Get real work out of AI.

Madlytics is a data and AI consultancy in Sydney. We help companies work out where AI is worth using, and then we build it with them. Once it's running, we keep the costs down.

Where the experience comes from

Madlytics is Alex Scriven's practice. Alex has spent more than ten years in data science, analytics and AI, in some quite different organisations.

He leads every engagement, and brings in senior specialists in AI strategy, data engineering and software when a project calls for them.

Large technology companies
Alex has been a lead data scientist and tech lead on Atlassian's AI data science team, and has worked with other large technology companies too. That work covers agent workflows, the information agents rely on, running experiments and measuring results.
Government
Alex was a senior data scientist in a large state agency, where he ran machine learning projects from scoping through to delivery. He also built analytics for senior decision makers and tools for non-technical teams.
Agencies
We built a software platform for marketing agencies, where AI agents work from a shared company brain. We also set up AI workflows for client delivery and bring agencies' AI costs down.
Social enterprises
We've built AI for organisations on tight budgets, including a news hub where AI agents support advocacy.
Universities
Alex has spent twelve years teaching and doing research. He built and delivered Masters-level courses in machine learning and AI at UTS, and held a data science research fellowship at UNSW.
Startups
Alex is a founder, a Startmate First Believers investor and mentor, and a moderator of Sydney Startups, a community of 13,000.

About Madlytics

Where are you up to with AI?

Starting out

A sensible first step

You'd rather not lose a year working out where to begin with AI. We find the work where it'll help most and get your team using it.

  • Pick the first use cases by what they're worth
  • Set up the first agents
  • Train the people who'll use them
  • Agree up front how you'll judge whether it worked
Already running

Get more for what you spend

AI is in daily use and the bill keeps climbing. Some teams get far more out of it than others. We find what's worth paying for and fix the rest.

  • See what you spend and what you get back for it
  • Cut token costs
  • Rework the workflows that aren't worth what they cost
  • Give leadership numbers they can check

The difference between a skill and a loop

A lot of what gets called an agent is really a skill: a set of instructions someone wrote once, which runs the same way every time. A skill is one and done. It can't improve, and as the business around it changes it slowly gets worse. A loop is a skill with the extra parts that let it learn.

A skill: a request goes in, the skill runs, an output comes out, and nothing comes back to improve it. Request Skill Output

A skill

The output goes out and nothing comes back, so the skill has nothing to learn from.

A loop: do the work, check the result, record the lesson, update the agent, then do the work again. Do the work Check the result Record the lesson Update the agent

A loop

Every run gets checked, and what was learned is written down for the next run to start from. That's more work to set up. It's also why a loop keeps up with the business and improves the more it's used.

Most AI overspend is in five places

We see the same ones in most companies that have been running agents for a while. In a review we go through each one and tell you what it's costing you.

Model choice
Routine steps running on the most expensive model.
Context
Agents re-reading material they don't need.
Caching
The same long instructions, paid for in full on every run.
Retries
Agents stuck retrying, with no rule for when to stop.
Workflow
Lots of small calls where one planned pass would do.

We also teach, and wrote a book on data strategy

The book

Data Science and Analytics Strategy: An Emergent Design Approach

By Kailash Awati and Alex Scriven (Routledge, 2023). It's a guide to building data capability without betting everything up front, and it draws on interviews with data leaders across Asia-Pacific.

See the book at Routledge

12 years
teaching and research at universities
100,000+
enrolments across four DataCamp courses
Hundreds
of colleagues trained in data and AI inside a large technology company
Published
research on AutoML in industry, big data in education and group reasoning

We start small, and only grow the parts that work

  1. Audit

    We look at how your team works now, what you spend and where AI would help.

  2. Plan

    You get a short roadmap: what comes first, who owns it, how success is measured and what it'll cost.

  3. Build

    We set up and build alongside your team, one useful piece at a time.

  4. Hand over

    Your people run it from here. We train them, and we stay on call while it beds in.

Tell us where you're up to with AI.

One conversation is usually enough to know whether we can help, and what a sensible first step looks like.

Talk to us