AI Gold Rush: How to Build a Better AI Startup
Business + ManagementConferencesStartupsWest 2018posted by Elizabeth Wallace, ODSC September 25, 2019 Elizabeth Wallace, ODSC
We’re in the midst of an AI “gold rush.” It’s not just people working directly with AI, but the entire ecosystem that’s changing. Even fields affected by AI are innovating within the infrastructure, causing new use cases and new ways of thinking about AI.
It’s an exciting time to be working within AI. A lot of money is going into the space while startups continue to revolutionize the entire stack. There are lots of challenges, however, and starting something in AI because that’s what everyone is doing isn’t quite the answer. The biggest challenge in the coming years will be the uniqueness of creating something in this space.
[Related Article: 4 Examples of Businesses Solving Problems with AI]
Divya Jain, Director at Adobe Sensei, hopes to help give startups some advice for building a successful venture in this new space.
Providing uniqueness has a few constraints. The AI gold rush is filled with people navigating the same space, so you’ll have to consider these things while building your product or service:
- Problem complexity: There are so many angles. Providing value will come in the form of a unique approach to a problem or offering a unique perspective.
- Resource availability: You can’t compete with organizations like Google or Amazon when it comes to data. Your compute power can’t keep up either. Scaling out will take some creativity when it comes to data and the computing power you need. You’ll need the capital to support your growth and scale.
- Timelines: Experiments can take a while still with some algorithms. You must balance the fast pace of the business world with the slow nature of working with big data. You need a good feedback loop, and you must plan accordingly.
Despite these challenges, there are plenty of reasons to keep working with AI. The field is an exciting venture regardless of those challenges, but for Jain, there are four different things vital to making progress.
The main focus is to figure out how AI can solve a problem. The biggest question would be, “Do you need it in the first place?” Many business problems can be solved without it. Just because a difficult problem doesn’t have an easy solution, it doesn’t mean you automatically use AI.
You must consider your market when you’re creating your AI solution. If your initiative overcomplicates the problem simply to work in a buzzword, it’s not going to survive. If your initiative doesn’t have a clear ROI, it’s not going to work.
However, if you have the right kind of data and a simple solution, you might have a workable idea. Business needs value, and so much of that depends on the historical data available. Even if you have a viable solution, if the cost is too high, it’s still not going to work.
The takeaway: If you can think through a whole problem and not just the immediate problem, you can potentially create business value.
Looking at your problem in a vacuum isn’t going to help you get to your product. Instead, leveraging the ecosystem will allow you to grow and scale faster. Building from scratch may seem like a solid idea, but you’ll be behind your competitors, especially those enterprise companies.
The takeaway: Using free data sets or pre-trained tools will get you from point a to point B faster.
Your end to end pipeline should consider the existing ecosystem and have a sustainable flow. Even more so, considering what kind of answers your AI gives you and how it arrived at that solution will help your solution.
In a business setting, you’d never be settled with wondering how a colleague arrived at an answer and receiving a simple “Because I know.” You’d interrogate that answer until the methodology was clear. AI is no different. If you can’t interrogate your algorithm for the same why, you may never discover why your solution doesn’t quite work.
The takeaway: Interrogating your process and your algorithms
Doing work with other organizations can help you scale. Your solution potentially complements the work others are doing, and that could open doors for you to acquire the data you need even when you’d normally never have access.
Jain doesn’t believe you always have to have an eye on enterprise partnerships. Keeping an open mind about who might complement your solutions and vice versa could open up your access in ways that you’d never reach on your own timeline. For scale, that’s a critical factor.
[Related Article: 10 Startups Killing the AI Game]
The takeaway: There are no lone wolves in AI development.
Providing Unique Value During the Rush
Divya Jain believes that startups can still offer unique perspectives to our problems, but they must keep in mind her four core components. Trying to build a solution from scratch all on your own may make a good story, but it’s likely to put you behind the curve in an already fast-moving ecosystem. Keeping an open mind and building relationships could be the way you finally begin to stand out from the crowd.