-By Raghavendra.K.A SVP and Global Head of Engineering – Integrated Product Development, Engineering Systems & IoT
Gestational age, or how far along a pregnancy is, sets the timeline for prenatal care. In a University of Oxford study, clinicians estimated gestational age from ultrasound images. Supported by an estimate from an AI model, their average error fell from 23.5 days to 15.7 days. On accuracy, which was the measure that mattered, the technology worked.
Adding visual explanations of how the model reached its answer improved accuracy marginally further. Those explanations raised clinicians’ confidence, but had no significant effect on how much they trusted or relied on the system.
That finding about trust is the more instructive one. Being right is often not the same as being acted on. What decides the difference is how a product presents itself, not just what it calculates. The more capable a product becomes, the more its value rests on a boundary that engineering has only recently begun to design for.
Such analyses offer engineering leaders actionable insights about the human factor that can drive product success.
Move beyond functionality
Giving center stage to usability, safety, and human experience is the work of Human Factors Engineering (HFE), a discipline built on the premise that a system’s design has to account for human cognition, not assume it away. An explanation that is technically complete but adds to cognitive load for the person has failed.
The purpose of a product is to help a person trust a system when it is right and question it when it’s not. A product can be accurate and still be unsafe if the person using it cannot tell which situation they are in.
Design for intuitive interactions
HFE and experience-led design are essential in industries like healthcare, where outcomes depend on how intuitively humans interact with technology. In particular, healthcare shows the problem early because how a person responds to information, is part of the result itself. More than 104 crore health records are now linked to over 93 crore Ayushman Bharat Health Accounts, according to the Ministry of Health and Family Welfare. At that scale, the outcome of an interaction compounds across millions of encounters.
India’s public health programs increasingly run on digital portals. Their value depends on how easily frontline staff can use them through the course of a busy work day. Training, workflow fit, and systems that connect to one another decide whether a platform serves patients, staff, and administrators and is a reliable, technical foundation that supports the needs of good healthcare programs.
Regulation is moving in the same direction. In July 2026, India’s Central Drugs Standard Control Organisation (CDSCO) issued its guidance on medical device software, with expectations covering usability validation, risk management, AI bias assessment, and clinical performance evaluation. This indicates that software is now assessed on how it behaves in use, in addition to what it computes.
The pattern extends beyond healthcare. The Telecom Regulatory Authority of India (TRAI) issued a direction in February 2026 requiring operators to act on AI-flagged spam rather than only alert subscribers to it, having observed that alerts alone had not acted as a deterrent. The direction specifies the wording of the notice that a flagged sender receives.
Break out of silos for success
The strongest products come from three disciplines working as one team. Design, engineering, and behavioral insight usually sit apart, and aligning them is a leadership decision before being a delivery one.
A recent Infosys Knowledge Institute research on human-agent collaboration sets out what teams must be able to do: validate outputs, detect drift, and recognize when a decision belongs with a person. Whether a person can do any of that largely depends on what the system chooses to reveal, and that is decided at the architecture level.
Understanding of the system is key
Measurement follows the same logic. Override rates and time to verify belong alongside latency and uptime; offline capture, regional-language input, and legibility on entry-level devices decide whether a robust product is usable by the people it serves.
Experience-centric engineering brings the same rigor to a person’s understanding of a system that engineering has always brought to the system itself. Performance and comprehension are the same measure, seen from two sides. This understanding is something the product is built to produce.