Anneal / Data science practice

Commercial decisions, built into systems that run.

Anneal is a data science and AI practice. We take the commercial decisions a business makes every day, in risk, marketing, pricing, operations and customer service, model them, optimise them under real commercial and regulatory constraints, and put them into production with the testing and governance needed to keep them running.

Led by Doryan Gowty: 20 years at NAB, FICO and the Australian Bureau of Statistics.

Get in touch See the work →

Core capabilities / 06

Built for decisions with real operational consequences.

  1. 01

    Decision optimisation

    Model the decision, then optimise it within the limits the business actually works under: competitors, regulators, budgets and capacity. That might be a price, a marketing budget, a credit policy or an operational plan. Includes a review of the analytics stack and the workflows that carry a decision from model to market.

  2. 02

    Production ML & agentic workflows

    Working across the operational lifecycle from custom modelling to MLOps pipelines and autonomous agentic workflows. Predictive models deployed to drive business-critical decisions and agents with scoped permissions. Operational AI that is fully productionised, continuously monitored, and resilient.

  3. 03

    AI agent evaluation and governance

    Model and LLM selection, caching and cost control, then the testing that decides whether an agent is ready: LLM-as-judge evaluation, human review, shadow-mode deployment and permission scoping. That testing is the distance between a demo and something you can let near customers or money.

  4. 04

    Commercial analytics and decision strategy

    Designing and building data-driven decision strategies that bring fragmented data and predictive insights together. Custom model development and strategy design, operationalising these decisions directly into your existing enterprise systems rather than building static dashboards nobody opens.

  5. 05

    Business experimentation

    Designing data and systems around a continuous testing and experimentation rhythm. Every initiative is tested before it's promoted, and findings are fed back into the loop to build better models and strategies.

  6. 06

    Fractional data science leadership

    Senior capability without building a full internal function: setting the analytical direction, choosing what to build and what to buy, and carrying delivery until a permanent team makes sense.

Selected work

Case study / 2026

DTC

Beverage brand

Agentic AI and advanced analytics for eCommerce.

A platform Anneal built and runs end to end for an operationally complex DTC beverage brand, automating customer service, delivery exceptions, fulfilment and marketing-spend decisions in production.

  • MMM
  • Agentic Workflows
  • Optimisation
  • Inventory Management
  • Marketing Spend Management
Read the case study →