Just over a year ago, SAS announced that its AI investment strategy is set to span across the next three years, with the company’s AI solutions including analytics, machine learning, deep learning, natural language processing (NLP) and com-puter vision. The latest IDC report said the worldwide spending on artificial intelligence systems was forecast to reach US$35.8 billion in 2019.
“I am a big user of AI myself, and areas such as voice to text, text to voice and image recognition, are where the major use of AI is. We can actually improve some of the models that we used to do and make them a little bit better using some AI techniques,” says James Goodnight, CEO of SAS Institute. It was a necessity to know more.
“All the image recognition that we are doing is going to end up within medical. We are doing work to look at CAT scans of liver cancer patients and determine whether or not the cancer is growing or shrinking,” he adds.
“A lot of times when liver cancer is discovered, the tumours are too big to operate on and the patient will be put on a chemotherapy regimen to try to shrink the tumours so that once they get down to the proper size and they can operate. You need to know from scans whether or not the tumours are shrinking, and we’ve got to where we could do that extremely well.”
SAS is betting on AI at a time when several industry verticals are making investments in AI solutions.
SAS recently released Viya 4, available in late 2020, and is engineered to take advantage of the latest cloud technologies. Designed to be delivered and updated continuously, the new architecture helps bring powerful analytics to everyone, everywhere.
Because SAS Viya integrates the art of decision making with the science of artificial intelligence (AI) and analytics, organisations will be able to make better decisions, faster. The innovation behind this release underscores SAS’ commitment to helping organisations transform data into intelligence. The release of cloud-native SAS Viya is one outcome of SAS’ $1 billion investment in AI.
“The goal was to continue to emphasise in our R&D group with the development of AI. We are going to demonstrate GAN (generative adversarial networks), that is where you actually artificially generate a person talking, and that is interesting because a number of our insurance companies are interested in helping detect fakes,” he explains.
“People say they were in an accident and send in the picture of their car, then it turns out to be somebody else’s car or the car was generated by computer.”
Microsoft meets SAS
This summer, Microsoft Corp. and the analytics firm announced an extensive technology and go-to-market strategic partnership. The two companies will enable customers to run their SAS workloads in the cloud, expanding their business solutions and unlocking critical value from their digital transformation initiatives.
“We have to work with all of the cloud vendors because our customers want us to run in the cloud. They took a survey of the ones planning to move to the cloud, and many of our customers are planning to do that. Microsoft was the most popular one with our users,” says Goodnight.
“It is just like any other platform; we need to run on that platform because our users want to be there. Thus, it is really a customer driven solution that we need to be ready for. We already run in Microsoft cloud, but this will give us the ability to take advantage of many other things as a company.
“We will be putting in much more work into making sure that we are the best we can be on Microsoft, but there are a lot of analytical vendors that will be moving there, and a lot of them are already there.
“Sometimes that is the only way they offer their software – through the cloud. Around two months ago, we saw many Silicon Valley start-up companies laying people off because venture capitalists were a little tired of putting money up and not seeing any return, so there is quite a bit of talent on the market right now.”
When asked if this is a bad or good thing, Goodnight explained that it is certainly the latter as the labour market had gotten really tight and it was previously hard to find people.
Working through the pandemic
SAS has close to 5,000 people working remotely in Cary, North Carolina. The landscaping crews are still going to work but they are outdoors, he explains.
“The rest of us who work indoors have been working from home for the most part and we are still paying everybody, even the people that work in the cafes and day care. A lot of the people that cannot work from home because their job is here, we have kept all of them on home salary. Our revenues are down about 6% so far this year but our expenses are down too, so that will balance out because we are not travelling,” says Goodnight.
“Everybody that we talk to says they are more productive from home. You do not have an hour to commute each way, you
do not have to get up and spend too much time getting dressed, therefore, people were claiming that they are actually starting meetings earlier than they used to.
“R&D is meeting all their timetables and deadlines; therefore, I think we might continue to stay at home. In fact, I am afraid if we ask everybody to come back to work, a lot would not because they are still afraid of the virus. Because of the virus, they have kids at home that are not in school, none of the day cares are open, thus it is going to be tough to get back to work until we have a vaccine.”
SAS’ future
When asked about the company’s five-year plan, Goodnight says SAS does not operate on looking five years ahead, instead keeping its horizon down to around two years, because that is about how long it takes to get a new product out of the door from start to finish.
“We want to be nimble and easy to move and have the ability to change plans swiftly. If you are in the middle of a five-year plan to try to beat some five-year or 10-year goal, you cannot be as nimble,” he says.
“I think we will continue to grow, and in the analytics market space or the artificial intelligence space, a lot of vendors are claiming they do AI, but anybody that does analytics can claim to be AI because it is not a hard thing to do.”
He concludes: “People are using the word left and right, as a substitute for analytics, but we actually do a lot of AI ourselves. This is going to be primarily used in the voice-to-text area. Google does a great job of that, and when you ask Google a question, they get 99% of my questions correct.”
