Any procurement discussion today in the patient voice or outcomes research space is focused on two words: time and cost. These are the jaws of the modern project vice. Between them sits the fragile balloon labelled ‘Project’. It is inflated with expectations, timelines, lowering of costs, KPIs and templated methods. The vice tightens: they want it done faster and cheaper. Something has to give.

You never know which part of the balloon bursts first. Sometimes it is not directly related to the reduced budget or the tighter timeline. Often, it is the room for nuance and rigour that ruptures first. This is the breathing space that is required for actual thinking. In other words, quality. Research, increasingly, is squeezed until it becomes indistinguishable from compliance. The stuff of the checkbox, of painting-by-numbers.

One reason why this happens is that standardisation, on a surface level, looks like quality. A well-documented method, a neat deck of slides and a compliant report all suggest competence. In practice, though, standardisation can also mask homogeneity. Everyone says they do cognitive debriefing or usability testing or regulatory affairs. Everybody has a methods slide. Everybody has their AI-generated ‘how-to’ blog article. Everybody is fluent in regulatory-speak.

AI is now part of this vice. Used unconsciously, it accelerates sameness. It mass- produces polished outputs that look like thinking but bypass the actual process of it. It flattens difference. It mimics insight. It can become a turbocharged engine for compliance. But used with intention, AI can be a pressure release valve. It can clear the deck of more labour-intensive activities – grunt work – such as transcriptions, formatting, box-ticking and give researchers back the space to work better, ask better questions and reach the human heart of things. But this is only of benefit if we slow down enough to notice.

As Thomas Hübl has observed, the issue is not just speed, which is entirely manageable, but acceleration. By this, he means the continuous ratcheting up of speed over time. It is acceleration, not motion, that creates whiplash. It is not that we are moving fast but that we are never allowed to stop accelerating. And the human system, just like the research system, cannot integrate information at that pace.

Procurement teams, under pressure themselves, may find it hard to distinguish between true research and the checkbox delivery. What if the checkbox is all they are looking for? What if they prefer to do business with the salesperson they like best? This could be the only differentiator: who schmoozes the best.

The result? A market where sameness is rewarded. If the output looks the same, time and cost become the only meaningful variables: who can do it faster and who can do it cheaper.

And so begins the downward spiral: lower costs, shorter timelines, tighter margins. The turning of the screw, the jaws of the vice tighten. Eventually it feels like we are reduced to acting like rats fighting over a pizza slice in a Manhattan dumpster.

Frozen methods and flattened outputs

As detailed in the previous article, when methodologies become codified and uncritically adopted, they cease to be useful tools of inquiry. They become frozen rituals. The same templates are used. By everybody. Flair and insight are quietly engineered out. All the original thinkers are retired now. Who holds the fire now? It’s not the priesthoods or the cartels.

In time, even genuine research-based companies start to conform. They do what the RFP asks. No thrills. They have to mimic the formats that win. They stop pushing. They stop being the preacher in the desert, because this takes time and time does not fit the budget. So, all that additional energy to capture more of the patient voice, usually done as a gesture for free, for the sake of completeness, are all dispensed with.

Here too, AI can act either as a ritual amplifier or a ritual breaker. It can harden templates, automate flattening and perpetuate the ‘just enough” that passes for “good”. On the other hand, it could also offer provocations and lateral connections, if researchers are allowed the time and scope to slow down and think, to listen and not just to label.

The Iron Triangle

Project management theory warns about the dangers of fixed constraints. This is the classic ‘Iron Triangle’ of Time, Cost, Scope. You cannot compress all three without compromising something.

In reality:

·       If you compress time and cost, scope shrinks.

·       If you keep cost low and scope high? Time extends

·       If you want scope on a tight timeline? Cost soars.

However, in health research, all three are too often locked in from the outset. Timelines are non-negotiable. Costs are capped. Scope is set in stone. Something has to give and that is usually quality or depth.

All projects operate under constraints. Pressure is not the problem. Compression is. When time and cost are treated as something that can be shrunken and every project is squeezed through the same funnel, a research project can lose its oxygen. And a vacuum does not engender innovation.

This is the logic of acceleration. We build systems that no longer allow time for digestion, pause or reflection. We reward the shallow quick win over the difficult long view. And so even the most powerful tools, like AI, become just another jaw in the vice, unless we reclaim the right to breathe.

When critiquing the Iron Triangle through lens of 2025, some might argue that there are conditions where the balloon does not have to burst. You can loosen the vice of time and cost by moving the work elsewhere, to regions with low living costs, cheaper labour and increased reliance on AI. Could a project be saved this way? Could the slow reflective researcher be replaced by a team overseas, assisted by AI, working around the clock and well within budget?

This logic can seem, at first glance, compelling. You save money. You meet deadlines. You manage scope. But this model carries hidden costs and they show up in the very places we claim to care most about. In Patient Voice Research, what gets outsourced is not just transcription or formatting, but it is context too, as well as cultural understanding and the ability to interpret a patient’s behaviours or attitudes adopted in the interview. These can be easily overlooked, but they constitute the real language of illness. AI, trained on biased data and shaped by dominant norms, cannot do that. Nor can the human who is detached from the communities being studied, reduced to a vendor working against time zones and margin.

And then there is bias, which will inevitably show up in analysis, translation and storytelling. Offshoring and automating too much of the work just pushes these biases and distortions further downstream, where they harden into research outputs, clinical decisions and policy assumptions. The balloon may not visibly burst, but it leaks. It leaks trust, meaning and validity.

So, yes, we can replace the “slow human” in theory. However, in doing so, we risk removing the very ingredient that makes patient voice work meaningful: the capacity to care, to understand and to respond. That cannot be mimicked by algorithms or managed by a procurement team. The solution to the Iron Triangle should not be to ‘squeeze harder’ elsewhere and through automation. It should be to resist acceleration, to create just enough room for reflection, relationship and depth.

If we want patient voice work to remain meaningful, we need to loosen the grip on the vice. It does not have to be much: just enough to let the balloon breathe. Maybe just enough to slow down speed and interrupt acceleration. Maybe enough to invite the human back into the method. A research system that prioritises speed and cost above all else will eventually collapse under its own superficiality. What will be left will not be innovation, but sameness, disguised as progress.

A burst balloon is no good to anybody. And neither is shoddy research.

Thank you for reading,

Mark Gibson, Leeds, United Kingdom, July 2025

Nur Ferrante Morales, Bonilla de la Sierra, Ávila, Spain, July 2025