The enterprise AI conundrum: Build or buy? (VB Live)

• Updated on: 2019-04-25
• Read original article here

It’s do or die time: enterprises are sacrificing competitive advantage without AI. And the first challenge is, build or buy? Learn from experts in the trenches how to identify business needs and get started in the most cost-effective, profitable way when you register for this VB Live event!

Global business value from AI leapt by 70 percent from 2017, increasing to $1.2 trillion. By 2022, that number will triple, going up to $3.9 trillion. AI isn’t the future of business any more — it’s the urgent right-now, dividing the enterprise companies with significant competitive advantage from the companies that are going to be left behind.

The first step isn’t deciding whether to go all-in on an AI strategy, because the market has already made that decision for you. It’s where and how to begin bringing AI on board. The first question, and the foundational one, the one on which the success of your entire AI game plan rests, no pressure, is should we buy, or should we build?

Of course there’s no one-size-fits all solution, the question isn’t an easy one to answer, and fun fact, making the wrong choice can cost you money, delay your time to market, and hit your market share hard.

It’s also not a this-or-that, black and white question either. Your options vary widely and range over a spectrum of choices and decisions.

In-house party: Is your enterprise large enough to throw your weight behind the build-in-house model of AI implementation? It means open-source toolkits and a series of expensive, but essential new hires to build a team that not only understands AI and plays it like a fiddle, but also knows how to most effectively leverage AI and machine learning to accomplish your business objectives.

Custom-built, for a price: There’s also a wildly proliferating number of startups that offer full-service custom AI consultation, set-up, and service, but in a crowded field, which vendor do you go with, and do you even know what services you should really pay for?

Pre-fab(ulous): Coming in fast, and heading to the top of many enterprises’ wish lists with a bullet, are the artificial intelligence as a service (AIaaS) and machine learning as a service (MLaaS) vendors. Companies like Amazon, Microsoft, and Google are of course getting in on the game, offering cloud-based solutions, ready-to-order data sets, and pre-trained algorithms for that new, prefab AI smell.

In-house AI solutions reduce overhead, in theory, because you’re only building what you need, when you need it, and extensive fiddling with options and features isn’t necessary. Once you’ve established a data science team and an AI-first mindset in your company, in-house builds offer the freedom and flexibility to build and explore on the fly, launching projects as fast as inspiration strikes. And any one of those projects can become a competitive advantage not just in the market but as intellectual property, which increases shareholder value.

But finding qualified, brilliant talent is wildly difficult in this market, and the vast mountain of data many enterprises have isn’t plug-and-play — it also needs to be cleaned, processed, and then used to train your algorithms, and that is sometimes an enormous undertaking. It might even be that you don’t actually have the data you need — and gathering it and training it can be a big time suck.

Buying your AI can mean cost savings since you’re not shelling out expensive financial and employee resources out of pocket. Vendor solutions also come with pre-vetted, clean, shiny and new solutions, because quality assurance stands between a vendor and going bankrupt for breach of contract. Your solution comes in a high-quality, stable package, and will sync with whatever systems it claims to be compatible with on the package. Security is high, and ongoing customer service is almost always built-in, which means upgrades, enhancements, training, and more is almost always part of the deal.

The problem of course is that the number of vendors leaping into the scene means sorting through a universe of pinky-swears about the efficacy of their solutions and the value they can offer you. Finding a vendor that offers all the compatibility and functionality you need is also a long and arduous proposition. Super-customizing those solutions can be tricky, and cost you time, money, and take your focus away from where it needs to be: innovation, development, and implementation.

While packaged offerings are ubiquitous, many businesses need unique solutions. The need for custom solutions requires additional cost and time for implementation.

So how do you choose? It’s never a close-your-eyes-and-point decision, more’s the pity.

As ever and always, it starts with identifying the need. What problem in your enterprise do you need to tackle, and will throwing AI at it be the most cost-effective way to handle it? What kind of AI solution do you envision solving your business need and directly impacting your ROI? And perhaps most importantly, does that solution already exist?

When you’re choosing between various flavors of build or buy, what’s driving that decision? It isn’t as simple as “cost!” or “my brother just launched a startup!” There are a lot of really essential, very key criteria to keep in mind as you work through your decision.

The category is: Build or buy. The questions to answer are:

Getting the right solution requires in-depth conversations with stakeholders, including leadership, technology, marketing, and sales. And transparent, hard-hitting discussions with the vendors or a potential new team you are considering bringing onboard. Also, insight from the experts already in the trenches, who have made the decisions, lived with the results, and have the real-world case studies and results at hand to help you make the best decision for your unique business needs and enterprise. Sign up now for this VB Live event and get up close and personal with a panel of those carefully chosen experts, ready to break down their experiences and answer your questions.

Speakers to be announced soon!