Start an AI Project: An Open Letter to Business Leaders
Dear Leader of Your Business:
You see the headlines like ‘AI is the new electricity.’ You hear that AI software will create $13 Trillion of GDP growth before your kids head to college, and before you run a marathon and maybe even before you make CEO.
When you click on Amazon or Facebook, you already know that the AI powering underneath is creating all sorts of extraordinary conditions for them, to create even more value for themselves. You’re smart. You see AI coming for your industry too and you even see the Tech Companies with their troves of data, themselves, maybe coming for your industry. No, no, please don’t take flight. Oh, and don’t just freeze either.
Stay. And fight, but do choose to fight small battles with intelligence. With 2019 out the door, here’s a shortlist of how to move forward in 2020. If you’ve got data, yes it may be quite unorganized and in silos, start with that. However, if you have some data, then plan for Machine Learning to help you automate tasks for you to gain greater clarity and insight over your business. Start small, start smart, but just, just start.
A Pilot Project is a Fine Place for a Beginning
..and the following are what to keep an eye on to keep this effort both strong, and manageable, to get both, a win in quickly (under six months), and with high success probability. Maybe not start with one, but rather three, increasing your odds of barreling success. You already know success will make it easier to bring along the rest of your team.
For an OTA, our team is providing an AI-Ready Platform helping them manage integration with existing IT systems. Being able to work with them closely and dynamically responding to their need to be AI ready has helped them move quickly and aggressively in this direction.
Focus on your Own Arena
..so this pilot AI project applies to what your company and your industry actually do, but writ small. If you focus Artificial Intelligence projects on sideline activities to your main business, guess what, I’d bet, there’s already someone doing it, and doing it better and really it doesn’t move your company’s needle. If you’re a manufacturing company, maybe try finding anomalies through AI, not AI for hiring which you can buy easily if you need.
The Right Partner here is crucial, as you need deeply intelligent, iterative people with experience to build strong AI, and just as importantly, who take responsibility and ensure it fits seamlessly with your company’s existing technology. At the beginning you won’t already have a team in place, and waiting to grow the team would leave you lagging where you should be leading.
Make sure the AI is ‘Creating Value’
Either costs need reducing and/or revenues need increasing for value to be created. Generally this happens through prediction systems, which can increase sales or allowing insight to help you launch new lines of business. Look for value, which can be found even in sparse data, or by automating tasks, or outcomes or budgets. Figure out the right combination.
To really pull this off, look for a leader who can work cross-functionally, and is respected through the organization – hey why not you? You’re not running the whole 10K yet, just doing a half walk, half fun run to learn, practice, get strong. The aim is to get ready to take on the whole 10K, where the benefits far outweigh the pitfalls.
Test, Test, Test
..to check actually, how the project is doing at creating business value. The principles of good software development – technical teams who examine the data, and the conducting of small-scale experiments don’t change. A small team, which works really well together, would involve more bang, for less resources from your side. We saw this during an initial pilot for a solar plant company where we helped one of their plants to help them predict outages and also reduce the number of outages. Based on its success, we plan to expand to all their solar plants, and then by leveraging prediction and anomaly detection algorithms.
Be the Team’s PR Agency
..by making sure both Business Teams and the Technical Teams involved get internal kudos, plaudits and of course both time and attention in ‘CEO sunshine.’ This will be crucial in getting more people and teams ready to join in with you.
Keep your goals simple, lead your company in learning this new AI vocabulary, work with the right partners, and finally, create value by really being focused on AI for your industry sector. Run three pilots. See how they go, then plan on ramping up to apply AI algorithms and see where it all leads you.
Loads of good luck! Oh and Happy 2020!Share on Facebook Share on Twitter Share on Pinterest