Building worlds first commodity AI agent.
Enabling better decisions for CFOs.Chief Financial Officers.Risk Managers.Procurement Leaders.Proc. Leaders.Trading Strategists.Data Scientists.
From volatility to strategy.
We help industrial leaders turn volatility into strategic advantage with event‑based forecasting and AI agents designed for real‑world complexity.
We track what matters - Curious about KPIs like MAE, RMSE, or sMAPE? Let’s talk!
Transforming Events Into Decisions.
1 – Baseline
Building robust daily forecasts on time-series data to challenge existing methods and establish a reliable benchmark for decision-making.
2 – Events
Adding live event data — from markets, geopolitics, and supply chains — to identify peaks, lows, and anomalies that traditional models miss.
3 – Integrations
Operationalising event‑driven forecasts — embedding results into procurement, finance, and risk workflows for real business impact.
4 – Agents
At scale, multiple agents combine forecasting, event extraction, and scenario intelligence to automate decisions and augment teams.
How Leaders apply AI Forecasting.
Real-world use cases where AI forecasting delivers measurable results.
Make Procurement Data-Driven
Use forecasts to plan purchasing and negotiate contracts with data‑backed insights.
Stabilize Financial Planning
Bring reliable price forecasts into budgets and financial plans to reduce volatility risks.
Anticipate Market Shocks
Model geopolitical and market events to understand price impacts early and act proactively.
And you can go even further.
Improve Trading Strategies
Leverage real‑time forecasts and scenario analytics to guide trading and arbitrage decisions.
Optimize Hedging Decisions
Use AI-powered forecasts to time and price your hedging strategies with confidence.
Automate Risk Reporting
Provide ready‑to‑use reports with KPIs, forecasts, and scenarios for risk and management teams.
Operating from around the world.
Building AI That Works For You.
DSA brings together experts in Data Science, Finance, Procurement, and Trading to build AI agents for commodity forecasting and decision intelligence.
Empowering Market Leaders with AI Confidence.
Built by researchers and operators, we combine cutting-edge AI, deep industry expertise, and hands-on execution to transform how businesses navigate volatile commodity markets.
Commodity expert, data scientist, or decision-maker?
Frequently Asked Questions
Commodity markets are highly volatile — prices move with geopolitical shifts, regulatory changes, and supply-chain disruptions. Traditional tools often fail to capture these dynamics, leaving companies exposed to sudden shocks.
Our Commodity price forecasting uses advanced time-series models enriched with real-world event data to predict price movements more reliably. This helps procurement, finance, and risk teams make better decisions, reduce uncertainty, and protect margins.
Traditional forecasting tools often stop at historical data. Yet markets move on news, policy shifts, and disruptions that traders interpret in real time, drawing on countless information sources and experience.
Our approach combines this human intuition with technology: We digitize existing processes and build AI agents that learn how markets react to global events. The result is a tool that translates signals into actionable forecasts, supporting traders, procurement leaders, and finance teams in making faster, more confident decisions.
Forecast accuracy can be measured in different ways — whether by MAPE, RMSE, MAE, or hit rate. Which metrics matter most depends on your use case and objectives. We align our evaluation with the KPIs that are relevant for you.
In current projects, we achieved intraday and day-ahead accuracy levels above 99% for liquids and around 98% for metals. Reported error metrics include MAPE values below 1% in liquids and around 2% in metals, RMSE of 0.27, MAE of 0.21, and consistently high hit rates.
But the success of a project is never defined by KPIs alone. Forecast data must first be operationalized — embedded into procurement, finance, or risk workflows to create measurable value. In an initial discovery call, we jointly define what success looks like for your organization, set milestones, and establish how results will be assessed.
Our methodology applies to any commodity with sufficient data and market signals. Among others, we currently cover:
- Energy: Brent & WTI crude, refined products, TTF gas, and European power (day-ahead, intraday, baseload & peakload)
- Metals: aluminium, copper, steel benchmarks (e.g. Aluminium, HRC Steel), as well as nickel and zinc.
- Other raw materials: extendable where data quality and liquidity allow
Getting started is simple: send us a message and we’ll schedule a short Teams call. In this first conversation, we introduce the DSA approach and discuss potential use cases.
If there is a fit, we bring together your stakeholders and our project team to define scope and expectations. Based on this, you receive a tailored proposal.
From experience, most projects move from first contact to kick-off within 3–6 months — the time clients typically need for alignment and preparation.
In the first two project phases, we provide forecast data in the format you prefer (e.g., Postgres or similar). In the third phase, integration into existing systems is addressed if required. The concrete setup depends on your objectives and use cases, and is defined together during the project.
The goal is not to create another standalone tool, but to embed forecasting data into your teams’ workflows where it delivers real value.
Many clients begin with a pilot to validate results and build trust. Once proven, solutions are scaled into recurring processes, reports, and decision systems. From experience, pilots often evolve into long-term engagements as forecasting becomes part of daily operations and delivers sustained business value.
We see ourselves as your partner throughout this journey — whether you want to explore how AI can enhance your forecasting results, make your teams more efficient with advanced technology, or integrate AI solutions into your tool landscape for the long term.
The investment depends on scope, data requirements, and integration depth. A short-term pilot will differ from a fully embedded long-term solution.
We keep pricing transparent and align it with measurable business value. In practice, this means costs are tied to defined deliverables and the KPIs agreed at project start.
Our goal is not to sell another tool, but to deliver tangible improvements in forecasting, decision-making, and risk management.
In the first two project phases (typically 3–6 months), the internal effort is very limited. In most cases, it is sufficient if one person provides about one hour per month. We take care of data collection and forecast modelling, while your team gives feedback, answers relevant questions, and validates results with us.
As a data science partner, we see it as our role to challenge existing approaches, uncover potential biases, and complement your team’s expertise with data-driven insights. This keeps the internal workload low while ensuring maximum business value.
A conversation to begin - Let's explore what's possible
No obligations, just a conversation.
