What data do you need to start with a decision engine?

Which minimum data set lets a decision engine start?

The minimum is a question, the options, and the facts that move the answer. Write the question in one sentence, such as whether to sign a contract or which site to open first. List two or three options. Then note the five to ten facts that would change your mind if they moved: a price, a date, an exposure, a headcount. Keep the sheet to one page, so that anyone on the team can read it in two minutes.

This short sheet is enough for a first read, because a first score tests the shape of the problem before it tests the detail. If the sheet is hard to fill, that is already a useful result: it shows which facts the team has to collect before any engine can help.

Which types of data feed a decision engine?

Three groups of inputs appear in most cases. Measurable facts include revenues, dates, volumes, and exposure. Judgments are ratings from people who know the situation, given on a scale. Context covers the rules around the decision, such as regulation, currency, and the countries involved. A balanced file holds some of each. Mixing the three groups gives the model both the numbers and the reading of the numbers.

Quality matters more than volume. Ten inputs with a clear definition and a named source serve a model better than a hundred loosely defined figures. For decisions across Asia, add the currency and the jurisdiction beside every number, so that the model compares like with like.

Which documents can you share to start the decision engine?

From the Study upwards, you can send internal and external documents: balance sheets, contracts, meeting minutes, presentations, emails, and raw data (EQOS Asia pricing page, as of October 2026). The method page lists internal data from finance, HR, operations, sales, and product, across entities and countries, plus qualitative documents such as presentations, emails, and meeting notes.

Any format is accepted, digital or physical, and multi-format material such as PDF, video, audio, and raw data can be ingested (EQOS Asia method page, as of October 2026). External sources such as LinkedIn, ICIS, SimilarWeb, and Crunchbase are analysed alongside. The more you give, the more accurate the model becomes. Share what you hold today, and add more files as the questions sharpen.

How much data does each tier of the EQOS Asia grid use?

Data volume grows with the tier. A Reading at US$358 starts from one specific question and answers in 24-48 hours. A Decision at US$910 integrates the internal and external documents of one case in 48-72 hours. A Study at US$1,369 uses the documents you send and delivers in 5-7 days (EQOS Asia pricing page, as of October 2026).

A Mapping at US$22,976 integrates 100% of the data you provide, from finance and HR to market, contracts, and communications, over 6-8 weeks. Choose the tier whose data appetite matches what you can gather in a few days. A smaller, well-sourced file beats a larger one that arrives late.

How do you prepare and date your data before sharing it?

Give every fact four labels: its source, its date, its unit or currency, and the person who can confirm it. Put the labels in a simple table with one row per fact. Date matters most of all, because a figure from last quarter and a figure from today describe two different situations. Write currencies with their three-letter code to avoid doubt.

Group the documents by theme, such as finance, contracts, people, and market, and name the files in plain words. Remove duplicates, and flag the documents you consider incomplete. A short cover note that says what each file proves saves the first call from becoming a file hunt.

Can you start the decision engine with a self-assessment first?

Yes. The Λ simulator on the EQOS Asia site asks you to rate each of 9 dimensions between 0 and 100 and returns a score in real time (EQOS Asia simulator page, as of October 2026). Your honest self-assessment feeds the engine, so the exercise starts before any document is gathered.

The privacy policy states that simulator data is retained only for the session, unless you request a deeper analysis (EQOS Asia privacy page, as of October 2026). Use the result as a map: the dimensions with the lowest ratings show where documents will add the most value in the next step. Keep a note of your ratings to compare them with the first full score.

How is the data you share kept confidential?

A systematic NDA is signed before any exchange. The standard workflow starts with the NDA and a framing call of 60 to 90 minutes by video in Singapore time, then moves to structured collection in an encrypted, isolated environment over 2 to 5 days (EQOS Asia method page, as of October 2026).

Questionnaire answers are encrypted in your browser before they are sent, and data is hosted in Singapore under PDPA-aligned handling (EQOS Asia pricing page, as of October 2026). Share documents once the NDA is in place, and ask for the pre-diagnostic when you want to discuss the file first.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile