AI research for business

Turn idle data into a productive asset.

Kunkafa is an AI research lab building custom models for tabular and time-series data, turning business questions into intelligence you can evaluate and use.

Talk to Kunkafa See the live record
Use cases

Purpose-built models for your business question

Custom research and model development around your investment mandate, operating priorities and available data.

01

Investment & portfolio

  • Strategy-specific return distributions
  • Position sizing & dynamic exposure
  • Rebalancing & position management
02

Risk & market regimes

  • Market-state transition probabilities
  • Volatility, drawdown & tail-risk models
  • Stress scenarios & hedge allocation
03

Liquidity & execution

  • Liquidity & spread forecasting
  • Slippage & market-impact estimation
  • Execution timing & order allocation
04

Retention & customer value

  • Churn & renewal probabilities
  • Customer lifetime value
  • Retention priorities & response propensity
05

Sales & commercial planning

  • Demand & revenue forecasting
  • Lead-conversion probabilities
  • Pricing sensitivity & inventory planning
06

User behaviour & operations

  • Activation & engagement propensity
  • Conversion & journey modelling
  • Anomaly detection & case prioritisation

Your data stays yours. Access, hosting and handover are agreed with you in advance.

Why numbers

Most of the investment went into words. Most of the data is numbers.

Language models have transformed how organisations read, write and search. They were trained on text, and text is what they are good at. The data most organisations actually run on is different: sales by week, transactions by the second, inventory by location, sensor readings by the minute, prices by the tick. It is structured, noisy and numerical, and the same investment has not reached it.

Kunkafa’s research is concentrated on that data. The models are built for structured records and sequences over time, they say how likely an outcome is rather than which outcome sounds right, and they are tested on data they have never seen before anyone relies on them. Language models and numerical models can serve complementary roles inside one organisation. The point is to match the model to the question.

Our live showcase

Models you can watch and evaluate

Kunkafa’s market analysis is the lab’s research applied in public, where anyone can compare a forecast with what happened. Seventy AI experts, a mixture-of-experts neural network, assess every market and duration.

Direction

Up and down, always both

Magnitude

The size of the move

Duration

How long it has to get there

Kunkafa’s confidence

The chance price touches the level within the duration

Why markets? A demanding test for AI

Noisy signals

Separate patterns from noise

Shifting regimes

Recognise changing conditions

Rare events

Learn from limited examples

Evolving relationships

Track changing dependencies

  1. Forecast
  2. Observe outcomes
  3. Evaluate models
Calibration first

Confidence, checked against live outcomes

Live statistics are published daily. Weekly and monthly reports give the deeper market-by-market breakdown.

127,380,418

live forecasts evaluated across all markets

Updated 6 November 2026

~10 billion

training data points

~2 billion

held-out test points

Gold

By Kunkafa’s confidence reached the level · All time
0%25%50%75%100%5%25%50%75%95% 89% at 90%

Kunkafa’s confidence 90% or higher 89% reached the level 419,773 forecasts

Each point counts every forecast at or above that confidence.

90% or higher

89.2%

reached the level

419,773 forecasts

95% or higher

94.6%

reached the level

60,037 forecasts

Live tracking by market

Settled forecasts at Kunkafa’s confidence of 95% or higher, both directions, from each market’s latest full report.

Gold

View report

4 September 2026 to 2 November 2026

34,367 settled forecasts

  1. went 300% or more11.0%
  2. went 200 to 300%8.9%
  3. went 125 to 200%28.0%
  4. went 100 to 125%, reaching the level44.4%
  5. went 75 to 99%1.8%
  6. went 50 to 75%1.9%
  7. went 25 to 50%1.8%
  8. went under 25%2.2%

S&P 500 Index

View report

4 September 2026 to 30 October 2026

55,982 settled forecasts

  1. went 300% or more14.7%
  2. went 200 to 300%12.1%
  3. went 125 to 200%31.1%
  4. went 100 to 125%, reaching the level36.0%
  5. went 75 to 99%1.5%
  6. went 50 to 75%1.4%
  7. went 25 to 50%1.3%
  8. went under 25%1.9%

Euro - Dollar

View report

9 September 2026 to 30 October 2026

82,163 settled forecasts

  1. went 300% or more25.0%
  2. went 200 to 300%12.2%
  3. went 125 to 200%24.9%
  4. went 100 to 125%, reaching the level27.8%
  5. went 75 to 99%2.0%
  6. went 50 to 75%2.2%
  7. went 25 to 50%2.1%
  8. went under 25%3.8%

SanDisk

View report

29 August 2026 to 2 November 2026

156,921 settled forecasts

  1. went 300% or more27.1%
  2. went 200 to 300%12.7%
  3. went 125 to 200%26.4%
  4. went 100 to 125%, reaching the level26.4%
  5. went 75 to 99%1.5%
  6. went 50 to 75%1.4%
  7. went 25 to 50%1.3%
  8. went under 25%3.2%

Percentages are shares of each market’s settled forecasts. 100% means the level was reached within its duration; 300% means three times that distance, not a price change. Forecasts can overlap. Past results do not promise future ones.

Explore all reports

One model platform

Three ways to use it

API

Integrate into your systems

MCP

Connect to AI applications

Interface

Explore with any level of expertise

Enterprise capacity for API and MCP is agreed with you.

Services

Across the model lifecycle

From feasibility and data assessment to deployment and monitoring. Scope and validation criteria reflect your business question and available data.

  • Research
  • Advisory
  • Custom models
  • Evaluation
  • Integration

How an engagement runs

  1. 01

    Scoping

    One conversation to name the question, who acts on the answer, and what a good result looks like.

  2. 02

    Data assessment

    A review of the data you hold: coverage, quality, gaps, and whether it can support the question at all. If it cannot, we say so.

  3. 03

    Success criteria

    The measure of success is written down before any model is trained, so the result can be judged rather than argued.

  4. 04

    Modelling and evaluation

    Models are built on your data and tested on data they have never seen, in the conditions they are meant to operate in.

  5. 05

    Handover

    A model your team can run, monitor and question, with its record kept from the first day.

Who you would work with

Kunkafa was founded by Kareem Farid, who built Kundera (2019–2022), an early large-language-model application for research and writing, before turning to structured data, and Saleem Mugal, whose long career in consultancy shaped how Kunkafa scopes, evaluates and hands over work. Engagements are led by the founders.

Moving AI from an experiment into day-to-day operations needs evidence that a model performs in the setting where it is meant to operate, not in a demo. Kunkafa builds every model the way the public product is built: both sides of the question, a measure of confidence, a test on unseen data, and a record kept from the first day. A model that is never scored in public should not be relied upon in private.

Put your data to work.

Tell us the question, the data you hold and where the answer would be used. Access, capacity and pricing are agreed with you.

Contact Kunkafa

Or write to contact@kunkafa.com. To see the models at work first, create a free account.