Recently, a PWC report claimed AI to be a mammoth contributor to the world’s economy and for all the right reasons.
Most processes we see around us have the capacity to be more efficient and productive by removing the human element out of them, increasing efficiency and reducing error.
With this size of market contribution, it’s poised to empower a lot more than just processes; even decisions and industries at large.
In the day and age of Artificial Intelligence, there’s very little room for an industry to not be impacted by its reach – making its application highly industry agnostic.
Whether you’re in the Education sector trying to predict student learning curves and outcomes to enhance learner experiences, or in the Human Resources sector trying to parse through millions of CVs and predict attrition; your industry is impacted.
It’s positioned to impact industries by creating a ton more jobs than it reimagines – increasing the demand of high quality and byte-sized experiential education and the constant need to re-skill and up-skill.
If it’s so powerful, why aren’t we experiencing a complete overhaul just yet?
The core problem of current AI systems is not so much that they are narrow — specialisation can be very helpful — but that they are narrow by design; they freeze in time.
For years, selling against buyer egos has always been harder since we as humans inherently believe we don’t yet need intelligent machines to tell us what to do; we’d rather tell one to act in a certain way.
AI as a tool for automation is how it’s currently being built, where it can do one particular task extremely well and at scale – a useful thought to consider while building it into your product!
In the real world things change all the time, and intelligence is by definition the ability to effectively deal with change; narrow AI stands incapable here.
Watch the Trend
In one of the best books I’ve read in my entire lifetime, one chapter particularly stood out for me; industry trend watching, it’s a skill.
Pick an industry – any one. Drill down into processes, usage and behaviour within that industry. Do you see it getting obsolete? Do you see a drastic change coming up soon?
That’s your spot. Begin there.
An internet company in the early 2000’s, well, caught on to the trend. A physical photo camera company in 2020? Not so much.
Trends are the ground beneath your feet; they move you up and down sometimes without you making your own moves first.
What do I keep in mind before I begin an AI venture, or build it onto my venture?
Before any of the things to keep in mind, find a problem that can be solved using AI – not the other way around. Every problem isn’t an AI problem. It isn’t a silver bullet.
Assuming you’ve covered the screening bit well….
Domain Specific Data
You have a venture in the EdTech space and wish to predict learner outcomes. You collect very limited data around how much time a student spends on the platform.
Your prediction will only take you so far, and benefit the learner even lesser.
Your friend has a similar venture: collecting data around time spent on the platform, course choices, completion rates and so on. They make more detailed predictions, giving the learner greater benefits and thereby having an edge.
Distance from the Buyer
This remains a particularly debatable slide, where I could be blamed for having an inside view.
Over the years, I’ve noticed how the value of your AI offering is amplified the closer it is to your buyer.
If your venture is building an AI solution to predict learner outcomes in EdTech and gives a detailed analysis of your strengths, weaknesses, and way forward to your mentor instead of you – it makes you dependent on the mentor to gauge all of that information and teach you better.
Whereas if it offers that same level of detail in its analysis to you, you’re more likely to pay for it.
Key Product Differentiator
One of the classics, doesn’t even need much of an explanation.
I understand that you have a pen that does above-average in writing smoothness, but has a microphone, a speaker and wireless calling; I do have a phone for that right now, your pen better write supremely well.
I have an AI venture in place, how do I scale my AI deployments well?
Based on Bill’s teachings at MIT Bootcamps, I came up with something I call the value creation filter that helps diversify into follow-on avenues from your current one; where your AI exists.
From where you currently are, moving in a different direction would be beneficial if your next prospective deployment passes the value creation filter.
Your deployment either saves 10x Time, 10x Money or enables 10x Reach.
Chrystina Russell at SNHU was kind enough to walk me through how they’re absolutely killing it with their AI deployment pilot to assess student work and assignments!
This pilot enables detailed assessment of student work, helping in better and more robust feedback for students in lesser time and lesser costs.
One of the main costs of attending universities is the cost of good quality educators, who do more than just teach. If the costs could be saved in this area, SNHU would use this to enable students from low-income regions to attend college as well; thereby increasing their reach!
[Related Article: AI and Small Business: It’s Not Just for Enterprises]
This initiative isn’t trying to replace human assessors with technology, but rather make them more efficient with their jobs while also addressing the age-old assessment problem of human bias.
Sometimes, you lose track of what really adds value to the customer when you’re obsessed with tech. It’s easy to be distracted by the capabilities if you’re new to it, and be stuck in an inside view where you think it’s perfect but nobody buys it.
Lastly, remember to directly impact visible KPIs. This is a thumb rule that’s almost said too much – if I can’t directly see where my purchase is adding value, I may not be buying it super quick.
Selling a prospective job to a student after your course is easier than selling them an “adaptive and personalised educational experience”.
Hope this has been helpful for anyone in a similar setup, planning to build and scale an AI venture or build it onto their pre-existing ones.
Special thanks to Aaina Singh for helping me put together some stuff on here! 🙂
Originally Posted Here