Techweek26 AI-at-Work Recap: Why AI's Biggest Impact Isn't Speed
If AI started feeling less like a novelty and more like infrastructure this year, you weren't alone in noticing it. During Techweek26, Enlighten Designs opened its doors for our session: AI at Work, AI Productivity from Experiment to Execution. The room was full, and the conversations before and after were lively.
CEO, Damon Kelly, presenting at Enlighten Designs Techweek26 Event: AI-at-Work
Our objective was to share what we've been learning while using AI inside our software, design and marketing work. Here's what we shared.
Agentic AI is for the challenges you don't know how to solve yet
Most companies are still asking, "what can I automate with AI?" but that may be the wrong question to start with.
Agentic AI is a category of AI that pursues goals by working out the steps itself, rather than running a pre-defined sequence. Automation handles known processes, where A to B is mapped and the tool just runs it. Agentic AI is for the challenges you don't know how to solve yet. You give it a goal, and it helps you think through a path to a solution.
That distinction reframes what agentic AI is for. Instead of speeding up your work, it helps you think differently about the work, the objectives, and how to achieve them. It's the shift from automation to AI-driven development, using AI to shape decisions, not just execute them.
Automation is for the known. Agentic is for the unknown. Most people are still treating AI like a faster way to do automation, and they're leaving the real value on the table.
How AI makes it dangerously fast to build the wrong thing.
One of the most useful lessons we shared was about a project where the obvious solution wasn't the right one. The real problem, surfaced by AI-powered UX research, saved the client a lot of money and delivered a better result.
A client wanted live customer-facing visibility. The instinctive answer was full real-time tracking. It was logical, enticing, and roughly twice the cost of what was actually needed.
The real risk wasn't the cost. The client wouldn't have complained. They got what they thought they wanted. They'd have quietly used it, the budget would have gone on the wrong feature, and the features that mattered would have been quietly cut from scope. Nobody would have ever known.
However, before any code was written, we used AI to analyse a large set of customer support data ... and it surfaced a different story.
The real problem wasn't what people were asking for. It was a downstream problem they were trying to manage. The right solution looked completely different, and cost half as much, because of what we found. Reframing the issue and using AI to analyze the situation changed the entire solution and created a better result for the client.
We used AI for the interview script and synthesis, but humans ran the conversations. The trust-building that made those interviews honest was entirely human. AI for rigour, humans for empathy. That's the working model we keep coming back to.
AI makes it faster to build the wrong thing. The discipline has to be doing the research before the building, because if you skip that, you can spend twice the budget and your customers will quietly use what you ship, and nobody will ever know you missed the real problem.
Design Lead, Adrian Neal, presenting AI in Project Research
Did AI make us more productive? It made our work more predictable, and better.
In November 2025, we began measuring AI's effect on our service delivery, using actual project records rather than self-reported surveys. At our Techweek session, we shared what we've been learning.
AI's biggest measurable effect on our delivery wasn't raw speed, it was consistency. Tasks completed closer to estimate, with fewer blow-outs. For a service business, predictability is arguably more valuable than the "do everything in half the time" story most AI vendors are still selling.
We're also seeing signs that AI lifts the quality of what we do, not just the timing. Time saved on overruns is going back into better test coverage, cleaner reviews, and fewer regressions. Predictability is the first dividend. Better work is the second, and the one that matters more in the long run.
Marketing is also benefiting from AI
The same shift in order of operations is shaping how we approach marketing.
We're about to launch a nurture campaign for Enlighten Aquarium, our Power BI data visualisation tool available on Microsoft Marketplace.
Before we wrote a single word of copy, we used AI to analyse the people downloading Enlighten Aquarium, segmenting them into cohorts based on real user behaviour, industry, and job role. From there, we used AI to draft a story arc, a tone of voice, and a set of fish characters - each carrying a different part of the message. The persona-driven content came last.
The strategy came first. The content came last. What we ended up with isn't a one-off campaign, it's a repeatable playbook we can use across Enlighten's other products and services.
AI can accelerate marketing, but only if you give it a strong strategic foundation first. We didn't write a single word of copy until we'd done the strategy and established the foundations. That order of operations is the difference between content at scale and content that actually works.
CMO, Catherine Lindner, presenting AI in Marketing
Wrap up: Pro tips for using AI in your work
We shared all of this at our Techweek26 session at the Enlighten office. If you're working out how to bring agentic AI into your own business, here are a few final tips we'd pass on, from our company to yours.
- Start with the research, not the build. AI's biggest measurable value lands before code or content gets created, not after.
- Use AI for rigour, humans for empathy. Let AI do the synthesis, the structuring, and the data work. Let humans handle the conversations, the judgement, and the trust-building.
- Don't confuse "agentic" with "unsupervised." Leave permission prompts on, and keep humans in the loop.
- Track what AI actually shifts. Don't be surprised if the benefits show up in places you least expect.
Let's keep the AI conversation going
The next round of our AI Productivity Research is on its way. If you'd like the findings, or you want to talk about what any of this could mean for your business, get in touch with our team.
