Artificial intelligence (AI), promises better solutions, quicker results, fewer wrong decisions, and predictive analytics. This all comes at a cost, but it’s also important as a user to determine what you are looking for in a solution. The main caveat is not all solutions are created equal, says Joseph Zulick, manager at MRO Electric and Supply.
There are providers who are determine your metrics for you and some that want to slant data to adjust to your goals. Realise that data is equal and that gathering data is very “matter of fact” in nature. We start to slant it when we decide to place importance on some information and ignore the other data. Welcome to AI and it’s solutions.
Let’s look at where we are today. Artificial intelligence is creating incredible results but is evolving slower than some people expected. There are many reasons for this from support programs to computer language and security blocks. How accessible is the data and the source connectivity to real world data. Some systems already filter the data and the sensors before the back end office systems read it in raw form.
Not all AI is created equal; It’s one thing to make a stand alone AI system that you build from the ground up, hardware and all, but it’s another to analyse other systems’ data.
Analysis. Can the system analyse the data effectively. Does the system have the programming to achieve quality results providing you, the user, with reporting that is effective and accurate for your needs. Some systems have ‘Analysis Paralysis’ built in which leaves you with too many choices of data that is ineffective at best and inaccurate at worst.
Connectivity. Can the system connect to the data you are looking for in a needed solution. Today in the IoT 4.0 world and in the AI world there is a tremendous amount of finger pointing. “It’s not our fault we can’t analyse data that is not accessible,” says one company. The counter side of this conversation includes comments like, “We can’t provide access to third party companies or companies with questionable security protocols.”
These are discussions happening regularly in systems that require other systems for data. It’s frustrating to hear after you have just plunked down a large sum of money for a system. Cooperation is a must if the system is not a ground up provider. This includes compatibility and language acceptance.
Who are some of the top providers of AI in different categories and technologies? Here are a few.
Many of the companies are a who’s who in computing, analysis and hardware such as IBM with their Watson system, Microsoft with Azure and their partner systems of OpenAI and Artificial General Intelligence may lead the way with many heavy hitters getting in on this partnership; and Google, of course with their search engine optimisation providing tools to users, sometimes listed under their parent of Alphabet. Nvidia whose chipset optimisation and semiconductor business is forging ahead by partnering in non typical computer areas like automotive.
In the SaaS area salesforce has long been a leader but now is developing their systems to incorporate AI through their cloud based system.
Amazon has been spearheading the AI world, naturally, with their hugely successful backing of the database system tied to the most successful home shopping system but is only a small portion of their AI business which is unleashed through their AWS business systems, analytics and logistics.
Facebook‘s algorithms have continued to develop and have experienced much hullabaloo over their computer communication with each other which is as much of a cautionary tale as it is a breakthrough of what is possible.
Some of the lesser known but no less important players in their own fields are Twilio, the add on interface company specialises in making interface related software tools for voice and connectivity. Tencent is specialist in the Chinese market of social tools but inevitably will weave into a world wide player in this area if they so choose. Alteryx is a newer and lesser known company in the AI world, but their data and analytics interfaces provide information and communication for many products especially in SaaS.
From built in.com, Here are their top companies, they have many of the same but a lot of smaller up and coming companies.
ThirdEye Data, theft detection. DataRoot, AI system solution. DataRobot, automated machine learning. Master of Code Global, chatbots. H2O, open source system designed around Nvidia. Hatchworks Technologies, managed cloud infrastructure
CloudMinds, open end to end robot system. Fayrix, software developer and big data scalable solutions. STX Next, NLP, speech recognition and AI design. Xicom Technologies, custom AI provider
As you can see there is a lot of opportunity in the AI world but you will need to work your way through to see who can deliver to you a solution that meets your company’s needs and is scalable for all size companies. Lastly, who will be around and not be providing an obsolete system a year from now. These companies provide some of the most important products and have the long term backing to be in it for the long haul.
Remember, some of these companies are the backbone of a system that you may not even know exists. Their hardware and programs provide the means to gather, analyse and process the information we use but may even be in the cloud as part of another system provider.
Realise that many of these are creating either a very customised solution or a very vanilla stripped down product. I use product very intentionally here since many of us are looking for a solution and a solution requires a lot of time and customised metrics. The reality is that if you knew the metrics that are required to provide the solution, you probably wouldn’t need AI!
Most companies will expect you to already know an awful lot about your process and have continuous improvement solutions currently in place. You should know what some of your Key Performance Indicators (KPI), are and how they are valued in the process analysis.
If you expect to just plug in a solution and it spits out an answer, you may be disappointed, unless it’s the answer you already wanted in which case the system is just there to provide support for a theory of improvement. I strongly suggest that you undertake a level of continuous improvement, Lean and / or Six Sigma so you have a baseline of information. Some companies will include this but you can pay a college intern to hold a stop watch just as easy as paying $150/hr (€126.50/hr) to an MBA grad to write down TAKT times. The more turnkey the solution, the more cost.
Deployment. Is your system truly AI or is it really just data mining? When you launch your system you should be seeing that the AI interface is actually developing. Is it building new decision trees, assessing value of data and deploying improved solutions?
Some considerations; Hardware limits, who can deliver the hardware as well as the analytics. If you have to piece together a system, do you have the expertise to handle the scope of the project? If not, plan to spend more for the solution with a provider who can supply a turnkey system. Software limits, is the software only for their system or can it integrate into systems like Salesforce to leverage their cloud based CRM? If you have 5 standalone systems you will not like the work it takes to piece together a single page report with results from each system.
Looking at the top AI companies the word that comes to mind is opportunity. Opportunity to move ahead. The opportunity that all of the companies provide to us is a strong future in AI that we can all leverage for our future.
The author is Joseph Zulick, manager at MRO Electric and Supply.
Follow us and Comment on Twitter @TheEE_io