Faster, Better, Cheaper. If Only.

There’s a saying in marketing that you can have faster, better or cheaper,  but only two of the three at any one time. It’s a neat way of describing the trade-offs we all face when delivering work. Over the last year or two, AI has been widely presented as the first technology capable of breaking this rule. Faster? Often. Cheaper? Frequently. Better? Well… that depends.

The reality is more nuanced. AI can certainly be faster when it produces what you actually need, but speed has limited value if time is then spent validating, correcting or reworking the output. It is often cheaper, but not universally so,  particularly once quality checks are taken into account. And whether it is better depends entirely on whether the result improves on what you or your team could realistically produce within the constraints of a real project. The promise is attractive, but outcomes vary considerably depending on how and where AI is applied.

In market research, this tension is especially visible. Our industry has always involved a significant amount of manual work: designing questionnaires, managing fieldwork, preparing and checking data, producing analysis, and turning results into something coherent and meaningful. Each stage requires care, judgement and experience,  and each stage introduces opportunities for inefficiency or error. AI can help, but only when it strengthens processes rather than weakening thinking or clarity.

For users of MRDCL, the biggest demands on time are rarely the mechanics of writing code. They lie elsewhere: preparing analysis structures, managing and reshaping data, checking outputs, correcting errors and refining workflows. MRDCL has always been a powerful scripting environment, but power alone does not guarantee efficiency. What makes a substantial difference is how that power is applied, the workflows chosen, the order of operations, and the ability to adapt processes to suit different projects.

This is where flexibility really matters. Poor workflows don’t just make analysis 10% or 20% slower; they can multiply effort, introduce risk and make projects far harder to manage than they need to be. MRDCL gives users far greater freedom to design workflows that suit their data, their teams and their delivery requirements, and that freedom can make an enormous difference to speed, accuracy and repeatability.

At Brown & Blue Circles, we want to build on this strength. Our aim is for MRDCL users to be able to prepare analysis more quickly, with fewer errors, and with greater confidence that data will flow cleanly to wherever it needs to go,  whether that is another platform, a reporting system, a client environment or an internal application. If MRDCL is to sit at the centre of a modern insights ecosystem, it must reduce friction and remove avoidable work without limiting choice.

That is a high bar to set. And it is important to be clear: we have no interest in faster-and-cheaper solutions if they come at the expense of quality. That approach does not serve clients, analysts or the reputation of the industry. Market research cannot afford to become a button-pressing exercise that produces superficial outputs. Logic, structure and detail still matter, and they always will.

So what are we working towards?

• Automation that removes friction rather than hiding logic, making complex processes smoother while remaining transparent.
• Agentic AI that helps with difficult or error-prone tasks, particularly those that are time-consuming or repetitive.
• AI-assisted workflows that reduce manual effort, allowing analysts to spend more time thinking and less time managing mechanics.
• Automated outputs that are accurate, consistent and ready for the next step, whether that is tables, charts, data structures or integrations.
• The ability to create tools and processes that can be repeated or shared across teams, improving consistency and scalability.
• Clean, reliable data pipelines, enabling MRDCL to connect intelligently with the tools and systems that matter to your organisation.

Will all of this be fully realised by the end of 2026? No, and that would be neither realistic nor desirable. But the quieter work of the last three years has laid the foundations for a very different MRDCL experience. We are now in a position to make meaningful advances during 2026 and to build a platform that supports the next era of data processing, data management, analysis and reporting.

Which brings me back to the faster–better–cheaper consideration. The aim is not to break the rule for its own sake. It is to improve those parts of the workflow where speed and cost genuinely matter, while strengthening the processes that underpin quality and flexibility. Used well, AI becomes part of that system,  not the centre of it, with MRDCL providing a robust, structured core around which better workflows can be built.

We are entering a period where automation and intelligent assistance will reshape how insights work is done. The challenge is to use those tools in a way that reinforces rigour, supports expertise and improves outcomes, rather than chasing shortcuts that offer short-term gains at long-term cost.

Next month, I’ll look at why many research teams end up wasting money on software, and what to consider instead when choosing tools that are meant to last.