Sachin Dev Duggal is the co-founder and CEO of Builder.ai, a UK-based human-assisted artificial intelligence (AI) platform to build, run and scale tailor-made software. 

As more and more companies race towards a digital-first strategy, there is now an existential adoption problem for them. There isn’t enough supply of great talent and there isn’t enough technical understanding by this new customer.

The axis to this problem lies in three parts. First, it will be about removing large swathes of work needed so you can balance the supply and demand of expert hours needed for either designing or building your app. Second, it will be about traversing the wide knowledge gap between those who can and those who cannot. Finally, it will be about choosing the right expert to close the creative and expert loop.

We have seen this play out often in history - with car manufacturing (or any automated manufacturing to that extent), autopilot in planes and even something as mundane as spell check in workplace communications. Today, the very fabric (software) that was the catalyst to automation (from ordering a taxi to predicting when your coffee beans will run out) is likely to be semi-built by itself.

Automation is a key technology trend impacting business growth and success. Over the past few years, it has gone through an accelerated evolution - from being used for basic automation of repetitive tasks to being ingrained in intelligent, context-aware machines that learn and respond to human behaviours. These AI robots are not likely to entirely replace humans any time soon; partly because our capacity for creativity, empathy and innovation remains sacrosanct to anything more than simple flows. And so, the role of AI /deep automation and its immense capability to perform tasks in seconds will free the “humans” from the humdrum work and allow them to focus on things more aligned with the human spirit.

According to a McKinsey Institute report, realising automation’s full potential requires people and technology to work hand-in-hand. This is a reinforcement of my own belief that true automation will be either “AI-assisted humans” or “human-assisted AI”.

The benefits of this human-AI partnership are hard to ignore. There is even a term for it: cobots, or collaborative robots. Cobots are designed to work with humans instead of replacing them entirely, augmenting human capabilities with their strength, precision, and data processing power. This happy marriage of human ingenuity and raw machine power unlocks so much potential and myriad opportunities, which could cement this partnership as one of the most significant developments in the continued evolution of both humans and technology.

Smarter machines, expanded capabilities

One of the greatest strengths of AI is its ability to understand and properly respond to context – as well as being able to fine tune its responses by learning. A properly trained AI can save human users hundreds of hours that would be spent on repetitive work, particularly in the software development field. Not only can AI shorten the amount of time it takes to perform complex tasks like QA testing and UI/UX testing, it also does not sacrifice on the quality of work and accuracy level. How can that be possible, you might ask? Simple - because it has been taught by humans; not in the literal sense but by studying “good” and “not so good” behaviour in a particular task and being able to detect the pattern.

There is far more to AI than repetitive task work. As we have seen with Alexa, Siri and Microsoft’s Cortana, AI can add tremendous value to our lives by acting like a smart assistant and providing recommendations based on pre-trained patterns. Spell checking software like Grammarly is a good example of how AI can be useful in improving the quality of work communications. But that functionality is amplified in coding where even a missing parenthesis or semicolon could cause an error. AI systems can be trained to automatically spot these mistakes and suggest replacements, saving hours of time that would otherwise be spent debugging.

This capacity for context also lowers the language barrier between human and machine, which has historically been a key communication inhibitor. Typically, computers can only recognise and execute instructions that are written in programming languages, which are difficult for humans to learn and use. However, with Natural Language Processing (NLP), AI machines can convert ideas written in natural language into code that machines can understand. That, in turn, simplifies the coding process and turns ideas into workable software much more quickly.

And these facets of technology can help us in building software itself; from translating a single one-line statement “I want to build an app to let people order from my restaurant and see delivery times”, to asking questions about how you want to manage your inventory and even gluing a prototype together or choosing the right expert to tailor a particular feature. 

The importance of the human touch

AI-enabled systems shine brightest as a bridge between human understanding and software power. The ability of AI to comprehend and convert human intention into software instruction is what will make the next generation of no-code software development an eventuality; not the “buy your oven, do 10 bootcamps to make a pizza” kind but the “Domino’s style of ordering a pizza, put on Netflix and eat in 30 minutes” type where you really do not need any pizza making expertise (inherent or trained).

The move will be to let humans “order to create” rather than “build to create” through visual interfaces instead of text coding or spending days dragging widgets on a screen. The greatest simplification of making digital dreams come true will be to completely remove the notion of knowledge and “building”; greatly simplifying the app creation process and lowering the barrier to entry for software development.

With AI taking care of the vast swathes of repetitive and often mundane work streams related to app development, humans can then be free to create, innovate and be problem-solvers, rather than solution-led. It is like Simon Sinek tells us – focus on the why and how, not the what.

The other powerful lever in this approach is that everyone who is “ordering” their app gets the benefit of the collective knowledge of everyone who has built before. The biggest challenge today in building or ordering software is that your knowledge is capped at the aggregate knowledge of the network working with you (albeit an expert or an advisor). With a human-assisted AI approach, you have the serendipitous talents of the expert network coupled with the holistic benefit of everyone that has built on the platform before you (the training set for the recommendation or decision being made by the AI).

This means that suggestions of best practices and user preferences based on collective knowledge can be entirely automated, while humans leverage those insights to build apps that provide a better experience or optimise operational efficiencies. Such is the growing popularity of this NorthStar that Gartner anticipates app development will not be done by a traditional engineering team by the end of 2025.

Retaining the element of human touch in app development is important because they are ultimately designed to be used by humans and solve human issues. While AI excels at driving efficiencies, accelerating development, and improving cost-effectiveness, it still needs human input and interaction to deliver exceptional, personalised customer experiences. Being able to anticipate user behaviour and incorporate features that people would use is a large part of app development, and that remains the domain of the human builder.

An essential and beautiful friendship

Digital is at the heart of nearly every experience today. This is especially true as the world grapples with Covid-19 social distancing rules. Apps and software will soon become the foundations on which our governments, economies and societies operate.

But the digital divide is still a very real issue for many countries. This became particularly evident as millions of children worldwide had to shift from in-school learning to virtual Zoom classes due to the coronavirus pandemic. For families without financial means and access to digital devices, learning became incredibly difficult. In this type of situation, a human-AI partnership could do wonders for bridging the divide and removing such barriers to inclusion in a digital world. At the same time, it could open up new possibilities for talent and innovation to solve such problems.

Existing narratives on humans and AI often focus on how AI can do tasks just like a human but better, leading to a common, not-so-unreasonable fear that AI machines someday will replace humans in the workforce. In reality, that outcome would produce only short-term productivity gains and would not leverage the full potential of both AI machines and human workers. Both have their strengths and weaknesses. To quote Helen Keller: “Alone we can do so little; together we can do so much.”