How much time are LLMs saving Engineering & Software Development Firms during Design & Development?
Poll results and a few opinions from an engineer turned marketer
Yeah, that’s right, I’m an engineer… at least I used to be. Though I’ve been told many times: “once an engineer, always an engineer”. I suppose at some level that’s true. I certainly mostly still think like an engineer, for better, AND for worse 😊.
I thought it would be useful to gather some feedback on how much of an impact LLMs are having on engineering and custom software companies on their actual day to day dev work.
So I created a poll and put it up on RocLogic’s website.
The question I asked:
“Engineering & Software Dev leaders: How much time do you think LLMs are saving you during your design and development process?”
Response options offered:
- I think it’s actually costing us additional time.
- ~0%
- < 10%
- 10-20%
- 20-30%
- 30-50%
- >50%
The raw data from the poll:
Total number of responses = 50.
The poll ran from ~November 2025 – August 2026.
A note worth mentioning: take this informal poll with a grain of salt. I’m inferring that the majority of the people that submitted poll responses to this poll are in the engineering / software dev world based on the reality that the vast majority of RocLogic’s website is very focused on helping engineering and software dev companies with their marketing, along with the framing of the question itself. In addition, the responses from this poll are most likely geared toward companies that develop a lot of software (whether they label themselves as engineering or software dev companies). I believe that to be the case mainly because I placed this poll on pages that are geared specifically toward software development companies.
Poll Results – saving time during the engineering and software design and development process

The time savings appear significant
The median response is 20-30% savings, while the most common answer was 30-50% savings.
Overall, a whopping 68% of responders reported 20% or more improvement in time savings during the design & development process! And 22% of responders indicated they’re seeing >50% improvement in their design and dev process.

That’s an enticing value prop. I think that any new tech that can consistently save you 20% or more during the design and dev process is worth considering.
To be fair, the people that are more likely to try to leverage LLMs are those that had reason to believe they could see significant savings or else those companies probably wouldn’t have initiated the experiment.
But not everyone is seeing savings

20% of responders reported that they’re seeing at most a negligible amount of improvement, i.e.:
- those that saw <10%,
- ~0%,
- or those where it’s actually costing additional time.
I’m interested in understanding the “why” behind this:
- is it just working through updated process kinks?
- Does this group develop applications that somehow don’t lend themselves to leveraging LLMs?
- Are they finding too many bugs?
10% of responders suggested that leveraging LLMs is actually taking them MORE time during design and development! On first blush that may seem surprising, but after sitting with it for a moment, this makes sense, especially if you’re in the first couple iterations of evolving your design & dev process:
- learning edge cases,
- what new risks to worry about,
- what guardrails to put in place,
- how much oversight to provide,
- what new unit level and integration testing to perform and how.
In 2026, my sense is that most companies that existed before LLMs took center stage are still trying to figure out how best to utilize AI, so I think it’s fair to assume that this is a useful snapshot, but the time savings will likely evolve and converge to some median savings over time.
In case you’re wondering why RocLogic would put the time and energy into creating such a poll. A few reasons:
- RocLogic is laser-focused on marketing for engineering and software dev companies, so understanding where my audience’s head is at is very important to me.
- It’s a way that I felt like I could provide something useful to my audience. I felt like my audience would be interested in seeing how their peers are adopting AI.
- The internet thrives off new information. This in-turn helps RocLogic’s exposure from a marketing perspective.
Next Steps
In data mode? Here are the results of a couple more polls:
- 5 roadblocks that stop tech services companies from moving forward with marketing
- Niche selection obstacles
In learning mode? Check these out:
Poll results table and pie chart in more “raw” form (if that interests you)
A table of the data:
| Reported Time Savings | % |
|---|---|
| ~0% | 4% |
| < 10% | 6% |
| >50% | 22% |
| 10-20% | 12% |
| 20-30% | 20% |
| 30-50% | 26% |
| I think it’s actually costing us additional time. | 10% |
A raw pie chart:
