IT leaders know the significance of ITSM, which aims to keep IT operations up and running 24/7. However, the service desk managers continue to realize the pain of working with monolithic and traditional ITSM.

The convergence of AI into ITSM gives them some opportunity to leverage automation and turn the traditional platform into AISM or AITSM.

Although AISM helps facilitate internal operational resilience and improve user experience, they sometimes have difficulties achieving overall ITSM efficiency.

Well, there is a new superpower in the block, Generative Artificial Intelligence or Generative AI for ITSM, making IT service management highly efficient, automated, adaptive, and scalable.

From creating new protein designs to generating new content and writing codes for new software or website design to summarizing long texts or drafting emails, Generative AI has everyone exploring different use cases to improve business outcomes.

Depending on its capability to generate new and innovative data and create human-like responses, Generative AI provides great opportunities for ITSM leaders to leverage the unprecedented benefits of the underpinned technology of GenAI—large language models and exponentially boost service management experience.

ITSM or IT service management describes critical steps, processes, or actions for employees or service desk agents to handle issues.

ITSM generally encompasses guides to help people log issues, communicate, escalate, or self-repair problems using knowledge resources or tools.

In a typical traditional scenario, a user can handle a common problem, such as a request for a new laptop or apply for leave using predefined forms or knowledge bases.

However, if it is a Tier-1 problem, agents must handle it manually by going through log reports, analyzing its importance, prioritizing, triaging, and finally addressing it.

With AISM or AITSM, the fundamental structure remains similar to the traditional one except for its AI-powered automation.

Automation to streamline ITSM workflows is known as AISM or AITSM.

Self-service automation removes manual efforts, reduces workloads on human agents, and boosts the pace of problem resolution in many instances.

However, AISM or AITSM can evolve to huge potential with the application of Generative AI, which augments the existing state of automation and extends the pace of task management.

According to Salesforce, 86% of IT leaders expect Generative AI to soon play a prominent role in their organizations, and 57% believe it is a game changer.

Another study from Pulse Report observes that 62% of organizations have already implemented AI into their ITSM strategy.

It is as simple as AISM or AITSM. Generative AI in ITSM refers to the application of Generative AI to ITSM processes to enhance and transform IT service delivery and efficiency.

Generative AI can efficiently expand ITSM's current capabilities into more redefined structures that remove manual work and redundancies by allowing employee support to leverage contextual yet consolidated data to speed up information discovery and resolve problems.

In essence, Generative AI aims to drive overall efficiency for ITSM, remove automation friction from self-service, elevate agents’ productivity and efficiency, and expedite growth.

Traditional ITSM is static and reactive. It builds cases using historical data and provides solutions based on known scenarios. It responds to post facto issues, meaning it can suggest recommendations only after downtime. When sculpting the best strategies, it relies on historical data.

The threat landscape is ever-evolving with traditional ITSM—

• The same type of issues are repetitive and frequent

• Downtime-related delays stay longer

• Self-service lacks flexibility and convenience

• Growing workloads on service desk agents across all tiers

For example, training large language models with millions of IT issues and internal enterprise knowledge bases such as ITSM or ESM becomes an easy workaround for enterprise leaders to enable ITSM automation.

• It can automate workflows for predefined scenarios

• Cross-functional workflows still need human assistance

• Knowledge articles remain static and lead to self-service friction

• Manual efforts continue to stay except for a few cases

• Agents work hard to capture data and prepare reports

Generative can move IT leaders out of the evolving challenges of traditional or AISM problems and bring a transformative change to ITSM.

One of the key differentiators of Generative AI is neural networks and deep learning, subsets of neural networks with many layers depicting human brain functions. Built upon large language models, neural networks enable Generative AI to process vast amounts of data, recognize patterns, and mimic the human brain’s technique to create memories and evolve to react to unknown scenarios.

Given that GenAI replicates the human brain, it can understand natural language and perform certain tasks exceptionally.

This further makes Generative AI learn patterns in given datasets without relying upon explicit programmed rules like conventional AI. As a result, if an unknown scenario appears, GenAI can leverage the existing training datasets and create new yet innovative content pieces or responses that maintain the similar characteristics of the given datasets.

Let’s say, in the context of ITSM, you build your data repository using ITSM communications between a user and an agent. GenAI can learn to replicate this pattern and use the existing data to create new responses for different situations.

Suppose a request for a new headset allotment comes up. In that case, the GenAI interface can check previous conversations on a similar topic and generate contextual and human-like responses to help in a meaningful way.

Given that GenAI generates intent and context-based answers, they do it by using GANs or VAE algorithms.

• GANs, or Generative Adversarial Networks, consist of generator and discriminator nodes. The generator attempts to generate data while the discriminator continuously assesses the generated output's authenticity, ensuring the validation of the produced data.

• VAEs or Variational Autoencoders in GenAI encode a piece of data to redistribute or restructure it in various outputs, maintaining the original properties, yet in new and novel shapes.

Using GANs and VAEs, GenAI can create novel and realistic outputs.

Again, GenAI is able to simplify complexity in prompts by turning them into embeddings or numeric representations using vector databases, thus increasing the pace of generating intent and context-based responses or content.

The ability to recognize patterns, simulate new scenarios, and create new data helps ITSM at large scale to predict upcoming threats and mitigate them before they escalate.

This helps service desk managers complement ITSM objectives to deliver best-in-class ITOps or employee support experience.

Nancy Gohring, research director for ESM Observability and AIOps Software at IDC, claimed that ITSM or ITOps vendors already apply Gen AI to different use cases to improve tool adoption, accelerate ITSM response, and expand use cases.

Let’s know the existing challenges of traditional ITSM and address them with GenAI.

Use Case # 1 - Predict and triage issues automatically in real-time

Challenge: Historically, ITSM is designed to alert the service desk after an issue has occurred. Unfortunately, it involves subject matter experts spending time understanding the case and its urgency and then escalating it to the right team. Not only does this delay triage and resource allocation, but it also prolongs resolution. The manual intervention does really pain a lot.

Generative AI solution:

With the power of simulation, GenAI can visualize upcoming threats to user experience before they escalate.

For example, suppose an employee's app expires in a day or two. In that case, GenAI can notify and alert him to take appropriate steps to mitigate lost productivity using existing knowledge articles or actions provided previously.

Besides self-help augmentation, if it is a critical scenario like an app outage, GenAI can simulate it and build predictive analytics in advance.

With the ability to leverage the perpetual learning curve, GenAI can learn and recognize patterns to strengthen its understanding. With each interaction and utilizing history, it can automatically refine its prediction and offer better suggestions to triage, escalate, and mitigate issues in real time.

Use Case #2 - Generate a clear and concise incident log report for tickets with GenAI

Challenge: Providing a clear log report of the IT incident is critical to reducing the mean time to respond and the mean time to resolution or MTTR.

However, IT alerts can be difficult to interpret and offer a simple step for IT responders to take. The challenge in extracting context out of IT incident alerts lies within 一

• Lack of fluency in observability languages for a variety of correlated and disconnected alerts and events

• Multiple alerts from a large application stack for one incident

• Delayed analysis of what is impacted and what is the root cause because of manual synthesis

• Improper large language model training for correlated alerts, including insufficient alerting enrichment with proper metadata

As a result, an IT incident log report can contain misinformation and not include the change update to provide accurate information for root causes and systems impacted. Instead of the right responder checking on the issue, the wrong team may devote time and inadvertently leave the issue unchecked.

Generative AI solution:

When Generative AI is used for ITSM, large language models must be trained on with millions, if not billions, of parameters related to enterprise ITSM information, i.e., IT tickets, IT alerts, change data, SME insights, and alerts data from configuration management database system (CMDB) to pull information across complete infrastructure and topology.

When an event occurs, Gen AI uses its large language model to combine a variety of alerts and data points into one integrated rich extraction of technical information and summarize them into easy-to-digest inputs.

Integrating the ITSM platform with LLMs delivers summarized yet accurate IT incident information to the service desk agent or ITSM manager. This reduces time to visualize key information for root causes and systems impacted and escalate it to the right team to mitigate the incident.

Use Case # 3 - Create a summarized version of the action taken for stakeholder communications

Challenge: Traditional ITSM requires massive manual effort and time to create a summary of an issue's mitigation report. Service desk agents must combine information from disparate tools, check with people for real-time information capture, and manually create summaries that otherwise require writing skills; hence, it often feels like a burden.

Generative AI solution:

Generative AI chatbots can seamlessly feature an agent window for ITSM. Since GenAI features capabilities such as summarization and content creation using its LLM models, it can be trained to give natural language responses to agents and help them create summaries and other attributes in real time.

By sending a request, agents can generate a summary of an issue addressed and capture user experience to share it with stakeholders or create reports.

Use Case #4 - Continuously upgrade ITSM knowledge articles with Generative AI

Challenge: Traditional ITSM requires continuously fine-tuning knowledge articles to offer real-time information and help improve employee support efficiency. Unfortunately, traditional ITSM creates a small volume of data that can be used as a reference to update knowledge bases.

Generative AI solution:

‘With generative AI capabilities, knowledge workers can benefit from the power of the first draft,’ said Julie Mohr, a senior analyst at Forrester.

Elaborating on the above aspect of knowledge creation through the first draft could refer to the ability to learn from ticket resolution experiences continuously and offer unique insights for specific IT incidents rather than limiting them to subject matter experts only.

Through LLM training, an automated workflow can be created to combine various data points, such as incident logs, a solution provided, and the root cause analyzed, to prepare a draft. This approach can remove the knowledge gap even if incidents increase daily. With a manual approach, there are fewer chances for a unique resolution to be converted into a knowledge article.

Generative AI eliminates the need to create a knowledge article from scratch. Instead, it provides enough opportunities to create unique solutions for IT responders through rapid approval of drafts in real-time, thus efficiently facilitating knowledge management and ensuring the delivery of current information to users.

Additionally, by helping reduce time in creating knowledge resources for ITSM, Generative AI helps IT leaders conform to ITIL 4 policies and maintain a healthy ITSM infrastructure.

Use Case #5 - Reduce service desk workload with Gen AI-powered self-service ITSM support

Challenge: The chatbots in traditional ITSM follow predefined scripts, limiting the capacity of the self-service bots to mitigate common problems. AISM also meets a similar fate. Hence, most simple ITSM issues related to password resets, PTOs, VPN settings, and others escalate to agents and increase their workloads.

Generative AI solution:

By creating AI workflows for as many as 80% of repetitive IT support tasks, enterprise leaders can easily streamline processes, reduce the workload of the IT helpdesk, and reimagine support services for speedier end-to-end IT service delivery.

For example, training large language models with millions of IT issues and internal enterprise knowledge bases such as ITSM or ESM becomes an easy workaround for enterprise leaders to enable ITSM automation.

As a result, you can build a modern ITSM with conversational AI and GenAI, which empowers your users to use a hyper-automated self-service capability. Using LLMs or ChatGPT-like architecture, your self-service can refer to historical data and current context to understand query patterns and supply information for an unexpected situation.

As a result, GenAI-powered self-service can answer new questions even if predefined dialog templates are not set up.

Additionally, with the ease of flexibility to connect GenAI and conversational AI-powered chatbot to a familiar collaboration channel like MS Teams, Slack, or web widget, employees get answers to queries anytime without needing to know login credentials.

For example, if a user has a problem with the monitor, she can directly communicate in a chat interface about her issues. The process can be automated and streamlined by creating app workflow automation to reduce manual intervention. As a result, she can get a replacement, or if circumstances allow, a new purchase order could be initiated.

There are more instances where you can create app workflow automation and streamline repetitive and mundane IT support.

Create a workflow to allow your users to reset passwords

Enable users to manage leaves through workflow automation

Allow workflow automation to update payment information for credit cards or new bank accounts

Workativ offers a unique approach to facilitating ITSM automation through its conversational AI platform backed by Generative AI to allow faster time to market, enabling you to ramp up the employee experience while improving customer satisfaction.

Connect with Workativ Sales Experts to explore opportunities across the ITSM space.

Use Case #6 - Improve knowledge search for ITSM issues using Generative AI

Challenge: Be it traditional ITSM or AISM, these platforms accommodate knowledge articles or data repositories for known scenarios. Oftentimes, they feature data of certain knowledge cut-off periods. It can mean self-service can have scalability challenges, which can see serious consequences for users retrieving stale or rudimentary answers and continuing to face downtime

Generative AI solution: GenAI seamlessly uses Retrieval Augmented Generation (RAG) to extract accurate information across an external database in case CAI fails and retrieves a response that matches a user's intent.

Regarding RAGs, GenAI helps employees communicate with third-party data resources and find answers that match their queries. It further means that when GenAI and conversational AI utilize the RAG approach, it helps improve search performance for company-specific queries, find accurate answers, be able to solve problems, and increase productivity.

Conversational AI virtual assistant solution providers such as Workativ integrate the power of large language models through Hybrid NLU, which improves faster chatbot development and knowledge search for ITSM issues.

1. Effortless problem-solving

Generative AI enables the development of unique knowledge base management at scale, offering comprehensive knowledge solutions to improve self-help capability and rapidly solving IT service desk issues.

2. Enhanced user productivity

With unsupervised learning capabilities, Generative AI seamlessly improves intent detection entity, which enhances user search experience, helps derive the right information at speed, and pushes for greater productivity levels without any obstacles.

3. Decline in service desk requests

Human-like conversations enhance self-service capability for most mundane IT support issues. By providing real-time auto-resolutions for users, Gen AI reduces the workload on the service desk by eliminating IT tickets.

4. Reduction of operational costs

Generative AI does the heavy lifting on your balance sheet. By automating workflows for self-help IT support, the technology reduces the reliance on agents and helps you minimize costs for L&D for IT agents. With that brilliant cost savings, you can allocate resources to improve ITSM response.

5. Enhanced user experience

Generative AI-powered ITSM empowers users to resolve issues independently. The chat interface, which retrieves responses in real-time by simulating human language, enhances user interactivity with the platform and improves adoption.

Generative AI is notorious for creating novel challenges for users. There are many risks that ITSM can face from Generative AI, and they can impact user experience. However, a careful approach to using Generative AI in ITSM can help you leverage its benefits and augment employee support capabilities.

Communication biases:

Generative AI trains on massive datasets. If ITSM data contains bias for a particular community, race, or gender, GenAI can display biased behavior and ruin workplace harmony.

For example, if a particular dataset contains scenarios where a bot prefers male candidates and rejects applications from female candidates, GenAI in ITSM can learn and depict this behavior. Further, it may prefer not to communicate with female users whenever they have questions.

The best approach is to clean datasets to remove biases and use them to train workflows.

Security concerns:

There is a high chance that you can expose your company data to GenAI during RAG training for ITSM workflows. These datasets can include confidential HR policy, employee salary structures, or client details. If knowledge articles inadvertently use these pieces of information to generate answers, GenAI can pose huge threats to companies.

A thorough data check is essential before you prepare your data pipeline for training. Ensure that humans are kept in the loop for continuous data cleaning oversight and improvement.

Discrepancies in communications:

Generative AI can sometimes produce faulty or arbitrary answers based on assumptions. This can happen if there is not enough data to train LLMs.

There can be discrepancies with the quality of answers GenAI produces.

Your employees should be trained enough to recognize faults in communications and flag them before they harm your company culture.

At the same time, you must build a continuous feedback loop to report discrepancies and take appropriate action to prevent further damage.

Although some limitations can result in ethical issues due to hallucinations, let’s not forget that time is the best element for learning and improving. With continuous monitoring and training, GenAI can improve response delivery and accuracy, thus simplifying knowledge sharing for the self-service capability that helps resolve ITSM issues at scale.

Generative artificial intelligence is making significant strides to transform ITSM for enterprise use cases. With that, IDC suggests that more and more ITSM and ITOps vendors are eagerly assimilating the power of generative AI into various use cases. This pinpoints the growing interest of enterprise leaders in adopting GenAI and transforming employee support and experience.

By providing a massive set of IT resolution knowledge across internal enterprise settings, Generative AI promises to improve efficiency in IT service in real-time and eventually improve user satisfaction, which ensures operational resilience and drives business success.

Want to know the best way to leverage Generative AI for your IT Service Management? Ask Workativ for a personalized demo today.

How does Generative AI handle complicated and IT domain-related incidents received from alerts and events from various systems?

To gain customized capabilities from a GenAI-powered IT incident tool, you must train LLMs using millions of billions of data related to industry-wide IT incidents, recommendations provided, historical data of IT tickets, alerts, and change data, including SME insights. By combining various alerts and event data points into an integrated summary, Generative AI can help you visualize the root cause of an impacted system, enabling you to make data-driven decisions for faster incident resolution and minimizing business impacts.

Does implementing Generative AI for ITSM raise potential risks or drawbacks, such as data privacy or security?

Undoubtedly, Generative AI unlocks significant benefits for ITSM users. However, the potential risk factors of data privacy and security cannot be denied while using Generative AI to surface IT incident management responses. This requires robust data-handling practices to mitigate data breaches and protect sensitive information, whereas continuous monitoring and training are of utmost necessity to tackle hallucinations and inaccuracies.

Is it seamless for Generative AI to adapt to different organizations’ unique needs and workflows, or is it necessary to train or customize it for seamless adaptation to business needs?

Generative AI can learn from experience and thus adapt to new requirements and workflows of different organizations. However, the customization on top of the LLM model can enable Generative AI to unlock better efficiencies for ITSM needs. Organizations need to evaluate the extent of customization for effective implementation, which also requires significant investment and integration efforts to maximize performance across the ITSM environment.