I registered Knowsis as a business in 2021. But the thinking behind it had been accumulating for considerably longer than that.
Twenty years of working at the intersection of analytical research and strategic consulting had taught me something that the industry rarely says out loud: the tools that researchers use and the frameworks that make research genuinely valuable are not the same thing. You can have access to every statistical package ever written and still produce research that changes nothing. And you can produce research that transforms a client's strategy using methods that would have been recognisable to researchers working twenty years ago, if the frameworks you are applying are grounded in a genuine theory of human behaviour.
The technology matters. But it is not sufficient.
Knowsis was built on that conviction. And it is why the business has two distinct tracks rather than one.
The Problem With How Research Gets Organised
The research industry has always had a fragmentation problem.
On one side are the analytical tools: platforms for data collection, software for statistical analysis, visualisation tools for output. These have become dramatically more powerful over the past decade. The shift from manual analysis to machine learning, from static reports to autonomous analytical pipelines, represents a genuine transformation in what is technically possible.
On the other side are the strategic frameworks: proprietary models for understanding brand equity, consumer motivation, and behavioural prediction. These take years to develop, require academic grounding and commercial validation, and cannot be replicated by purchasing a software licence. They are the accumulated intellectual capital of researchers who have spent careers thinking deeply about specific problems.
The industry has tended to keep these two sides separate. Analytics providers sell tools. Strategy consultants sell frameworks. The gap between them is where research value often gets lost, in the translation from what the analysis shows to what it means, and from what it means to what you should do.
Knowsis was designed to close that gap. Not by combining existing tools and existing frameworks, but by building both from scratch with the explicit intention that they would work together.
Track 1: The Data Alchemy Platform
The first track is the analytical framework. The Data Alchemy Platform is an insight operating system: infrastructure for research, not a collection of separate tools.
The platform currently consists of four analytical engines available now, with two additional self-service applications also available:
Description provides exploratory data profiling and survey structure analysis. It is the foundation that every other engine builds on, understanding what you have before deciding how to analyse it.
Distillation performs dimensionality reduction and KPI index creation. It solves a problem that every researcher with a large survey battery has faced: twenty questions that are all measuring variants of the same underlying construct, producing a table that is impossible to interpret and a client presentation that makes no one any wiser. Distillation finds the underlying structure and creates clean, interpretable indices.
Catalyst is the driver analysis engine. It identifies not just what matters statistically but what would actually move the needle commercially. The Importance versus Impact distinction separates strategic investment from wasted spend. Catalyst is available directly on the Data Alchemy Platform, giving any researcher access to the kind of driver analysis that previously required a dedicated data science engagement.
Segmentation performs outcome-based behavioural clustering. It builds segments defined by what people do and why, rather than who they are demographically. The resulting segments are designed to be analytically meaningful and strategically actionable from the moment they are created.
Synthetic Projection and Imputation are both live as self-service applications on the platform. Projection scales small-sample surveys to population-level datasets, governed by the PRISM validation framework described in the first post in this series. Imputation fills missing data using the same generative approach, with the same five-dimension quality audit. Both are available now, independently of the Alchemist Agent.
These engines do not just coexist. They are designed to chain. The output of Distillation informs Catalyst. The output of Catalyst shapes Segmentation. The Synthetic Data Engine prepares and scales data before it enters the analytical pipeline. The whole is considerably more powerful than the sum of its parts, because the analytical logic flows continuously rather than requiring manual handoffs between stages.
The Alchemist Agent is a channel for accessing those engines autonomously, through conversation rather than self-service configuration. Rather than a researcher manually configuring each engine and passing outputs between steps, the Agent orchestrates the entire workflow from a plain-language brief to a branded strategic output. Integration of Synthetic Projection and Imputation into the Agent workflow is in development.
Track 2: Intelligence Products
The second track is the proprietary IP suite. These are standalone products built on psychological frameworks that took years to develop and validate, and they are genuinely different in kind from the analytical tools in Track 1.
The Resonance Engine measures brand equity through four psychological dimensions: Signal, Fit, Affinity, and Gravity. As I described in the third post in this series, the framework has its intellectual roots in the Conversion Model developed by Jan Hofmeyr and Butch Rice, and has been behaviourally validated across two different commercial contexts, online gaming and grocery retail, more than a decade apart. The cross-category consistency of that validation is what distinguishes the Resonance Engine from brand measurement frameworks that are theoretically interesting but commercially untested.
The Impulse Engine profiles consumer motivation across eight states organised in four dimensional pairs. It grew directly from Honours-level cognitive science research into motivational orientation in high-risk sport, and the central insight from that research, that the same physiological experience is interpreted completely differently depending on dominant motivational predisposition, runs through everything the framework does. The eight states are not a post-hoc clustering of market research data. They are a commercially developed model grounded in a genuine theory of how human motivation is structured.
Resonance Radar extends the Resonance Engine to web-scale brand intelligence, analysing how brands are positioned across the open web and measuring narrative share against competitors. Where the Resonance Engine measures brand equity as consumers experience it, Resonance Radar measures it as the information environment constructs it. Together they provide a complete picture of brand health: what consumers think and feel, and how the broader digital world is shaping those thoughts and feelings.
Why Two Tracks
The two-track architecture reflects a conviction about what research needs to be genuinely valuable, which is two things simultaneously.
It needs to be rigorous. The analytical engines in Track 1 are built on sound statistical methodology, validated benchmarks, and transparent quality frameworks. The PRISM validation standard is an example of the kind of rigour that the industry too rarely applies to its own methods. You should be able to audit the output. You should be able to know what quality means. The analysis should be defensible.
And it needs to be grounded in a theory of human behaviour. The intelligence products in Track 2 are built on frameworks that start from how people actually think, make decisions, and relate to brands, not from statistical patterns in historical data. This is the distinction I described in the previous post as the difference between pattern-matching and modelling. Pattern-matching finds what has been true in the past. Modelling explains why it was true and what that implies for the future.
Most research tools deliver one or the other. The analytical rigour without the behavioural theory produces outputs that are precise but not meaningful. The behavioural theory without the analytical rigour produces outputs that are compelling but not defensible. Knowsis was built to deliver both, because that combination is what produces insight that is simultaneously credible and commercially actionable.
The Convergence Point
The two tracks are not parallel lines. They are converging.
The Alchemist Agent is the convergence point. In its current form, it orchestrates the analytical engines of Track 1 autonomously, from brief to branded output. As the intelligence products of Track 2 are integrated into the platform, the Agent will be able to include brand equity measurement, motivation profiling, and web-scale brand intelligence in the same autonomous workflow.
A research brief that begins "understand what drives our brand equity among our highest-value segments, and identify the motivational profiles that make those segments distinctively responsive to our brand" will produce not just a segmentation and a driver analysis, but a full brand intelligence output. Resonance Engine equity dimensions, Impulse Engine motivational profiles, Catalyst driver prioritisation, and Persona Engine archetypes, all chained autonomously and delivered as a strategic output.
That is the long-term vision: a Research OS that makes the full analytical and intellectual capability of Knowsis available to any researcher, regardless of the size of their team, through a conversation.
The Research OS Category
The IIeX community has been developing the concept of the Insight OS, the idea that research infrastructure should work more like an operating system than a collection of separate tools, with a common layer that different analytical capabilities can plug into, and a conversational interface that makes the whole stack accessible without specialist configuration.
Knowsis is building exactly that. The Data Alchemy Platform is the analytical layer. The intelligence products are the specialist capabilities. The Alchemist Agent is the conversational interface that makes the stack accessible to any researcher.
This is not a niche positioning. It is a response to the structural challenge that the research industry faces: too much data, too little actionable insight, and too wide a gap between the analytical capability that exists and the capability that most research teams can actually access in practice.
The Research OS closes that gap. Not by replacing researchers, but by giving researchers access to the full depth of analytical and psychological intelligence that was previously available only to the largest teams with the deepest specialist resources.
Twenty Years in the Making
I said at the start of this piece that the thinking behind Knowsis had been accumulating for considerably longer than the business has existed.
The Conversion Model training I received from Butch Rice and Jan Hofmeyr in 2004 planted the conviction that brand relationships are psychological phenomena, not statistical ones. The validation work I did across online gaming and grocery retail showed me that psychological measurement could predict commercial behaviour with a precision that purely behavioural models could not match. The rock climbing research I had done before any of that gave me a framework for understanding motivation that turned out to be more commercially useful than I could have imagined at the time.
Twenty years of client-facing work across financial services, FMCG, retail, and consumer credit taught me which questions actually matter when a business is trying to make a decision, and what kind of insight is genuinely useful versus what kind produces a well-received presentation and no change.
The technology has finally caught up with the vision. Autonomous analytical pipelines, generative AI, and large language model orchestration have made it possible to deliver the kind of research infrastructure that I have been building toward for two decades. The Alchemist Agent does not just make the analysis faster. It makes the full depth of what Knowsis offers accessible in a way that was simply not practical before.
That is what two tracks, one vision means. The analytical rigour of Track 1 and the psychological depth of Track 2, converging in an infrastructure that any researcher can use, from the first question to the final boardroom slide.