Kallidin AI

Most AI tools give you an answer. KAL gives you an answer that has already been argued with. Before anything leaves the system, a second set of agents tries to take the first answer apart. Wrong data source, a join that double-counts, a definition of "customer" that doesn't match the one finance uses. If it survives that, it comes to you with the challenge and the response recorded alongside the working. It's the same thing a good analytics team does when a senior analyst reviews a junior's work before it goes to the board. The difference is that it happens every time, in minutes, and the review is written down. That's what we mean when we say an answer has been checked. Not that a model sounded confident. That something tried to prove it wrong and couldn't.

What makes enterprise AI so difficult? Warwick's answer comes down to two things: governance and trust. Governance, because AI projects are popping up everywhere and organisations can't get a handle on what's actually working, or what it means for the business. And trust, because it's hard to trust an answer when you can't see the decision-making behind it. That's the nature of AI, and it's a huge barrier. Both are what we built KAL to solve.

The problem with an AI that answers questions about your business isn't getting an answer. It's knowing whether to trust it. An answer with no working behind it is an opinion. It might be right. But nobody is taking an opinion into a board meeting, and no finance director is signing anything off on one. So KAL shows its working. Every answer comes with the steps behind it: the data it used, how it got from the question to the conclusion. You can check it the way you'd check an analyst's work, because that's what it is. That's the difference between an AI that sounds confident and one you can actually use.

What does it mean to be human when AI can do the analysis? John Brodie, our Co-founder, has an answer: it means asking better questions. Most thinking in most roles is regimented. Done the way it's always been done. And when people first use AI, they ask the same questions as everyone else. But get an answer back in minutes, and it leads you to another question, then another, until a couple of hours in you're asking something you couldn't have reached before. John's bet is that the leaders of the future will be creative thinkers. Kallidin exists to give them the room.

We asked John Brodie, Warwick Beresford-Jones and Samuel Riddington four questions. Describe Kallidin in three words, favourite feature, most overused words in AI, and the best thing about working here. You'll notice the overused words and the three words describing Kallidin have nothing in common. That's deliberate. Also, Sam described the team as 'old and grey, still acting like little children.' We're taking it as a compliment. Watch below.

KAL is live. Kallidin is the home of the Autonomous Data Office, and KAL is the product we've been building towards for a long time. Every data team knows the queue. A question gets asked, turned into a brief, dropped into a work stack, and answered weeks later, if it gets answered at all. One analyst means one problem at a time, and most organisations have far more questions than analysts. KAL runs those questions in parallel. Ask it what you'd ask your data team and it works through the analysis, at whatever scale the business needs, with the decision-making behind every answer laid out for you to check. No black box, no guessing why the model said what it said. AI you can understand. Answers you can trust.

Kallidin is really two things brought together. On one side, over a hundred years of combined analytics experience across the team, delivered into organisations of every shape and size. On the other, what AI has only just become capable of. Our Co-founder and Chief Product Officer, Warwick Beresford-Jones, on what that combination builds: the autonomous data office, KAL 👋 Analytics that's faster, accurate, and actually delivers value.

A board asks a simple question. Which of our clients are actually profitable? Someone in finance logs the request. An analyst picks it up the following week. The data lives in four systems, so there's a fortnight of reconciliation before anyone trusts the numbers. Six weeks later the answer arrives, and the board has moved on. The analyst did nothing wrong. The question was reasonable. The cost sat in the gap between the two: every decision made during those six weeks was made without the answer, and most of them couldn't wait. Firms are good at measuring the cost of producing answers. Headcount, tooling, consultancy days. Almost nobody measures the cost of waiting for them, because waiting never lands on an invoice. We built Kal to close that gap. Ask a question in the morning, interrogate the answer that afternoon, with a full audit trail of how it was reached. The waiting was always the expensive part.

We went through a lot of names before Kallidin. Kal-El is Superman's original name, and two of our co-founders, John and Warwick, pushed for it. Kallidin came from there, as a name that could carry a company rather than just a product. Then we looked it up. Kallidin is a peptide in the human body, cleaved from a larger protein by an enzyme called kallikrein. It dilates blood vessels, widening the channel so that blood reaches the places that need it. That's the same language we had been using for months to describe the problem we work on. Data sitting inside a business, unable to reach the people and the systems that need it in time for it to matter. Adding more data doesn't fix that. It lengthens the queue. What has to change is the width of the channel. The name was settled after that, and here we are.

Our Co-Founder, Samuel Riddington, spent 25+ years in enterprise consulting firms before starting Kallidin. That experience shaped a clear view: traditional consulting is structured around effort, with engagements that typically end at the recommendation. Sam wanted to build a firm that stays accountable for what happens after it. That's the thinking behind Kallidin. We build autonomous data teams that carry the work through to a measurable outcome and stand behind the result, with every step checked and auditable along the way. Here's Sam on the experiences that shaped it.