What successful ML companies do differently?
While there are technologies like VR and Bitcoin, that are in the primary phase of the technology adoption life cycle, machine learning and AI are obviously in the early majority phase. And it’s very probable that they will grasp the late majority phase very soon.
Every business would be using machine learning and AI for business applications very soon. By now, VCs have a hard time captivating software pitches seriously if they don’t include ML/AI as a sustaining. If you remember, SaaS is now known as a delivery mechanism and a business model shift, AI and machine learning will be known as enabling technologies that make every application (and every user of an application) keener and more proficient.
In the same way that there’s nothing exclusive about SaaS, we trust that over the long term there will be practically nothing exclusive about the software that permits AI/ML. By now, Google, Amazon, Microsoft, IBM, and many others make their ML frameworks readily accessible. And that’s because they identify that the adoption of ML in software is mainly to their benefit.
If all software depends on ML, there is one factor that becomes critical to the success of that software ─ data. Already the big daddy’s in the industry like Google, Facebook, Amazon, or Apple enjoy a clear and substantial data advantage with deep data.
So as a startup what do we do differently? At Brainfuel, we are focused on implementing ML/AI in companies that have two crucial elements:
- Data that is mostly hidden from the big players and,
- Cost-effective but scalable approaches
Our thought is that combining these two elements leads to long term benefits and a competitive platform that guards against both the established players and other startups that enter the market.
There are certain areas where data is not easily available for the big daddies and Industry leaders. We are targeting those domains and verticals where the big players don’t have access, experience, or focus.
For instance, say we work on a project with healthcare providers to access medical records data, patients result in data and insurance claims data to forecast and avoid acute health events amongst the providers’ patient populations. This data is not only unobtainable to other companies, but it also needs the difficult work of aggregating, cleaning, and normalizing the data, which comes from many sources and in many forms.
We actually go so far to generate the data we need for our ML models. Through the modularized training app, the company measures how well customer service reps know the information that is required for their jobs. And then we use that data to build models to forecast when each rep is likely to have forgotten specific pieces of data so that the right training lessons can be resurfaced at the right time to the right rep.
At times we have data that doesn’t happen within the domains of the tech behemoths and isn’t simply available to them given that their significance lies elsewhere.
But as you may have observed, we depend upon getting access to data from our clients. This is why we trust that win-the-market strategies are the second constituent of achievement for ML/AI software businesses.
We think of winning the market as different from merely going to market because the latter thinks that just selling is significant. We believe that selling is essential, but not enough, for winning, especially in ML/AI-driven markets. As vital as selling is delivering customer success.
In fact, we’d go so far as to say that in most software markets it’s not the best product that successes, but the one that solves the real business problem. When we estimate win-the-market strategies we look at three specific components to measure the potential for success.
The first is whether the business is addressing a grave pain point, what we often state to as a “hair on fire” problem. If your hair’s on fire, you’re possibly going to put that fire out before you do anything else! In the same way, if you’re a business client that is dealing with many different challenges, you’re going to concentrate first on the most urgent need and the potential solution to that problem.
The second component is whether the software needs a change in user behavior. Getting people to do anything different than what they’re doing today is tough. And that’s even more true in a business context when numerous people or teams may be involved in a detailed workflow or process.
Further, it’s much easier to access a budget that’s already made for a precise problem than to generate a budget for something that seems new. Fitting into a potential client’s existing business and budget makes for a simpler sales process and a quicker sales cycle.
The third component is whether the software will make the client a hero. It’s one thing to get somebody to buy your software, it’s another thing for her to be effective using it. The power of ML/AI is it can help give clients speed, accuracy, and insight that meaningfully affects their businesses.
So the solutions that are most likely to succeed are the ones that not only leverage vast amounts of data but gives users the ability to provide insights on this data and makes the users a superhero in their organization.
