I recently presented at LEAP in Saudi Arabia, where I spent several days talking with founders, startup teams and people building businesses around emerging technology. Unsurprisingly, AI was everywhere.
Much of the conversation centred on what AI could help people do faster: research that once took hours could be summarised in minutes, first drafts could appear almost instantly, routine analysis could be automated and teams could produce more with fewer delays. The enthusiasm was understandable, because the efficiency gains are real.
What I kept wondering, though, was what happens to workload once individual tasks become faster.
If a report that once took an hour now takes 20 minutes, does that create 40 minutes of genuine capacity, or does it simply create room for two more reports?
That distinction sounds small, but psychologically it is not. If every efficiency gain is immediately absorbed by additional output, the working day does not become easier. It becomes denser.
For startups, this deserves more attention because speed is already embedded in the culture. Teams tend to work across broad roles, priorities shift quickly and people are often expected to absorb new responsibilities as the business grows. AI can make that environment more efficient, while also making it surprisingly easy for expectations to expand without anyone explicitly deciding that the job itself has changed.
A person may retain the same title while significant parts of their role are generated, summarised or analysed by AI, with expectations about turnaround and volume changing almost overnight. The technology may be adopted quickly, but human adjustment rarely works quite that cleanly.
Recent research reflects some of this complexity. A 2026 study involving 541 employees in Chinese technology firms found that greater use of generative AI was associated with both increased confidence about taking on broader responsibilities and increased role ambiguity. Employees could feel more capable while simultaneously becoming less clear about where their role began and ended.
That combination is worth paying attention to. When AI allows someone to produce more, take on more and move faster, it can be tempting to read that as straightforward progress. Yet people generally function better when expectations, autonomy and responsibility remain reasonably clear. When those boundaries become blurred, additional cognitive effort is spent simply trying to work out what the job now requires.
This is particularly relevant in startups because ambiguity is often already part of the environment. Roles are broad, people move between functions and formal job descriptions rarely capture everything someone actually does. AI can increase that flexibility, although it can also make it harder to notice when a role has quietly expanded beyond what was originally expected.
Also Read: SEA’s venture capital shifts from mega-rounds to AI and SaaS
There is also the question of control.
Research published in 2025 examined workers using AI decision support and found that partial AI assistance could support autonomy, competence and meaningfulness, while more complete automation reduced those experiences over time. The implication reaches beyond the specific tasks used in the study because how AI is introduced appears to affect whether people continue to feel that they are exercising judgement or simply supervising output.
That shift can be subtle. A task may become easier while also giving the person less say over how it is done, and over time this can change how much ownership they feel over their work even when the technology itself remains useful.
This is why I would be cautious about treating productivity as the only measure of successful AI adoption. If a team is producing more, but the workday has become more compressed, responsibility less clear and meaningful judgement thinner, then efficiency is only telling part of the story.
Research into employee adoption of generative AI is already showing that people actively reshape their roles around the technology, particularly when AI affects their sense of control and whether their work feels meaningful. That makes AI adoption as much a work-design question as a technology question.
Also Read: The AI productivity paradox: Why finance must move beyond automation
For founders, the practical issue is whether the organisation is consciously redesigning work or simply allowing expectations to expand around the technology.
If AI makes a task faster, what happens to the time that has been saved? Does it create genuine capacity, better thinking, more recovery between cognitively demanding tasks or more space for work that requires human judgement? Or does the organisation simply increase the volume expected from the same person?
There is also a longer-term question about capability. If AI increasingly performs the early thinking involved in a role, organisations need to consider how people will develop the judgement required for more senior work later. Expertise usually develops through repeated exposure to problems, mistakes, uncertainty and decisions, so removing too much of that developmental work may produce efficiencies now while creating different problems further down the track.
None of this requires startups to slow down their adoption of AI. It does require them to pay attention to what happens after the efficiency gain appears.
One of the things I came away from LEAP thinking about was how much energy we are putting into imagining what AI will be capable of doing next. That conversation is moving extraordinarily quickly, but the more immediate question for founders may be much simpler: when AI makes work faster, what are you choosing to do with the time it saves?
—
Editor’s note: e27 aims to foster thought leadership by publishing views from the community. You can also share your perspective by submitting an article, video, podcast, or infographic.
The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of e27.
Join us on WhatsApp, Instagram, Facebook, X, and LinkedIn to stay connected.
The post Your startup has an AI strategy. Does it have a human strategy? appeared first on e27.