I recently spent three days at Collaborate Innovate 2026 in Perth, in rooms full of smart, motivated people who genuinely care about Australia's R&D future. The strategy conversations were excellent and the optimism was real, but I kept feeling like something was missing.
A post-conference reflection circulating this week named it well: almost every challenge discussed at the conference ultimately came down to invisible labour. Relationship building, translation between worlds, listening, synthesising across disciplines, managing expectations, repairing trust when it breaks.
"Innovation moves at the speed of trust" was the line that kept circling the room, and it landed every time. But no one really explored it with any specificity. That specificity is a job, several jobs, and the people doing them are largely invisible.
The translation layer
Australia's R&D ecosystem has a structural gap between the people setting strategy and the people executing it. The strategy conversation happens at the top: CEOs, chairs, directors, policy leaders. Conferences like CI26 reflect that naturally, and those conversations are important.
But the actual work of research translation happens in the middle, with the project managers, partnership leads, research development staff, commercialisation managers, communicators, coordinators, and ecosystem builders. These are the people making complex science legible to an industry partner who has 20 minutes to decide whether to commit. They're the ones holding the thread when a researcher and an industry lead have interpreted a project scope in two completely different ways, building trust before the ink is dry on a collaboration agreement, and doing the quiet repair work when things go sideways. It's skilled, difficult work, and it is chronically undervalued.
Why it stays invisible
Part of the problem is prestige. The translation layer doesn't fit neatly into traditional academic or industry hierarchies. There are no obvious career ladders, no formal credentials, no consistent job titles, and the people doing this work are often embedded in institutions that don't quite know what to do with them.
Part of it is that the work only becomes visible when it fails. When a collaboration breaks down, when an application misses the mark, when an industry partner walks away frustrated, people notice the gap. When it's working, it looks effortless, which is rather the point.
And then there's the "soft skills" problem. Active listening, synthesis, relationship intelligence, expectation management, the ability to read a room and translate across registers: these get called soft because they're hard to quantify, not because they're easy to execute. They're genuinely difficult to learn, difficult to teach, and difficult to perform well under pressure, and the label does a lot of damage to how the sector values the people who have them.
There's a gender dimension here that the sector rarely talks about directly. Australia's R&D ecosystem skews male at the top: the named researchers, the CEOs, the chairs, the plenary speakers. The translation layer skews female, with the project managers, RD officers, communicators, and coordinators. That's not a coincidence, and it's not unrelated to why this work is undervalued. There is a long and well-documented history of work predominantly done by women being classified as intuitive, relational, or "just part of the job" rather than skilled and strategic. A session at CI26 on women in leadership made this explicit, with participants naming invisible work directly: the expectation of flexibility, the filling of gaps, the keeping of things moving without acknowledgement or formal authority. That conversation deserves a much bigger room, and the R&D sector is not exempt from the pattern it was describing.
AI has not solved this. It has raised the bar.
There is an assumption creeping into the sector that AI will absorb the translation work: that relationship management, stakeholder communication, and synthesis will become easier, faster, cheaper, and that the humans in the middle layer will need to do less. The evidence points the other way.
AI is already a baseline expectation for anyone working in research translation. Grant writers, R&D staff, comms people who aren't using it are slower than their peers, and that bar will keep rising. But what AI cannot do is read the room. It cannot sense when a collaboration is starting to fracture, navigate the political dynamics of a multi-partner project, or translate between a vice-chancellor's priorities and a startup's risk appetite. It cannot repair a working relationship that broke down over email, and it cannot build trust.
The effect of AI in this space is to automate the more mechanical parts of the work, which means what remains is harder, more relational, and more dependent on exactly the skills the sector has been undervaluing. The humans in the translation layer aren't becoming less necessary. They're becoming more so.
What the sector needs to reckon with
The Australian R&D ecosystem is ambitious. We want sovereign capability, deep industry-research collaboration, genuine commercialisation, and technology adoption at scale, and none of that happens without the translation layer functioning well.
If we're serious about innovation moving at the speed of trust, we need to get serious about the people building that trust. That means recognising this work as expertise rather than just experience, investing in the skills involved because they are learnable and teachable and worth improving, building career pathways that don't require people to move into management to feel like they're progressing, and making the work visible in funding structures, in organisational design, and in the conversations we have at conferences.
The strategy is solid and the policy intent is there. The gap is in execution, and execution is a people question.
At Straight Up, we work in this space every day: grant writers, science communicators, and project strategists at the interface between researchers, industry partners, and funding bodies. We know what the translation work involves because we do it, and we think the sector is ready for an honest conversation about what excellence looks like here.
Hellie