The core hurdle in manufacturing digitalisation is that factory conditions are rarely predictable

Anubhi Khandelwal, Co-Founder, Terafac.

What exactly is the concept of ‘Physical Intelligence’ that Terafac uses enabling robots to see, adapt, and make decisions without programming? Is this vendor neutral?

‘Physical Intelligence is to robots what cloud was to computing, a common layer that unlocks scale.’

It is the idea that machines should not just follow coded instructions but develop a physical intuition to perceive, reason, and act in dynamic environments. Globally, it’s seen as the next step beyond rigid automation, where robots evolve into adaptive agents.

At Terafac, we combine AI-enabled perception with adaptive real-time control. Instead of pre-written programs, our vision models interpret the environment, detect variations, and generate adaptive actions. Unlike traditional robots, our intelligence layer enables robots to:

● See – use sensors to detect weld seams, parts, and obstructions.

● Adapt – adjust paths if parts shift or defects occur.

● Decide – use AI to choose the best corrective action.

This allows robots to weld, paint, or glue with human-like intuition, even under variability.

Our first product, Weld-T, solves for welding. It gives robots the ability to adapt and weld consistently without reprogramming, designed for real shop floors where misalignments and tolerances vary.

Importantly, our system is vendor-neutral. The intelligence sits as a smart layer over any robot, brand, or power source, ensuring manufacturers can scale without being locked into one ecosystem.

How does this AI-enabled platform uniquely combine robotics and machine vision to address key pain points in manufacturing digitalisation?

Robots have long been general-purpose actuators, but scalable automation never took off due to heavy programming needs. Terafac’s AI-Vision system changes this by turning perception directly into motion intelligence, removing dependence on coding.

The core hurdle in manufacturing digitalisation is that factory conditions are rarely predictable: parts shift, designs change, humans add inconsistencies. Rigid automation collapses under such variability, worsened by:

● High-mix production – frequent product changes.

● Skilled labour shortages – bottlenecks in scaling.

● Programmer dependency – every change needs coding.

● Scaling failures – pilots don’t roll out widely.

Our Software-as-a-Skill platform fuses vision with robotics to create adaptive intelligence. Robots perceive, learn, and adapt like skilled craftsmen, handling imperfections without new programming.

The system is vendor-neutral, it works across different robot brands, and shop-floor setups. Adoption is scalable and investment friendly: start with one robot and one skill (e.g., welding), then expand to gluing, painting, or buffing without system redesigns.

In short, Terafac makes automation practical, scalable, and accessible. It solves today’s pain points while future-proofing factories for tomorrow.

What are the main challenges in integrating AI, robotics, and machine vision into existing manufacturing setups, and how does Terafac overcome them?

Factories were built for repetition, not intelligence. Legacy machines don’t connect, rigid robots fail under variability, and upgrades demand expensive programmers. Key challenges include:

● Fragmented infrastructure – heterogeneous robots and power sources.

● High integration costs – custom coding and redesigns.

● Unpredictable shop-floor conditions – variability breaks flows.

● Skill gaps – lack of skilled labour and programmers.

● Uncertain RoI – heavy investments without clear outcomes.

Terafac flips the script. Our intelligence layer adapts to factories as they are. We plug into existing robots and power sources without new hardware. Machine vision enables adaptability, while AI removes programming needs.

Adoption is cost effective and progressive: start with one robot, one skill, then scale across lines and plants, showing RoI at each stage. Being vendor-neutral, it integrates across different setups seamlessly.

We turn integration from a costly overhaul into a layered upgrade, making intelligent automation accessible and scalable.

How adaptable is your platform across different manufacturing segments, from small-scale units to large, highly automated plants?

Manufacturing is diverse: a 20-person shop struggles with labour and costs, while a 2,000 robot plant struggles with scaling and consistency.

Terafac works across both ends.

● For MSMEs: Start small with one robot and one skill (welding, buffing, gluing). No programmers or heavy investment needed. Our system works standalone, delivering RoI from day one. Small shops can access automation once reserved for OEMs.

● For large plants: Our vendor-neutral layer scales across robot brands, lines, and plants. Robots adapt to variability, solving bottlenecks in high-volume production. Plants move from fragmented automation to a unified adaptive system.

Most importantly, the same framework works globally, from fabrication shops in India to automotive lines in Europe or aerospace in the US.

This is not incremental improvement but a step-change in accessibility, much like the iPhone moment for manufacturing: making sophisticated technology accessible to everyone, everywhere.

How does your system leverage production data to drive real-time decision-making and process optimisation?

We treat production data not as records but as context for action. Each weld, bead, or pass contains insights, but traditional automation wastes them.

Our platform captures and acts on this data in real time, creating self-optimising feedback loops. For example, Weld-T collects data on geometry, alignment, and process parameters:

● If joints are misaligned, it corrects paths.

● If heat input is high, it adjusts speed or voltage.

Over time, these corrections reduce rework, improve quality, and cut downtime.

This extends to gluing, buffing, and painting: thickness detection, surface pressure, or spray patterns feed into decision-making. Each cycle adds to a data spiral where the intelligence layer grows sharper over time.

Automation doesn’t just run; it learns and compounds with every job.

What measurable improvements – such as productivity gains, defect reduction, or cost savings – have your clients reported after implementing Terafac’s solution?

We are still early in pilots, but results are promising:

● Up to 3x productivity compared to manual welding.

● Greater consistency and reduced rework.

● Faster setup compared to programming-heavy automation.

One customer even shared that with this flexibility, they could set up a dedicated welding shop to drive additional topline revenue.

We are now deploying across industries from heavy fabrication to auto-ancillaries and large OEMs. Early results show potential for a fundamental change in manufacturing, delivering productivity, new business models, and wider access to intelligent automation.

Women representation is growing in the field of hardcore tech sectors like manufacturing in the digital era. What is your experience from industry and fellow innovators/entrepreneurs?

Early in my career, I delivered a workshop on digital twinning in the UK. Facing a room full of senior male engineers, I felt their skepticism. But as the session deepened, curiosity replaced doubt, and appreciation followed. That moment captured both the challenge and change underway.

Manufacturing has been male-dominated due to early exposure gaps, societal expectations, and workplace biases. Boys often get toys that build spatial skills, while girls are nudged elsewhere, creating confidence gaps that persist into careers.

But digitalisation and AI are reshaping the field. The work now blends intelligence and adaptability, creating a level playing field. More women are entering as engineers, entrepreneurs, and innovators. Each success story chips away at stereotypes.

I’ve also seen peers, both men and women, far more welcoming to diverse perspectives than a decade ago. The change may be gradual, but it is real. Competence and value earn appreciation regardless of gender.

As factories get smarter, the people shaping them will naturally grow more diverse.

Anubhi Khandelwal is among a rising league of women entrepreneurs in deep tech who are reshaping legacy industries through innovation. With a background rooted in problem-solving and a keen interest in robotics and artificial intelligence, Anubhi has always been driven by a single idea: technology should work with the real world, not just within fixed environments.

Her entrepreneurial journey took shape when she saw the massive gap between traditional manufacturing systems and the promise of intelligent automation. Most factories either lacked access to modern robotics or struggled with rigid, hard-coded systems that couldn’t adapt. Anubhi set out to change that to build a system where machines could think, learn, and adapt, just like humans do.

As the co-founder of Terafac, she’s not only bringing this vision to life but also carving a niche for women in the industrial automation space where representation remains minimal. Focused, hands-on, and forward-thinking, Anubhi leads with the conviction that automation should be intuitive, not intimidating.

About Terafac:

Co-founded by Anubhi Khandelwal and Amrit Singh, Terafac is a Chandigarh-based startup pioneering the field of Physical Intelligence. At the core of its innovation is a proprietary AI-Vision platform that enables off-the-shelf robots to perform complex manufacturing tasks without any programming.

Unlike conventional automation systems, Terafac’s solution empowers robots with real-time vision, decision-making, and adaptability. This means machines can identify objects, adjust to variations, and execute precision tasks across production lines with minimal human intervention. From assembly to quality checks, Terafac is streamlining workflows in industries where flexibility and speed are critical.

What makes Terafac stand out is its plug-and-play approach to automation, making advanced robotics accessible even to small and mid-sized manufacturers. As factories seek smarter, leaner operations in the age of Industry 4.0, Terafac is fast becoming a key enabler of this transformation. 

Industrial Automation Magazine delivers intelligent content for the industry