Every agency has at least one pathfinder. Most don’t call them that and often the people doing the job don’t know it’s their job.
But if you don’t have one, you should.
A pathfinder is the person who goes ahead and probes what’s out there. New tools, new techniques, new ways of doing the thing you already do. They run the feasibility project, build the proof of concept, get burned by the bad option so the rest of the team doesn’t have to. They come back and say “this works, here’s how” or “we tried it, it’s not ready, here’s why”.
At Filter, this is a core part of my job with one or two others alongside me. We’re the people who spend time on things that aren’t yet billable. We try the new platforms, the new models, the new agent framework, the new build pattern. Sometimes it ends up in a client project a month later. Sometimes it goes nowhere and the lesson is “don’t bother”. Both outcomes are useful. The cost of figuring this out at the start of a real project, with a deadline and a budget attached, is much higher than the cost of figuring it out in a side experiment.
The clearest current example is AI. Everyone is trying to work out what to do with it.
So the question I’d ask is: who at your place is actually using it day to day? Who’s running the experiments, hitting the limits, finding the integrations that save four hours a week and the ones that look great in a demo but fall apart on real work? Who is reading the release notes, watching the new models drop, building little tools to see what’s possible?
If there is no-one doing that, then you don’t have a pathfinder.
The good news is that many agencies do have someone doing this, but they don’t recognise them. And that’s an issue. If your pathfinder has to do this work in their own time, hide it from their line manager, and battle for the credit when it turns into something real, the you don’t lack one. You just aren’t supporting them.
My view is that every agency should give at least one person the latitude and freedom to do this work. Properly. With permission. With time on the calendar that isn’t accounted for against billable hours or sprint commitments. With the expectation that some of what they try won’t work, and that this is the point, not a failure.
If nobody at your agency is probing what’s next, you’re going to find out about the important shifts the way everyone else does: from a customer asking why you don’t already do the thing, or from a competitor announcing it first.
By then you’re catching up rather than choosing. The lead time on getting good at a new tool, technique or platform is longer than people think. The agencies that look like they know what they’re doing in 18 months are the ones fiddling with the awkward early version of the thing today.
Of course there are downsides.
It can look and feel like wasted time, and for agencies, billing hours is key. To anyone watching the timesheets, a fortnight spent on a proof of concept that gets shelved looks indistinguishable from a fortnight spent on nothing.
So you need a leadership team that understands the difference, or a culture that protects the work, or both. I’m lucky at Filter that we can hold this space ourselves. Smaller teams without that protection often lose their pathfinders to bigger agencies that will pay them to keep doing it.
You also need someone you trust to make the right decisions, and this is harder than it sounds. Look at their history of recommendations: what did they get right, what did they get wrong? Are they broadly in line with what actually happened? Finding the right person is really tough.
But it’s more than likely that you already have one and just haven’t recognised them yet. So look for the person who keeps bringing up new things in stand-ups, who’s got a side project in their personal time that’s suspiciously close to your business, and who answers AI questions in a way that makes you realise they’ve been further into it than you have.
Give that person more flexibility and leeway and see where they get to. And if they’re not right, then find someone else. The key thing is that every agency needs a pathfinder.
Who’s yours?