While we’ve said this before, considering the implications it could have on your business, it’s more than worth repeating. Unless you are an AI company, or can afford to emulate one, working with an AI provider is likely the most cost-effective means for you to access this technology’s invaluable power.
Of course, being an AI company ourselves, it might be hard for you to take Clarifai’s word for it. As such, we spoke to Alex Shah, founder of Garden Answers LLC, the makers of the popular gardening advice app My Garden Answers, to find out why he decided it was better to buy versus build.
Who is My Garden Answers?
Based in San Diego, Garden Answers LLC is a company that seeks to offer its customers cutting-edge plant identification technology for their mobile devices. In addition to their app, the company has built a team of gardeners and horticulturists who can quickly give their users expert advice and guidance for all their gardening needs and inquiries.
How are they using visual recognition?
With Clarifai’s visual recognition capabilities, the My Garden Answers app can recognize an extensive variety of flora from across the globe. The app provides users with detailed information on over 20,000 plants, including where each originated from and the types of environments they best thrive in.
Visual recognition sits at the heart of their app, enabling it to “see” and analyze the plant in an uploaded image then compare it to the millions of plant photos in Garden Answers’ database— in seconds. This, in turn, allows for visual search. Visual search lets users use images from their phone in a search query versus keywords or tags. The app then instantly finds the most visually similar match in Garden Answers’ image database, to help users identify the plant in question.
Why did they decide to buy versus build on their own?
As Alex Shah, the company’s founder and CEO, puts it, “I’m getting all this value without having to build it myself.”
With their team largely consisting of horticulturists, the experts they need to provide value to their users, to build their own AI in-house, Garden Answers would have been required to shift significant resources away from these core experts to pull together the right AI talent to get them going.
Back in 2017, Ernst & Young found that 56% of the surveyed senior AI professionals cited a lack of AI talent and qualified workers as being the greatest barrier organizations faced to implement AI. That trend continued into 2018 with a Deloitte study finding that even early adopters of AI were still without the requisite AI talent. The study further noted an important point. This AI talent is not just limited to technical roles, like research scientists and programmers. Most companies also lack experienced AI business leaders to help them identify the best use cases for the AI they want to deploy.
These struggles go hand in hand with the fact that though this technology is getting better and better, qualified AI talent remains in short supply. Getting the team needed to build, maintain and upgrade your AI in-house is competitive and costly for most businesses, including Garden Answers. So, when Alex determined he would need image recognition to create value for his customers, he chose to work with an AI provider.
After this, Alex moved on to the next step in the process of investing in AI: thoroughly researching the visual recognition market to see what was available for his needs. Not all visual recognition technologies are created equal or are appropriate for your use-case. Research allows you to get the lay of the land, so you can gain the context you need to feel confident in your choice.“I’ve done a good job researching other companies’ offerings,” Alex said. “And Clarifai has a good delta in terms of the value proposition.”
The amount of research required is entirely up to you. Architizer, another Clarifai customer, dedicated 3 years to researching the computer vision market before deciding we were the right provider for them. Other factors can make the decision-making process less intense. Staples Europe, for instance, knew they needed a visual recognition API that was easy to implement, accurate, and all-encompassing in what it could recognize. However, since they also needed that API to be fluent in the 12 different languages spoken across their customer base, their search pool was likely much shallower.
With interest in AI continuing to soar, some companies may be tempted to rush through this process, afraid they’ll be left behind. Having a well-defined use-case can make the research process easier and faster—after all, you’ll know exactly what you need. Still, stakeholders are encouraged to take their time here, as there’s more to evaluate than just potential provider’s visual recognition capabilities and accuracy. Pricing, the options for running your CV AI system, and how much support you’ll receive are among other factors you should consider before making your final decision.
For Garden Answers, Clarifai gives them just what they need.“Clarifai offers an absolutely, hugely valuable service to me even compared to other image recognition services.”
Utilizing existing platforms from particular AI companies can give you many of the same benefits you would get from building your own AI system in-house, including retaining control over your data, without dealing with the very real pain that comes with building one from scratch, maintaining and upgrading it continually. Instead, My Garden Answers can dedicate their time to improving their business where it matters. “Clarifai is providing good value that allows me to focus on what I should be,” Shah said. “Which is making sure our business differentiates against our competitors instead of engaging in academic research.”
At Clarifai, research is one of our most critical tasks. That’s partly how we differentiate ourselves from our competition. For most businesses though, resources could be better spent on your core product offering. Shah puts it well:
“We’re not an image recognition company, so working with Clarifai allows us to focus on what we’re really good at.”