Data backup tools often list AI capabilities in their feature lists.
The use of AI by software suppliers in this area is not a new concept. Backup software has been automated for many years.
Backup companies use machine learning and predictive analysis, in particular, to improve backup reliability and efficiency. This includes using ML to analyse logs in order to predict where a back-up might fail and to select the best times and storage locations for backups.
In addition, suppliers use generative AI to help customers manage and set up backup tasks. Agentic AI systems promise to automate further in areas like system configuration and recovery tests. AI is most commonly used to detect and remediate ransomware.
Backup management and IT management
Already used by IT teams AI can be used for disaster recovery (DR), resource allocation, and planning . Some use AI to mine historical data to predict when a system might fail. AI “copilots”on the other hand, can perform simple but time-consuming IT tasks like updating inventories and mapping data usage across networks and applications. AI can be used to perform tasks such as storage optimization, monitoring backups – especially for malware – and compiling risk assessment for disaster recovery.
Most backup tools use AI in three different ways. This includes helping users with setup, configuration and optimising the backup process. Anomaly detection is a feature that chief information officers and suppliers are increasingly looking for in backup tools.
The increase in ransomware, and the increasing ability of ransomware gangs to target backups has forced IT teams look at how they secure their backup volumes.
The backup software should detect and remove malware at the source before data copies. This is also called in-line malware scan.
Software should scan volumes to detect anomalies or indicators. This is now offered by most, but not all suppliers. Anomaly detection can also help IT teams locate the last clean backup and prevent organisations from reinfecting their systems with a compromised file.
IT Teams can also use AI anomaly detection to detect other issues, like data corruption or accidental deletion. They may not be caused by ransomware but they can cause data recovery to fail.
The increased focus on ransomware remediation and detection is part of an overall trend where backup and recovery providers are repositioning themselves as security companies rather than IT administration tools.
According to Jon Collins, chief technology officer of GigaOm, “Backup has become a part of broader data security solutions, which include security and threat detection.” “The AI that we’ve examined tends to cover more of this aspect. Ransomware is the biggest threat to which backups are a response. This makes sense.”
But suppliers also offer a wide range of capabilities that are not related to ransomware, which can make backups and recovery more efficient. Here are some of the key AI features in backup products.
Acronis[//19659016]Acronis has been using AI since 2017, when they began working on stack trace analysis of Windows OS. They then developed an AI static file analyzer to detect variations in malware samples and behaviour engine log detection. Acronis uses AI to ensure that backups are restored correctly and boot up. Acronis tools can predict SSD and hard disk failures and use AI-based monitoring. The supplier uses its own chatbot Acronis Copilot and “conversational AI” to support incident response.
According to the supplier, using AI to perform behaviour analysis is more efficient than older pattern-matching methods to ensure backups free of malware.
Cohesity (
Cohesity) uses AI in a variety of its tools. Cohesity Turing has AI-powered ransomware remediation, and AI and ML are used for discovery and reporting. Cohesity DataHawk offers ransomware protection via AI-based threat identification, “cyber-vaulting” and a machine-learning-powered data classification tool.
Cohesity Gaia offers conversational searching that allows IT teams more easily work with backup data by using natural language. Cohesity uses AI to plan capacity. This is not limited to backup and recovery but could reduce the cost of data protection.
Commvault Commvault’s Metallic AI uses AI and ML, as well as natural language processing (NLP), for data protection and recovery. This includes AI assisted anomaly detection and “AI enabled bursting” in order to recover data faster. Commvault offers AI-based data classifying to help organisations prioritise their backups and improve compliance.
Commvault utilizes AI for both on-premise workloads as well as Commvault Cloud. It states that AI is used for threat detection, recovery and risk management.
Druva (
Druva) Druva has two AI-based primary tools: Dru and Dru Investigator. The supplier launched Dru 2023 and positions it as an easy way for IT teams improve productivity and make better decision. Dru uses a conversational user interface to simplify administration, and improve reporting.
Dru Investigate is designed to help organisations secure backups and reduce threats. This includes displaying data at risk and accelerating investigations. Dru Investigates relies on large language models, a private retrieval tool and a generation tool that is augmented by retrieval to reduce the workload of security teams when monitoring backup environments.
Rubrik’s AI Assistant
Ruby, according to the supplier, helps organisations with cyber recovery and response. This includes AI-powered anomaly detectors, as well as advice on how to isolate infected data.
Rubrik claims to use machine learning to analyze deletions, modifications, and encryption of customer information, as well to provide alerts about any threats hidden within backup data. Rubrik can identify the cleanest backups across files and snapshots. It also helps IT teams to create clean VMware instances.
VeeamVeeam offers AI-based analysis, which includes backup performance monitoring, a predictive analysis to identify potential risks, as well as ML-based ransomware detection and threat detection. Its AI continuously scans backups to detect anomalies.
This supplier uses AI to manage data and classify data. It claims that this can help make “smarter decisions” about storage. Veeam also supports the Model Context Protocol, even though it’s not a backup feature per se. This open standard developed by Anthropic allows AI models to securely access enterprise data in Veeam repositories for tasks such a model training.
