AI has revolutionized how software gets developed. Code generation, prototyping, debugging, documentation writing, and finding technical solutions have all been made significantly faster by the use of AI.
It has become a fairly obvious implication for business - create more software in less time.
There is, however, one additional question that needs to be considered:
If AI has made the execution process cheaper, will making the right decisions on which software to create become increasingly important? I believe that this is the case.
The Cost of Software Development Is Falling
Software concept that previously took weeks to develop can now be turned into a functional prototype in days or even hours.
AI coding tools can assist developers to:
Generate application code
Build databases
Write tests
Fix bugs
Build interfaces
Clarify new technologies
Automate development processes
This is a huge boost in productivity.
However, this is also a problem that emerges unexpectedly.
With building becoming cheaper, just having ideas is not enough anymore.
More Ideas Lead to More Distractions
What would happen if a team could build ten prototypes instead of only two?
That sounds like a completely positive development.
What if none of these ten prototypes solves a customer problem?
The team will only get faster at making unnecessary things.
AI makes this problem clearer because it creates almost zero friction between idea and implementation.
"Should we build this?"
AI Is Very Good at Exploring
AI is especially helpful for exploring.
It can help you:
Come up with possible products
Compare alternatives
Find risks
Competitor analysis
Recommend features
Prototype
Question assumptions
All of which can lead to many more possibilities.
However, that doesn't necessarily translate to making better decisions.
Somewhere along the line someone has to narrow down the field.
The Power of Convergence
There is a crucial step that occurs between exploring and executing.
It is called convergence.
Exploring starts with:
What could we build?
Converging starts with:
Which of those should we actually build?
Executing starts with:
How do we build it?
AI is now very good at exploring and executing.
Convergence is hard.
Why?
Because convergence often involves judgment.
One has to know the customer, the market, the business constraints, timing, risk, available resources, and opportunity cost.
Not all of those are found in a data set.
Human Judgment is Key
AI can judge an idea from multiple viewpoints.
However, often the final call comes down to such questions as:
Does it matter to our clients?
Is the pain point big enough?
Is the market interesting?
Does it align with the strategy?
What if we are mistaken?
What do we give up by pursuing it?
Is it the right timing?
These are not technical questions.
They are decisions on value and priorities.
Faster Execution Alters Strategy
When development of software is cheaper, experimentation becomes easier.
Rather than spending months developing an idea and then finding out, teams can prototype it and validate assumptions.
This can provide a strong advantage.
However, speed should not come at the cost of validation.
In fact, a great strategy could look like:
Explore fast → eliminate ruthlessly → prototype inexpensively → test with users → develop thoroughly.
AI can improve almost every step of the way.
Humans still have to make a decision what avenue to pursue.
Human + AI Judgement
It does not mean that AI is not good at making decisions.
On the contrary.
AI can be an outstanding thinker.
It can uncover hidden assumptions, reveal risk factors, assess alternative options, and highlight flaws in your thinking.
However, there is a difference between helping someone to make a decision and taking that decision yourself.
And this difference matters even more as AI is getting smarter and more powerful.
For companies considering ways to use AI, IoT, and intelligent systems in practice, Aperture Venture Studio researches tech-based solutions for complex problems in industry and business. The New Bottleneck
For many years, software creation has been limited by engineering capabilities. Now AI is contributing to solving this issue.
However, the next limitation could be attention and judgment.
Businesses may very soon find themselves with too many ideas, prototypes, experiments, and software to handle.
While the advantage might not go to the company that creates things the fastest… … it may be taken by the company which knows how to decide what needs to be created in the first place.
Final Thoughts AI makes software creation faster.
It's useful.
But faster execution doesn't mean better results.
While the cost of creation decreases… …the value of good judgment increases.
It could be possessed by teams that are able to explore extensively, converge intelligently, and select strategically.
AI can help create virtually any product. The difficult part comes from selecting what to create.