Breaking through the "The Trough of Mediocrity"

Breaking through the "The Trough of Mediocrity"

The Trough of Mediocrity

As we see the furore around no-code AI tools and how we can now apparently build Enterprise grade apps with a couple of tools and some prompting, it’s seeming to settle into a pattern of shallow thinking and inievtiably leading product and product teams into a trough of mediocrity - where a 'good enough' experience can have a bad habit of settling.

The Natural Settling Point

The Trough of Mediocrity isn't new. It's where projects and products naturally settle when faced with competing needs, competing teams, time crunches, and the accumulated weight of technical and design debt. For as long as I've been working in this industry, there's been a gravitational pull toward "good enough."

How many products have you used that work fine, look professional, and solve the basic problem, but leave you feeling nothing? They're not bad enough to abandon, but not good enough to love. They've settled in the trough.

Design systems were our first attempt to elevate things systematically. Now GenAI is promising to revolutionize how we create products, offering "Enterprise apps in minutes" and designs that look ready to ship. But here's what we're missing: these tools are making it easier than ever to fall into the trough, and harder than ever to climb out.

The Facsimile Problem

While everyone is (rightly) excited about the new tools that can generate impressive prototypes instantly, we're creating something fundamentally hollow: the facsimile of design and product.

The GenAI-generated proof of concept looks amazing. It's coded, designed, and appears "ready to launch." As these tools get more sophisticated and can better pull from existing design systems, they increasingly look like products that could ship under a company's brand. Done in an hour, off to the pub for the rest of the day.

But underneath, it's just a facade.

The nuance around how people actually use products is missing. The emotional investment users might have (okay, maybe not in Enterprise SaaS, though I can get strangely passionate about email applications) simply isn't there. The tools are amazing and transformative, no doubt—but they're just tools. They’re really eager to please, and they'll generally look for the happy path.

Crap In, Crap Out. At Scale

But LLMs aren't magic boxes. They'll output at whatever level you input. If you're coming with surface knowledge, that's what you'll get back.

But think about your experience - all the projects you've worked on, the teams you've been part of, the users you've interviewed over your career. Think of the common pitfalls you've seen time and time again, the moments of delight you've witnessed, how you've had to juggle user needs against business needs and navigate those trade-offs.

You are a human LLM, and that model is far more attuned to the needs of users, products, and businesses. You're aware of nuance, can think through edge cases, and won't be content with only creating the happy path.

But if we don't bring that expertise to bear, if we don't elevate the input going into these tools, then countless products will settle into the trough. Everything will start to look and feel the same.

AI as Force Multiplier, Not Replacement

We use AI tools daily at Comedia. I love how Claude helps me synthesise information quickly and shapes my writing. But once it reaches a certain point, our Head of Content reviews it and makes it way better. We see AI as a force multiplier for human expertise.

The AI models are like an extremely keen, extremely confident recent graduate suffering from Dunning-Kruger syndrome. They can be confidently wrong but will happily pivot when called out. They're amazing force multipliers and valuable team additions, but without expertise going in, work will always slump and settle in the trough.

The RV Proof Point

We started as an innovation studio with a core promise: we were always "makers of the thing." In incredibly short timeframes, we could take complex information, product needs, and technical requirements to create working prototypes along defined narrative paths.

Take our pre-GenAI electric RV project: 10 weeks from standing start to functioning proof of concept mobile application that worked along the narrative path that our RV client could demonstrate at an RV tradeshow. These days AI tools help us move faster, but human expertise makes the critical decisions, understanding user context, anticipating edge cases, and crafting experiences that feel intentional rather than generated.

Now we could have created multiple PoCs and tested those to have an even more defined end application, but only because of our team's experience going into it.

The Choice Ahead

AI democratizes creation like never before, making it easier to produce "good enough" results at unprecedented speed. But this same capability creates a massive opportunity for those willing to climb out of the trough.

While everyone else settles for AI's confident mediocrity, companies that use AI to amplify human expertise, not replace it, will create products that people actually care about. Products with soul, nuance, and that indefinable quality that makes users think "someone really thought about this."

The question isn't whether AI will replace designers, developers, or strategists. The question is: Will you choose excellence, or will you settle in the trough?

Because in a world where anyone can generate the facsimile of a great product, actually creating one becomes the ultimate differentiator.