “is always on the lookout for signs of changes ahead” and continuously monitors signals.¹
A Gen-AI conversational interface to interact with the planning solution.⁵¹

These claims establish that KetteQ uses machine learning for forecasting and that it integrates a large-language-model-type interface. However, missing details include:

The classes of ML models used (e.g., gradient-boosted trees, neural networks, probabilistic models).
How forecasts are calibrated, evaluated and updated (e.g., error metrics, retraining cadence).
Whether lead times, returns and other non-demand uncertainties are modeled explicitly.

By contrast, Lokad’s demand forecasting FAQ and technology pages explicitly state that Lokad uses differentiable programming and deep learning, applied to detailed historical data and external signals where relevant, to generate probabilistic demand and lead-time forecasts; they also highlight the M5 competition results as external evidence of state-of-the-art performance.¹²²³¹⁸¹⁹¹⁴⁹ This does not invalidate KetteQ’s AI claims but underscores that KetteQ’s public disclosures are at a higher, less technical level.

PolymatiQ™ solver and optimization

The PolymatiQ™ solver is described as:

A “revolutionary, patent-pending” supply chain planning solver that runs thousands of scenarios, automatically tunes parameters and continuously learns from dynamic market shifts.⁵¹⁶⁷⁸
The “world’s first agentic AI engine” for supply chain planning, enabling adaptive planning by exploring thousands of potential futures in parallel.¹⁶¹⁰¹¹

Scenario-related claims are consistent across planning and execution pages: PolymatiQ™ is credited with automatically evaluating thousands of potential outcomes, scoring the resilience of plans across KPIs, and identifying strategies robust to disruptions.⁵¹⁶⁷²¹⁸ However, several technical questions remain unanswered in public sources:

Objective functions: what is being optimized? Cost, service level, profit, resilience metrics, or a weighted combination?
Constraints: how are capacity, lead times, budgets, and service-level targets modeled?
Algorithms: does PolymatiQ™ rely on mathematical programming (LP/MIP), constraint programming, metaheuristics (e.g., genetic algorithms, simulated annealing), reinforcement learning, or some hybrid?
Representation of uncertainty: are scenarios generated from explicit probability distributions or from heuristics / stress tests?

The Sourcing Innovation review adds a little insight by referring to KetteQ’s use of multi-pass optimization and a modern stack that can run many scenarios efficiently, but stops short of revealing algorithmic details; it is a qualitative endorsement rather than a technical deep dive.³¹¹

Lokad’s solver story, by comparison, is more explicit: its technology and documentation pages explain how probabilistic demand and lead-time distributions feed into stochastic discrete descent (for integer decisions) and differentiable programming-based optimization, and how economic drivers (holding costs, stockout penalties, basket effects, etc.) are encoded in Envision scripts to produce ranked decisions.¹²¹⁷¹³¹⁸¹⁹²⁶ Lokad’s differentiable programming docs and blog posts further elaborate how gradient-based optimization is applied over large-scale relational data to jointly learn forecasting models and decision policies.¹⁷¹⁸¹⁹²⁵

Given what is publicly available, it is reasonable to conclude that PolymatiQ™ does more than simple rule-based MRP or fixed safety-stock formulas. The emphasis on multi-scenario analysis, parameter tuning and resilience scoring suggests at least a simulation-driven optimization component. But in the absence of detailed documentation, PolymatiQ™ should be treated as a black-box solver whose internal workings cannot be independently assessed from public sources.

Deployment, clients and evidence

Named customers and sectors

KetteQ publicly highlights several named customers:

Coca-Cola, Carrier, Johnson Controls, NCR Voyix, Alliance Consumer Group and others are cited as brands “trusting” KetteQ’s platform in funding-related news coverage.⁷⁸⁹
Case-style website content mentions customers such as MobilityWorks (automotive mobility), a vending-operations company, and manufacturers using Salesforce, with quoted improvements in efficiency, precision and growth.⁵¹⁸

These references confirm that KetteQ has real, named enterprise customers across manufacturing, distribution and service operations. However, public case descriptions remain relatively high-level; they provide percentage improvements (e.g., 170% CARR growth, 3x precision, 2x growth trajectory) but do not expose detailed before/after metrics, time series or model diagnostics.⁵¹⁸⁹

Lokad’s case materials around Air France Industries MRO and other sectors (retail, aerospace, manufacturing) provide somewhat more detailed narratives, including historical horizons (10 years of data), data volumes (~1 million SKUs), and explicit financial outcomes (e.g., identified tens of millions of euros in divestment opportunities, reduction in working capital while raising service levels).¹³¹⁸²⁶²⁰²⁷²⁵ That said, even Lokad’s public case studies stop short of exposing full datasets, which is typical for commercial software vendors in this space.

Deployment & implementation practice

KetteQ emphasizes:

Salesforce-native UX leading to faster user adoption and easier IT governance.¹²⁴¹⁵
Direct connectors to Salesforce Manufacturing Cloud, enabling better alignment between sales forecasts and supply chain plans.²⁴¹⁸
A Gen-AI conversational interface to reduce friction in accessing insights and running scenarios.⁵¹

Independent sources (Salesforce-related PRs and podcasts) reinforce the narrative that KetteQ leverages Salesforce to improve cross-functional visibility and user adoption, but do not add much detail on configuration methodology, data-cleansing practices or change-management patterns.⁴²⁴¹⁸

Lokad, in contrast, describes a more explicitly programmatic deployment pattern: data is ingested via file/API, supply chain scientists write and iterate Envision scripts, and daily batch runs produce ranked decisions that are then integrated into ERPs and WMSs. Its Air France Industries case study and aerospace inventory pages detail a roughly 6-month implementation followed by a 6-month parallel-run phase, including mention of 10 years of historical data, 12 source systems and prioritization of investment/divestment lists.¹³¹⁸²⁰²⁷²⁵

The trade-off is clear: KetteQ offers a more conventional enterprise-app deployment (albeit on a modern stack), whereas Lokad offers a DSL-centric modeling project that can be more precise but demands more specialized modeling effort.

Assessment of technical depth and commercial maturity

Technical depth (as visible from public sources).

Strengths for KetteQ:
- Modern, multi-tenant SaaS architecture with Salesforce and AWS, independently corroborated by Sourcing Innovation.¹²³¹¹²²¹⁵
- Demonstrated ability to integrate tightly with Salesforce Manufacturing Cloud and to expose planning insights inside Salesforce’s UX, an advantage for organizations already standardized on Salesforce.¹⁴²⁴¹⁸
- Use of AI/ML for forecasting and scenario evaluation, with PolymatiQ™ orchestrating multi-scenario analysis and parameter tuning, implying non-trivial optimization logic beyond classic MRP or spreadsheet planning.⁵¹⁶³⁷
Gaps / unknowns for KetteQ:
- No public technical documentation of PolymatiQ™’s algorithmic internals (objective functions, constraint handling, optimization methods).
- Limited detail on how uncertainty is mathematically modeled and propagated (probability distributions vs. stress-testing).
- No external forecasting or optimization benchmark results (e.g., public competitions) to quantify performance relative to state-of-the-art.

Relative to Lokad: Lokad’s technical disclosures are considerably deeper: Envision’s language specification, probabilistic forecasting and differentiable programming docs, and detailed technology pages make it possible to understand the architectural and algorithmic choices, and M5 competition results provide an external benchmark for forecasting accuracy.¹²²³¹⁷¹³¹⁸¹⁹¹⁴⁹²⁶ Lokad also documents nuanced topics such as integer handling in differentiable programming and optimization over large-scale relational data, which indicate ongoing R&D investment at a fairly advanced level.¹⁷¹⁸²⁵

From a strict “state-of-the-art” standpoint, KetteQ has not yet provided enough technical transparency to assess whether PolymatiQ™ and its AI stack reach the sophistication of probabilistic-distribution-centric and differentiable-programming-based approaches documented by Lokad and some academic literature. It is entirely possible that KetteQ’s internal methods are sophisticated, but absent documentation or benchmarks, a skeptical, evidence-driven assessment must treat them as unproven beyond selected customer testimonials.

Commercial maturity.

KetteQ has:
- Multiple named enterprise customers in manufacturing, distribution and services.⁷⁸⁹
- USD 30.9M in venture funding and a growing team, with CEO and leadership profiles indicating experience in scaling supply chain software companies.¹⁰¹¹⁸¹⁵¹³¹⁴
- A clearly articulated product suite and go-to-market story focusing on Salesforce-centric organizations.
However:
- The company is 7 years old (founded 2018), with its flagship solver and branding evolving quickly over the last few years (e.g., the recent shift to “agentic AI” language).³¹⁰¹¹
- Public reference cases are relatively thin in quantitative detail and span a limited set of sectors compared to older APS vendors.

By contrast, Lokad has been operating since 2008, with over a decade of evolution from hosted forecasting to a full probabilistic optimization platform, and documented long-running deployments (10-year collaborations such as Air France Industries) that suggest higher commercial and technical maturity in certain verticals (e.g., aerospace MRO).¹⁶¹²¹³¹⁸²⁶²⁰²⁷²⁵

In summary, KetteQ is a credible, modern, VC-backed vendor with genuine traction and a technologically up-to-date architecture, but its AI and optimization capabilities remain somewhat opaque from a research-driven viewpoint. Lokad, while smaller and more specialized, provides more direct evidence of advanced probabilistic and optimization techniques and longer-running deployments in complex environments.

Conclusion

KetteQ delivers a modern, Salesforce-centric supply chain planning and execution platform with a clear focus on AI-driven scenario planning through its PolymatiQ™ solver. Public evidence supports the claims that it is multi-tenant, cloud-native, and architected on Salesforce and AWS; that it uses machine learning for forecasting; and that it has secured real, named enterprise customers and significant venture funding. Independent commentary (Sourcing Innovation) corroborates the modernity of its stack and confirms that the product is not a thin repaint of legacy code.

However, KetteQ’s public documentation stops short of exposing the mathematical and algorithmic substance behind PolymatiQ™ and its “agentic AI” branding. There is no detailed disclosure of objective functions, constraints, probabilistic modeling, or optimization methods, nor any public benchmark data comparable to forecasting competitions or reproducible case studies with full before/after metrics. As such, a rigorous, skeptical assessment must treat KetteQ’s AI and optimization claims as plausible but unverified beyond marketing material and customer quotes.

Compared with Lokad, which has publicly documented its evolution through quantile forecasting, probabilistic distributions, deep learning and differentiable programming, and has externally validated forecasting performance (M5) and long-running case studies in complex environments, KetteQ presents as a more conventional but modern enterprise app suite: easier to align with Salesforce-centric IT landscapes, but less transparent about the internal workings of its AI and optimization stack. For buyers, the choice is not purely about features: it is also about how much they value a programmable, white-box quantitative supply chain engine (Lokad) versus a packaged, Salesforce-native application with a strong UX and IT story (KetteQ). In either case, due diligence should go beyond marketing pages: ask for technical deep dives, model documentation, and concrete historical KPI improvements before accepting any claims of “agentic AI” or “world’s most adaptive” planning at face value.