{"id":26604,"date":"2022-12-29T23:45:30","date_gmt":"2022-12-29T23:45:30","guid":{"rendered":"https:\/\/mailinvest.blog\/index.php\/2022\/12\/29\/ai-goes-mainstream-but-return-on-investment-remains-elusive\/"},"modified":"2022-12-29T23:45:30","modified_gmt":"2022-12-29T23:45:30","slug":"ai-goes-mainstream-but-return-on-investment-remains-elusive","status":"publish","type":"post","link":"https:\/\/mailinvest.blog\/index.php\/2022\/12\/29\/ai-goes-mainstream-but-return-on-investment-remains-elusive\/","title":{"rendered":"AI goes mainstream, but return on investment remains elusive"},"content":{"rendered":"

<\/a>\n<\/p>\n

\n

A decade of big data investments, combined with cloud scalability, the rise of more cost effective processing and the introduction of advanced tooling, has catapulted machine intelligence to the forefront of technology investments. No matter what job you have, your operation will be AI powered within five years and machines may be doing your job in the future.<\/p>\n

Artificial intelligence is being infused into applications, infrastructure, equipment and virtually every aspect of our lives. AI is proving to be extremely helpful at controlling vehicles, speeding medical diagnoses, processing language, advancing science and generally raising the stakes on what it means to apply technology for business advantage.<\/p>\n

But business value realization has been a challenge for most organizations because of a lack of skills, complexity of programming models, immature technology integration, sizable up front investments, ethical concerns and lack of business alignment. Mastering AI technology and a focus on features will not be a requirement for success in our view.<\/p>\n

Rather, figuring out how and where to apply AI to your business will be the crucial gate. That means understanding the business case, picking the right technology partner, experimenting in bite-sized chunks and quickly identifying winners on which to double down from an investment standpoint.<\/p>\n

In this Breaking Analysis, we update you on the state of AI with a focus on interpreting the latest Enterprise Technology Research survey data around machine learning, AI and data. We\u2019ll explore what it means for the competitive environment and what to look for in the future. To do so, we invited into our studios Andy Thurai, vice president and principal analyst at Constellation Research Inc. Andy covers AI deeply, and he knows the players and the pitfalls of AI investment.<\/p>\n

AI assumptions<\/h3>\n

\"\"<\/p>\n

According to McKinsey, AI adoption has more than doubled since 2017, but only 10% of organizations report seeing significant return on equity (BCG and MIT).<\/p>\n

Part of the challenge of AI is it requires data, good data and data proficiency, which is nontrivial. Firms that can master both data and AI will have a competitive advantage this decade.<\/p>\n

The hyperscalers such as Amazon Web Services, Microsoft Corp. and Google LLC, as we\u2019ll show you with the ETR data, dominate AI\/ML mindshare, market share and share of wallet, and will likely continue to gain traction.<\/p>\n

Having said that, there\u2019s plenty of room for specialists, but they need to partner with cloud vendors for go to market productivity.<\/p>\n

Finally, organizations will increasingly have to put data and AI at the center of their enterprises, and to do that, most will have to rely on vendor research and development to leverage AI\/ML.\u00a0In other words, they\u2019ll buy it and apply it, rather than build it.<\/p>\n

Here\u2019s how Andy sees it:<\/p>\n

\n

First of all, the stat about only 10% of customers realizing a return on investment: That\u2019s so true because most companies are still in the innovation cycle. So they\u2019re trying to innovate and see what they can do to apply AI. Many times a model will be created to experiment\u00a0and there will be a big focus on the technology, but often there\u2019s not a good business case to solve. So they experiment, they realize the model is working and then they try to figure out where to apply it. So it\u2019s like you found a hammer and then you\u2019re trying to find the nail to hit kind of thing. That rarely works.<\/p>\n<\/blockquote>\n

Listen to Andy Thurai\u2019s additional comments on the premise.<\/strong><\/a><\/p>\n

Many customers easing up on strategic investments, including AI<\/h3>\n

Let\u2019s look at where ML\/AI fit relative to the other hottest sectors in the ETR data set.<\/p>\n

\"\"<\/p>\n

\u00a0<\/p>\n

The above XY graph shows Net Score or spending velocity on the vertical axis and presence in the survey, called Sector Pervasion, for the October survey. The squiggly line on ML\/AI represents the progression since the Jan. 2021 survey and you can see the downward direction. We position ML\/AI relative to the other hot sectors, Containers, Cloud and Robotic Process Automation, which have consistently performed above the magic 40% red dotted line for most of the past two years. And we\u2019ve included Analytics\/Big Data for context and because it\u2019s a relevant adjacency.<\/p>\n

Note the green arrow moving toward 40%. ETR gave us a glimpse of the January survey, which is in the field with currently more than 1,000 responses and AI is breaking the downward trend.<\/p>\n

Andy made the following key points about this data:<\/p>\n

\n

One of the things you have to keep in mind is when the pandemic happened, companies went into survival mode. So when somebody\u2019s in that mode, what happens? The luxury and the innovations get cut. And this is exactly what happened with AI. So as you can see in the last seven quarters, which is almost dating back close to pandemic, everybody was trying to keep their operations alive, especially digital operations. How do I keep the lights on? So while the numbers spent on AI\/ML is moving down on a relative basis, I still think embedded AI\/ML spend is happening but may not be as visible. Things like employee experience or ITOps, AIOps, MLOps\u2026 some of those areas actually went up. There are companies like Atlassian, even though it had some platform issues,\u00a0 still saw spending momentum because they are offering the solution that was not previously available. So there are companies out there doing embedded AI for things like incident management for example that may not show up in this data. A lot of companies using \u201chidden\u201d AI\/ML in some of those areas are growing unbelievably well.\u201d<\/p>\n<\/blockquote>\n

Listen to why Andy Thurai thinks AI is seeing some headwinds.<\/strong><\/a><\/p>\n

AWS, Microsoft and Google are dominant in ML and AI<\/h3>\n

Databricks stands out as the notable independent<\/h4>\n

Let\u2019s now unpack the ML\/AI sector in the ETR data and look at the players that hit the radar in the survey.<\/p>\n

\"\"<\/p>\n

The chart above shows the same XY dimensions and the players in the ML\/AI sector within the survey. The table insert in the above graphic shows how the companies are plotted \u2013 Net Score and N\u2019s in the survey. The hyperscalers are dominant in both market presence and spending momentum on their AI tools. Databricks Inc. is right in the mix, showing strong as a specialist. Then you go to a of seven AI-focused players including Anaconda, DataRobot Inc., H2O.ai Inc.,\u00a0Dataiku Inc., C3.ai Inc., SparkCognition Inc., and OpenText Corp. with Magellan. Then there\u2019s the large, legacy enterprise players, Oracle Corp. and IBM Corp., which have AI offerings.<\/p>\n

To the earlier point made by Andy, companies such as Salesforce Inc. and ServiceNow Inc., which offer AI tooling as part of their enterprise software offerings,\u00a0don\u2019t show up in the ETR ML\/AI taxonomy. it\u2019s similar with CrowdStrike Holdings Inc. and Palo Alto Networks Inc. in security, or Dell Technologies Inc. and Hewlett Packard Enterprise Co. in infrastructure, Splunk Inc. and Dynatrace Inc. in observability, and RPA specialists such as UiPath Inc. All of these companies and many others are actively embedding machine intelligence into their platforms and they won\u2019t show up in the ETR ML\/AI sector.<\/p>\n

Back to the premise: As technology vendors invest in AI through R&D to make their products more intelligent, benefits will confer to customers. But differentiation and direct ROI will require more than just deploying vendor products. Thinking through process improvements, human capital management and the overall business case will be critical.<\/p>\n

Here\u2019s how Andy interprets the hyperscaler landscape:<\/p>\n

\n

The first point is AWS is more about Lego building blocks. We\u2019ll give you all the components that you need. Whether it\u2019s SageMaker, MLOps or CodeWhisperer\u2026 whatever you want. They\u2019ll give you the blocks and then you\u2019ll build things on top of it. Google took a different path. Google did a lot of work with their acquisition of DeepMind and other things. They\u2019ve been way ahead of the pack when it comes to AI tech. Now, I think Microsoft\u2019s move of partnering and making a huge investment in OpenAI is unbelievable. Of course, everybody\u2019s talking about ChatGPT.\u00a0 Remember, as Warren Buffet says, that when my laundry lady is talking to me about stock market, it\u2019s heating up, and that\u2019s the way it is with ChatGPT.<\/p>\n<\/blockquote>\n

Listen to Andy Thurai\u2019s take on the hyperscaler competitive landscape in AI and the buzz around ChatGPT.<\/a><\/strong><\/p>\n

As it pertains to the smaller independents, they must partner with hyperscalers for a variety of reasons, especially go-to-market productivity.<\/p>\n

Listen to Andy Thurai discuss why smaller players have no choice but to partner with hyperscalers.<\/a><\/strong><\/p>\n

AI decision points are shifting from technical features to ROI<\/h4>\n

Before we move on, let\u2019s share some other findings from ETR on why people adopt AI. The data historically shows that: 1) feature breadth; and 2) technical capabilities were the main decision points for AI adoption. According to Andy, that\u2019s changing:<\/p>\n

\n

I can guarantee you, if you look at the latest data that\u2019s coming in now, those two will be secondary and tertiary points. The No. 1 would be about ROI. And how do I achieve it? I\u2019ve spent ton of money on all of my experiments. This is the same theme I\u2019m seeing across the board when talking to everybody who\u2019s spending money on AI. I\u2019ve spent so much money on it. When can I get it live in production? How quickly can I get a return? Because the board is breathing down their neck. You already spent this much money. Show me something that\u2019s valuable. So the ROI is going to become \u2014 I\u2019m predicting this for 2023 \u2014 the No. 1 factor in AI decisions.<\/p>\n<\/blockquote>\n

Listen to Andy Thurai\u2019s prediction about how buyers\u2019 decision factors will change in 2023.<\/strong><\/a><\/p>\n

Customer spending patterns for the top AI players<\/h3>\n

Hyperscalers show low churn in AI\u2026 not the case for Oracle and IBM<\/h4>\n

Let\u2019s look at some of the top players in AI and break down their spending profiles.<\/p>\n

\"\"<\/p>\n

The chart above shows how Net Score is calculated. Pay attention to the second set of bars showing Databricks where we\u2019ve annotated the colors. The lime green is new adds, the forest green is spending up 6% or more, the gray is flat spend, the pinkish is spending 6% down or worse, and the red is churn. Subtract the reds from the greens and you get Net Score, which is shown by the blue dots highlighted by the arrows.<\/p>\n

AWS has the highest Net Score and very little churn but notably, Databricks and DataRobot are next with Microsoft and Google also showing very low churn.<\/p>\n

But Oracle (N=46) and IBM (N=62) are showing much higher platform defections in the ETR survey.<\/p>\n

Andy had some pointed comments regarding IBM Watson:<\/p>\n

\n

A couple of things that stands out to me.<\/span>\u00a0Most of them are in line with my conversation<\/span>\u00a0with customers.<\/span>\u00a0One is\u00a0how bad IBM Watson is doing.\u00a0<\/span>So look,<\/span>\u00a0IBM has been at the forefront of innovating<\/span>\u00a0for many, many years now.<\/span>\u00a0And over the course of time,\u00a0<\/span>they moved from a product innovation-centric company<\/span>\u00a0to more of a services company.<\/span>\u00a0And they began\u00a0<\/span>making a majority<\/span>\u00a0of revenue from services.<\/span>\u00a0Now things are changing. Arvind [Krishna] has taken over,<\/span>\u00a0he came from research.<\/span>\u00a0So he\u2019s doing a great job<\/span>\u00a0of trying to reinvent IBM as a company.<\/span>\u00a0But they have a long way to catch up.<\/span>\u00a0IBM Watson, if you think about it, played J<\/span>eopardy like, what 15 years ago?<\/span><\/p>\n<\/blockquote>\n

Listen to Andy Thurai explain why Watson failed to live up to its early high expectations.<\/strong><\/a><\/p>\n

The surprising performance of DataRobot<\/h4>\n

Andy had these additional thoughts on the other AI players shown on the chart above:<\/p>\n

\n

AWS is No. 1, as it should be.<\/span>\u00a0<\/span>But what what actually caught me by surprise is<\/span>\u00a0<\/span>how DataRobot is holding.<\/span>\u00a0<\/span>I mean, look at that.<\/span>\u00a0<\/span>The net new addition [lime green] and\/or expansion [forest green],<\/span>\u00a0<\/span>DataRobot seems to be doing well.\u00a0<\/span>That surprises me.<\/span>\u00a0<\/span><\/p>\n<\/blockquote>\n

Listen to Andy Thurai discuss the momentum of AWS the surprise that is DataRobot.<\/a><\/strong><\/p>\n

Emerging AI players vie for mindshare<\/h3>\n

Databricks is the dominant private firm<\/h4>\n

We hit the bigger names in the ML\/AI sector, so now let\u2019s look at the emerging technology companies.<\/p>\n

\"\"<\/p>\n

One of the gems of the ETR data set is the Emerging Technology Survey \u2013 ETS. It\u2019s now run four times per year and is exclusively focused on private companies that are potential disruptors, M&A candidates and, if they\u2019ve raised enough money, acquirers of companies as well. Databricks and Snyk Ltd. would be good examples of buyers hoping to go public at some point.<\/p>\n

Above we show the emerging companies in the ML\/AI sector of the ETR data set. The dimensions are Net Sentiment on the Y axis and Mind Share on the X axis. Basically the ETS study measures awareness and intent to buy \u2013 or not buy. So ETR uses a methodology similar to Net Score where the negatives are subtracted from the positives. And Mind Share is vendor awareness shown on the horizontal axis.<\/p>\n

The inserted table shows Net Sentiment and Ns in the survey, which informs the position of the dots.<\/p>\n

You\u2019ll also notice we\u2019re plotting TensorFlow. We know that\u2019s not a company, but it\u2019s there for reference, since open-source tooling is an option for customers.<\/p>\n

We\u2019ve also drawn a line for Databricks to show how relatively dominant they\u2019ve become in the past 10 ETS surveys \u2013 going back to late 2018.<\/p>\n

You also see the position of about a dozen other emerging vendors including: Anaconda, H2o.ai, Scale AI, DataRobot, Dataiku, Domino, Hugging Face, Labelbox, ElectricfAI, Tecton, Anyscale and SparkCognition.<\/p>\n

Here\u2019s how Andy interprets this data:<\/p>\n

\n

Remember we talked about how Oracle is not necessarily the<\/span>\u00a0database of the choice [for AI]?<\/span>\u00a0Databricks is trying to solve some of the issue<\/span>\u00a0for AI\/ML workloads.<\/span>\u00a0And the problem is there is no one company that can<\/span>\u00a0solve all of the problems.<\/span>\u00a0For example, if you look at the names in here,<\/span>\u00a0some of them are database names,<\/span>\u00a0some of them are platform names,<\/span>\u00a0some of them are MLOps companies..<\/span>. a<\/span>nd some of them are feature-based companies like<\/span>\u00a0Tecton.<\/span><\/p>\n

It\u2019s a mix of those companies.<\/span>\u00a0You got Hugging Face in here,\u00a0<\/span>which does NLP.\u00a0\u00a0<\/span><\/p>\n

Most of these companies are going to get acquired.<\/span>\u00a0My prediction would be most of them will get acquired<\/span>\u00a0because look, at the end of the day,<\/span>\u00a0hyperscalers need these capabilities, right?<\/span>\u00a0So they\u2019re going to either create their own,<\/span>\u00a0AWS is very good at doing that.<\/span>\u00a0But the other ones, like Azure,<\/span>\u00a0they\u2019re going to look at it and say, \u201cYou know what,<\/span>\u00a0it\u2019s going to take time for me to build this.<\/span>\u00a0Why don\u2019t I just go and buy you?\u201d<\/span>\u00a0Or even the other players like Oracle or IBM Cloud..<\/span>. t<\/span>hey might even take a look at them.<\/span>\u00a0So at the end of the day,<\/span>\u00a0a lot of these companies are going to get acquired or merged<\/span>\u00a0with others.<\/span><\/p>\n<\/blockquote>\n

Listen to Andy Thurai\u2019s predictions on possible M&A activity in the AI market.<\/strong><\/a><\/p>\n

OK, let\u2019s wrap with some final thoughts.<\/p>\n

The AI winter is not coming<\/h3>\n

\"\"<\/p>\n

Despite the challenge of leveraging AI, we\u2019re not repeating the AI winter of the 1990s. Machine intelligence is a superpower that will permeate every aspect of the technology industry.<\/p>\n

AI and data strategies must be connected. Leveraging first party data will become increasingly important to AI competitiveness and shortening time-to-value.<\/p>\n

Getting cloud \u201cright\u201d matters in AI. Hyperscalers continue to capture more share of data, a key ingredient to AI success. They have native tools and integrated ecosystem partners that will continue to strengthen through the decade.<\/p>\n

Real-time AI inferencing (e.g. at the edge) will become an increasingly important force. It will usher new economics in silicon, new tooling, new companies, and could disrupt cloud norms by the end of the decade<\/p>\n

We gave Andy the last word and asked him to give his final thoughts.<\/p>\n

Listen to Andy Thurai summarize the state of AI in seven minutes.<\/a><\/strong><\/p>\n

Keep in touch<\/h3>\n

Many thanks to Andy Thurai of Constellation Research for his insights and participation in this Breaking Analysis segment. Alex Myerson and Ken Shifman are on production, podcasts and media workflows for Breaking Analysis. Special thanks to Kristen Martin and Cheryl Knight who help us keep our community informed and get the word out, and to Rob Hof, our editor in chief at SiliconANGLE.<\/p>\n

Remember we publish each week on\u00a0Wikibon<\/a>\u00a0and\u00a0SiliconANGLE<\/a>. These episodes are all available as\u00a0podcasts wherever you listen<\/a>.<\/p>\n

Email\u00a0david.vellante@siliconangle.com<\/a>, DM\u00a0@dvellante on Twitter<\/a>\u00a0and comment on\u00a0our LinkedIn posts<\/a>.<\/p>\n

Also, check out this\u00a0ETR Tutorial we created<\/a>, which explains the spending methodology in more detail.\u00a0Note:\u00a0ETR<\/a>\u00a0is a separate company from Wikibon and SiliconANGLE.<\/i>\u00a0If you would like to cite or republish any of the company\u2019s data, or inquire about its services, please contact ETR at legal@etr.ai.<\/p>\n

Here\u2019s the full video analysis:<\/p>\n