“Information-technology departments within corporations deal with a lot of busy work -- from answering tech support questions to giving employees access to laptops and phones.
That's exactly why one startup is aiming to use artificial intelligence and a "knowledge graph" [1] to automate those mundane IT tasks, building what it calls a "system of intelligence" [2] that can connect multiple data systems together to provide a comprehensive overview of what's going on inside a company's IT department.
XperiencOps, a startup also known as XOPS that has raised nearly $40 million from Activant Capital and FPV Ventures, is part of a new wave of tech vendors and venture capitalists that believe IT can be greatly streamlined by AI and other technologies.
The broader market for IT automation tools is currently dominated by vendors such as ServiceNow, IBM and Cisco's Splunk, analysts say. IT research and consulting firm Gartner predicts that by next year, 30% of enterprises will automate more than half their network activities -- that's up from less than 10% of enterprises in mid-2023.
That's because IT departments traditionally have fragmented data across various systems, making it hard for staff to track down which employees have access to which laptops and software, who should be eligible for upgrades, who should have the right access when traveling -- the list goes on.
The idea behind XOPS is that "human middleware," or individuals among a company's IT staff who manually connect data between disparate systems, have large amounts of menial work they shouldn't have to do, said Mayan Mathen, XOPS's co-founder and chief executive.
"What if we built software robots that did work that humans shouldn't have to?" Mathen said.
XOPS, which leverages software bots to manage a company's IT policies, first creates a so-called knowledge graph from diverse data sources like software systems, databases and data centers. This knowledge graph is like a database that represents information similar to the way maps do, and can show relationships between people, ideas and documents.
Stanley Toh, Broadcom's head of enterprise end-user services and experience, said the chip giant employs an IT staff of about 40 workers to support its roughly 50,000 employees. That means it needs the support of a technology solution to boost what its staff can do, he said.
To make the XOPS system work, Broadcom gave it 17 sources of data to build a knowledge graph on, including mobile billing, data-center monitoring and information security data. Taken all together, XOPS created a profile of each device within Broadcom -- making it easy for the human IT worker to track every change to an employee's laptop or phone.
Broadcom uses XOPS to autonomously manage the entire "life cycle" of an employee laptop, Toh said, replacing what was formerly a manual process in which IT staff were involved every step along the way, from selecting and delivering an employee's laptop to servicing and returning it when the employee leaves the company.
"The only time I need a human is to put the laptop in a box, slap on the shipping label and drop it at the receiving store to be shipped out," Toh said. "The other one is when the laptop comes back, they check it back in."
Toh said the benefit of XOPS is reducing the amount of downtime for Broadcom engineers who can't get work done without a functioning laptop. XOPS also enables Broadcom to eliminate the amount of unused laptop inventory in warehouses, as well as reduce the number of software licenses that aren't being used by employees.
Eliminating that software spend, Toh said, is millions of dollars each year in cost savings.
XOPS isn't alone in attempting to apply new approaches to the field of corporate IT. Prior waves of AI have given rise to the concept of AIOps for IT, or "artificial intelligence for IT operations" -- the idea that machine learning can be deployed to monitor sprawling, dispersed tech systems.
One difference now is that today's AI systems are much more powerful than the machine learning models of the past -- they are capable of "reasoning" or thinking across complicated topics, and "agentic" systems can take action on behalf of humans, said Arun Chandrasekaran, a Gartner distinguished vice president analyst.
But the problems that existed over a decade ago remain.
Enterprises and chief information officers need to keep good track of their corporate data in order for it to be ingested by AI, and IT staff must be trained alongside it to not fear a job takeover and trust the technology, Chandrasekaran added.
"There's a lot of fear that we are dealing with extremely mission-critical and business-critical systems," he said, "and even if the AI systems are capable of orchestrating some of that, the humans don't yet have the trust that AI is able to do that very effectively."
---
Belle Lin writes for WSJ Pro CIO Journal.” [3]
1. A knowledge graph is a structured way to represent information about the real world, using entities and relationships to connect facts and concepts. It's essentially a database that organizes data into a network of interconnected nodes (entities) and edges (relationships). This structure allows for more sophisticated data analysis, better search results, and the ability to uncover hidden connections between data points.
Here's a more detailed breakdown:
Key Components:
Entities:
These are the objects, people, places, or concepts within the knowledge graph. For example, in a graph about movies, entities could include "The Shawshank Redemption," "Frank Darabont," and "Drama".
Relationships:
These define how entities are connected. In the movie example, a relationship could be "directed_by" (connecting Frank Darabont to The Shawshank Redemption) or "genre" (connecting The Shawshank Redemption to Drama).
Organizing Principles:
These are the rules and categories that provide context and structure to the knowledge graph. They can be ontologies (In the context of Artificial Intelligence, an ontology is a formal representation of knowledge as a set of concepts and the relationships between those concepts within a specific domain.), semantic layers (A semantic layer acts as a bridge between technical data and business users, translating complex data models into familiar business terms.), or domain-specific knowledge (Domain-specific knowledge refers to the specialized expertise and information within a particular field or subject area, as opposed to general knowledge.).
How it Works:
1. Data Collection:
Knowledge graphs start with data from various sources, including structured databases, text documents, and even unstructured information.
2. Entity Extraction and Linking:
Data is processed to identify entities and the relationships between them. Techniques like named entity recognition and relationship extraction are used.
3. Graph Construction:
The extracted entities and relationships are then organized into a graph structure within a graph database.
4. Querying and Analysis:
Users can query the knowledge graph to find specific information, explore relationships, and discover new insights.
Examples:
Google's Knowledge Graph:
.
Google uses a knowledge graph to enhance search results, providing richer, more contextual information alongside search results.
.
Amazon uses a knowledge graph to organize its vast product catalog, enabling better product discovery and recommendations.
.
These are open knowledge graphs that provide structured information about various topics, available for use by others.
Benefits:
Improved Search:
Knowledge graphs enable more accurate and relevant search results by understanding the meaning behind search queries and the relationships between entities.
Data Integration:
Knowledge graphs can integrate data from disparate sources, breaking down data silos and providing a unified view of information.
Enhanced Analytics:
They facilitate deeper analysis of data by uncovering hidden connections and patterns.
AI Applications:
Knowledge graphs are crucial for powering various AI applications, including natural language processing, recommendation systems, and question-answering systems.
2. A system of intelligence refers to a modern approach to enterprise software that integrates data, analytics, and AI to create intelligent systems capable of making data-driven decisions and automating actions. It combines traditional systems of record (like databases and ERP systems) with real-time analytics and AI to provide insights and enable more informed and proactive business operations.
Here's a more detailed breakdown:
Key Concepts:
Integration:
Systems of intelligence integrate various data sources, including traditional systems of record, and leverage advanced analytics and AI to provide a holistic view of business operations.
Real-time Insights:
They aim to provide real-time or near real-time insights from data, enabling faster and more informed decision-making.
AI-powered Actions:
These systems can automate actions, trigger workflows, and make recommendations based on data analysis, moving beyond simple reporting to proactive intelligence.
Digital Transformation:
They are seen as a key enabler of digital transformation, helping organizations become more agile, responsive, and data-driven.
Core Components:
Data:
Systems of intelligence rely on a robust foundation of data, including structured and unstructured data from various sources.
AI and Machine Learning:
AI and machine learning algorithms are used to analyze data, identify patterns, and make predictions or recommendations.
Advanced Analytics:
These systems incorporate advanced analytics techniques, including predictive and prescriptive analytics, to extract valuable insights from data.
Natural Language Processing (NLP) and Computer Vision:
NLP and computer vision capabilities enable systems to understand and process human language and visual information, enhancing their ability to interact with users and extract insights from various data sources.
Orchestration:
Systems of intelligence often involve orchestration layers that coordinate various components and workflows to achieve specific goals.
Benefits:
Improved Decision-Making:
By providing timely and relevant insights, systems of intelligence empower users to make more informed decisions.
Increased Efficiency:
Automation and streamlined workflows can improve operational efficiency and reduce manual effort.
Enhanced Customer Experience:
By understanding customer behavior and preferences, organizations can personalize interactions and improve customer satisfaction.
Competitive Advantage:
Systems of intelligence enable organizations to differentiate themselves by leveraging data to innovate and adapt to changing market conditions.
In essence, systems of intelligence represent a shift from traditional systems of record to more dynamic and intelligent systems that can proactively drive business value through data and AI.
3. IT Departments Are Overloaded With Busy Work. Can AI Change That? Lin, Belle. Wall Street Journal, Eastern edition; New York, N.Y.. 18 Aug 2025: B4.
Komentarų nėra:
Rašyti komentarą