2025 was the year AI shifted from being an experimental tool to becoming a core part of how we work. It appeared across design, engineering, strategy, research, and even the small tasks that fill a workday. Not because it solved everything, but because, when guided thoughtfully, it accelerated work and expanded what was possible.
Across disciplines, AI helped teams explore more options earlier, test ideas faster, and reach alignment sooner. It also revealed its limits, showing that human expertise, judgment, and restraint remain essential to delivering quality work.
To understand the real impact, we spoke with two Twisthink teammates: Kyle Krueger, Senior Software Engineer, and Kent Pilcher, Principal Designer. Their reflections reveal where AI amplified their work, where it fell short, and what it means for how we build products moving forward.
Engineering: AI as a Reliable “First Draft” Partner
For Kyle, this year marked a noticeable evolution in daily development.
“AI became a reliable first-draft partner. Most of the team now uses it for coding, documentation, reviews, and tests. It helps us move faster, but we still rely on our own experience to make the right calls.”
AI excelled at the tasks that bog engineers down—scripts, boilerplate code, documentation, and test coverage. Offloading these tasks created more space for the work that truly requires depth: thoughtful architecture, system-level thinking, and debugging the gray areas where human judgment is essential.
But it wasn’t frictionless.
“Sometimes it wanders. It tends to over-engineer things or misread intent. Keeping prompts small and focused helps a lot.”
One example stood out from this past year:
“AI helped me build Azure Infrastructure-as-Code validation and deployment scripts. Within a few iterations, it produced something production-ready. That was a huge time-saver.”
For Kyle, the takeaway is clear:
“AI is a force multiplier, not a replacement. It accelerates the work, but we still have to guide it and verify the results.”
Design: AI as a Driver of Broad Exploration
In design, AI didn’t just speed up tasks, it widened the horizon of what could be explored.
“AI is now my primary divergent exploration tool,” Kent shared.
Where designers once manually produced dozens of early concepts, AI could generate broad aesthetic and functional variations in minutes. It also synthesized research faster, turning interview transcripts, articles, and raw inputs into themes and hypotheses to investigate.
This helped teams align more quickly:
“I used AI to create a spectrum of diverse product concepts. This visual range helped the team achieve faster strategic alignment on project goals and innovation levels, converting abstract debates into concrete decisions.”
But the speed came with a cost.
“AI generates such a high volume of text that the human effort required to sift, process, and curate for project relevance can negate the initial time savings.”
And Kent emphasized the risk of misplaced confidence:
“AI can create an illusion of validated accuracy. Its output must be strictly treated as a starting hypothesis for real-world research, not a replacement for it.”
One of his reflections captured the right balance:
“AI is an amplification tool, not a license to cut corners.”
Principles That Now Guide Our Approach
Kyle’s and Kent’s experiences surfaced a set of shared principles now shaping how Twisthink integrates AI across disciplines:
- Use AI to inform decisions, not make them.
AI is a strong starting point, but accountability remains human.
- Keep the bar high.
More output doesn’t equal better output. Quality still requires curation and refinement.
- Prioritize clarity over complexity.
If AI introduces noise, it isn’t adding value.
- Build new skills intentionally.
Prompting, editing, evaluating, and integrating AI outputs are now core competencies.
- Stay grounded in human-centered thinking.
Real human needs matter more than any model’s predictions or assumptions.
These principles reflect what the team learned repeatedly: AI changes the pace of work, but people still determine the quality of it.
“AI has become a real asset, but human expertise is still what drives quality,” says Kyle. “The magic happens when we combine our skills with AI’s speed.”
In Summary
Across engineering and design, AI didn’t diminish the importance of expertise—it heightened it. It handled the repetitive and the expansive so that Kyle and Kent could focus on the thoughtful, the nuanced, and the human.
The consistent message in their reflections is simple:
AI is a multiplier of good thinking, not a replacement for it.
It accelerates exploration, supports decision-making, and opens new creative and technical possibilities, but only when paired with the judgment, clarity, and human-centered mindset that define great work.
Twisthink is a team of designers + technologists that partners with companies to accelerate the design and development of connected solutions.
