Artificial Intelligence (“AI”) and Machine Learning (“ML”) technologies have been available for some time as tools to assist businesses in improving their understanding of data.  We are now on the cusp of a new era – one in which AI/ML capabilities will likely become pervasive and fundamentally impact the way enterprises use software to deliver products and services, operate their businesses and interact with customers and employees. 

The large and growing AI/ML market spans cloud providers with broad offerings, numerous open-source tools and functionality embedded into incumbent enterprise applications. However, we believe there will be compelling growth and value creation opportunities for mid-market infrastructure software companies that deliver solutions that help enterprises build, deploy and manage this powerful technology. Specifically, the DataOps sector – solutions that improve the efficiency, speed, and reliability of AI and ML processes – will likely be ripe for growth investment and strategic M&A activity. 

Fig. 1: DataOps 

Current market landscape 

1. Why businesses need AI/ML

Digital transformation makes businesses more reliant on data - AI/ML solutions are required to keep up with the rapid growth in data volumes, velocity and types.

Businesses are experiencing an explosion of data as they move forward with digital transformation initiatives. The quantity of stored digital content such as files, records, and databases and streaming data such as logs, events, and other metrics are growing exponentially. Collectively, these types of data can be harnessed by businesses to answer questions, solve problems, improve performance, increase efficiency, and drive automation.  Given the enormous volumes and high velocity of this data, it has become impossible for humans alone to fully leverage the value of all this data effectively.  AI/ML technologies have emerged to fill the gap. 

Subsequently, the market for AI/ML solutions is large and growth is increasing.  Gartner estimates that the market for AI software will be $134.8 billion in 2025 with a five-year compound annual growth rate (CAGR) of 24.5% with annual growth accelerating from 14% in 2021 to over 31% in 2025.[1]

2. How enterprises can use AI/ML

While generative AI offerings such as ChatGPT are currently capturing substantial attention and interest, that is only one of several ways for enterprises to leverage AI/ML:

  • Generate new data - OpenAI’s ChatGPT is an example of an AI model that learns to make predictions or generate outputs based on patterns and relationships in data. There are numerous other examples of using AI models to generate data such as responses to customer service requests, authoring emails, and making recommendations among others.
  • Better understand data - Optum Inc., a healthcare technology and services company, offers solutions that leverage AI to analyze medical claims data to identify fraudulent billing practices reducing fraud financial losses and dramatically improving the efficiency of fraud detection processes.
  • Automate tasks – Coupa Software Inc. offers invoice automation software that uses ML algorithms to extract data from invoices, automatically match invoices to purchase orders and receipts, and route invoices for approval and payment reducing the resources required to process invoices while increasing accuracy and efficiency.

Over time, we can expect that AI/ML solutions will be deployed in a wide range of industries and across a variety of use cases such as knowledge management, virtual assistants, autonomous vehicles, digital workplace, customer service, HR, analytics/BI, marketing, sales, and others to create seamless workflows, deliver information and operational transparency, and provide automation capabilities.  

3. The challenge of deploying AI/ML at scale

AI/ML adoption is complex, and enterprises will experience many of the same challenges as those posed by adoption of other innovative technologies such as public cloud.

The process of building and deploying an AI/ML solution is complex.  It involves the collection and transfer of data, building and training a model, deploying the model into a production environment, and exposing the model output through an interface or via integration with an application or workflow.  The underlying infrastructure resources are also complex and include cloud, edge, on-premise and hybrid alternatives. 

Fig. 2: AI/ML solution workflow

There is a large, noisy landscape of vendors offering capabilities to take advantage of AI/ML solutions including major cloud providers (AWS, Azure, Google), incumbent enterprise software vendors that integrated AI/ML functionality into their existing offerings and myriad open-source projects around AI/ML.

However, the process of building and deploying a scalable production AI/ML offering remains a significant challenge and exhibits a number of the same characteristics as the early days of enterprise adoption of public cloud computing:

  • Lack of skills and expertise - Similar to cloud native technologies, the number of people with specialized skills in AI/ML technologies is limited and not growing as fast as market demand for the solutions.
  • High costs - Implementing AI/ML requires resources that generally are not required in other areas of IT including high performance hardware, specialized software, and resources to train and maintain the systems.
  • Lack and location of data - Businesses need large amounts of high-quality data to build and train AI/ML models and require access to large volumes of data once in production.
  • Complexity - Implementation processes can be time-consuming and resource-intensive for organizations that are already strapped for IT resources and budget.

Emerging opportunities for growth investment and M&A activity

Delivering DataOps functionality provides opportunities for innovative software companies to leverage rapidly growing demand for AI/ML and create substantial value by enabling users to overcome challenges of adopting AI/ML and to optimize the value of AI/ML investments. 

We believe that enterprises will increasingly seek out solutions that enable them to overcome these challenges and take full advantage of powerful AI/ML capabilities.  DataOps is emerging to meet this demand.  DataOps solutions including both software and managed services to help enterprises by improving the efficiency, speed, and reliability of AI/ML processes across data engineers, data scientists and IT operations.

Fig. 3: DataOps solution categories 

  • Collaboration - Enable collaboration between data engineers and data scientists, as well as other stakeholders such as IT and business teams, to ensure that data-driven processes are aligned with business goals and objectives.
  • Continuous improvement - Leverage feedback and data to iteratively improve processes and outcomes over time.
  • Quality and governance - Ensure that data is accurate, consistent, and compliant with relevant regulations and standards.
  • Automation - Streamline and standardize data preparation, model training, and model deployment processes to drive operational efficiency.
  • Monitoring and management - Track model performance including metrics such as accuracy and precision, make updates or adjustments to the models, track changes to models and roll back to previous versions if needed.

We are tracking an emerging landscape of software and managed services companies targeting the DataOps opportunity and believe it will become a key category of the broader infrastructure software sector. 

Contact DC Advisory’s Infrastructure Software expert Joel Strauch here >

[1] Source: Gartner Forecast Analysis: Artificial Intelligence Software, Worldwide October 2021

For important information regarding this publication, please see our Insights and Publications disclaimer >

 

What to read next