Playbook for Successful AI Implementation in Service and Support

22 April 2025 - ID G00826617 - 21 min read

By Francesco Vicchi

Customer service and support leaders often face pressure to quickly implement AI initiatives, which can lead to ineffective outcomes. This research focuses on four critical pillars for the success of AI projects in customer service.

Overview

Key Findings

According to the 2024 Gartner AI Mandates for the Enterprise Survey, 18% of AI leaders or executives whose organizations deployed at least one GenAI use case, report that measuring value is among their top 3 implementation challenges.
Twenty percent of AI leaders report that finding the right use case is among their top 3 barriers to AI initiative implementation.
According to the survey, the primary obstacle to implementing AI initiatives is the availability and quality of data.
Organizations that effectively manage AI-driven change initiatives often see a significant boost in business outcomes, achieving at least a 20% greater impact ranging from cost reduction to revenue growth.

Recommendations

Align leadership, including both the project’s sponsoring board and stakeholders, by collaboratively establishing business objectives, a specific metrics hierarchy, related targets and anticipated outcomes from AI.
Define a technology strategy by identifying feasible and high-value use cases and developing a strategic AI roadmap to guide the implementation process.
Ensure AI data readiness by reviewing data and knowledge architecture along with programs to ensure their ongoing hygiene.
Prepare people for change and reassure them about the future by involving them in project implementation, and establishing effective change management and training programs.

Introduction

Customer service and support (CSS) leaders are under immense pressure to lead AI initiatives and showcase their value. According to the 2025 Gartner Business Buyer Survey, 93% of service and support technology buyers emphasized the importance of AI capabilities in their tech investments.¹ Amid ambitious promises and hype from numerous technologies and vendors, CSS leaders encounter overwhelming information, further complicating their already difficult selection and implementation decisions.

Many AI projects start with technology selection, which is itself a daunting challenge given AI’s constant evolution. However, even when managed correctly, technology selection alone cannot sustain the entire project. CSS leaders, who often spearhead these AI transformation projects (indeed, 79% of service and support leaders in a recent survey indicated having substantial control over their technology selection²), must understand that successful AI implementation requires more than just coordination between technology staff and the customer service team for process design and testing.

To achieve successful AI implementations, CSS leaders must focus on four critical pillars, only one of which is technology-related:

Leadership alignment: Ensure continuous involvement and coordination in target setting and metrics.
Technology strategy: Identify use cases, develop a comprehensive deployment plan, and decide whether to build or purchase the necessary technology.
Data readiness: Review data hygiene and architecture, and evaluate the knowledge base along with the knowledge program.
People preparedness: Establish effective change management and training programs.

While these four pillars are common to many technology projects, AI projects require a unique approach due to the technology’s rapid evolution, its potential impact on people and the noise generated by the current hype.

Moreover, while the pillars are often discussed in project meetings as concepts or intentions, they are rarely translated into tangible actions or structured efforts. This research provides AI-specific, actionable recommendations to guide CSS leaders toward successful AI implementation, following the structure of the four pillars mentioned above, analyzing each according to three fundamental aspects, and providing recommendations for each of these (see Figure 1).

Figure 1: AI’s Four Building Blocks and 12 Recommendations

Analysis

Align Leadership

Set Realistic Objectives

Despite the massive AI hype, turning the potential of AI into reality is not a given. Estimating the potential value and demonstrating the real outcome of AI in customer service remain elusive challenges.

According to the 2023 Gartner AI in the Enterprise Survey, 49% of leaders highly involved in AI report that estimating and demonstrating the value of the initiative is a key challenge to their AI implementation.³

This difficulty in estimating and demonstrating value often leads to misalignment within organizations. Without clear metrics and agreed-upon targets, conflicting priorities can emerge, causing inefficiencies as people and resources work toward divergent objectives.

When organizations fail to align their goals, they risk missing out on AI’s potential advantages. For instance, if a CEO focuses solely on cost savings while CSS leaders aim to enhance customer experience through resource reallocation, the lack of coordination can lead to suboptimal outcomes. The potential improvements in customer experience might even surpass the financial benefits of mere cost savings.

To address these challenges, CSS leaders should first build consensus among business stakeholders on shared objectives, such as cost reduction or enhanced customer experience, along with clear target values and metrics before project kickoff. Engaging leadership in scenario planning exercises can also foster a shared understanding of AI’s strategic importance. If differing visions persist, calculating the project’s ROI can help clarify benefits and align on a final business objective.

The following research provides guidance on predicting and evaluating the value of AI projects in customer service:

Get Better at ROI Projections for GenAI in Customer Service
Calculating the ROI on GenAI Business Model Innovation

Identify and Communicate the Appropriate Metrics to the Relevant People

After the scenario planning exercise, it’s essential to conduct more specific and targeted workshops for leadership, focusing on AI capabilities, limitations, and ethical considerations. Utilize case studies and real-world examples to highlight AI’s potential and challenges.

These workshops will help clearly define the goals, limits and boundaries of AI, enabling the establishment of metrics for measuring the success of AI projects and their alignment with customer service objectives. By setting specific customer service metrics for AI projects and linking them to strategic goals, CSS leaders can effectively monitor progress and communicate with stakeholders, fostering understanding and support for AI’s impact. Additionally, aligning with KPIs and setting boundaries will help address risks such as ethical concerns, biases and inaccuracies.

Figure 2 illustrates a practical example of how to hierarchically link individual AI metrics with the company’s business goals. It focuses on metrics for evaluating a customer-facing AI use case and its impact on customer loyalty through service and support.

Figure 2: Metrics Hierarchy for Loyalty Growth in Customer Service and Support for AI Products Implementation

The lower level of the metrics hierarchy is composed of the AI product implementation-specific metrics that must be monitored to have the sensibility of the ongoing project. For AI initiatives, they must be focused on five areas:

Adoption
Containment
Effectiveness
Performance
Compliance

Sample Metrics for the Five Areas of AI Initiatives

Measure	Description	Examples
Adoption	Adoption metrics measure how widely and frequently AI technologies are being used within an organization or by its customers.	User engagement rate Usage frequency
Containment	Containment metrics evaluate how well an AI system can handle tasks or issues without needing human intervention.	Failed utterances Escalation rate
Effectiveness	Effectiveness metrics assess how well the AI system achieves its intended outcomes and objectives.	Goal completion rate Accuracy rate
Performance	Performance metrics measure the operational efficiency and speed of the AI system.	Response time Throughput
Compliance	Compliance metrics evaluate how well the AI system adheres to legal, ethical, and organizational standards and regulations.	Regulatory compliance rate Data privacy compliance

Source: Gartner

Schedule Regular Alignment Meetings

During the implementation, form a project steering committee focused on employing AI in customer service that includes leaders from various departments to:

Oversee AI strategy implementation
Ensure cross-functional alignment
Discuss progress
Undertake demos

The committee should comprise active project roles like IT and operations, alongside sponsoring board members such as the CEO, and finance and business leaders. Additionally, stakeholders from legal, HR, and compliance should be included. The committee should leverage and coordinate with any higher-level, corporatewide steering or governance bodies overseeing AI use across the enterprise.

Given that many of the AI technologies and technical skills used may be immature — offering potential improvements during the project but also at risk of overpromising results — these meetings are vital. They allow for real-time adjustments to the initiative’s direction.

The excitement surrounding AI and interest in its potential can be leveraged to secure the leadership board’s commitment to regular meetings to monitor progress (see Figure 3).

Figure 3: First Pillar: Align Leadership

Diagram illustrating the first pillar: Align Leadership. It includes planning AI impacts early, defining KPIs, and forming an AI steering committee. Key takeaways involve engaging leadership in scenario planning, holding workshops on AI capabilities, and forming committees for strategy oversight.

Define the Technology Strategy

Choosing the right technology is neither the first nor the only step. The excitement over AI’s potential can overshadow actual needs, and new advancements often require a careful feasibility assessment rather than a plug-and-play adoption. Leaders must identify the right use cases (what), their roadmap (when), and only then the technological approach (how).

Selecting use cases and placing them in a precise and comprehensive roadmap helps in both choosing the right technology and engaging the right company resources, avoiding treating the project as a stand-alone pilot.

Identify Use Cases

According to the 2024 Gartner AI Mandates for the Enterprise Survey, 20% of AI leaders report that finding the right use case is among their top 3 barriers to AI initiative implementation.⁴

Risks deriving from faulty use case selection take three forms:

Pursuing use cases that promise significant value but face feasibility challenges.
Initiating a use case that is feasible but whose value has been overestimated.
Pursuing feasible and valuable use cases, allocating the necessary time and resources, only to discover later that technological advancements disrupt the path.

To mitigate these risks, the selection of use cases must:

Be aligned with the business objective to determine the target value, as outlined in the first pillar.
Take into account its applicability within the company’s processes and available technology to evaluate feasibility.
Follow a clear roadmap that balances value and feasibility.

Gartner has identified and ranked 20 different use cases for customer service and support, evaluating them based on value and feasibility (see Figure 4).

Figure 4: Artificial Intelligence Use-Case Assessment for Customer Service and Support

For more details, see AI Use-Case Assessment for Customer Service.

Create a Smart Roadmap

CSS leaders should select use cases based on the company’s specific needs and readiness. They must be integrated into a roadmap for AI use case deployment over time, starting from the “likely wins” quadrant. This ensures quick returns and mitigates risks by moving from internal to external use cases.

As the roadmap evolves, the company’s maturity level increases, making the team and the company more confident in the implementation process. Figure 5 illustrates a simplified depiction of a roadmap, where internal use cases are implemented before external ones, as is often the case.

Figure 5: Example of AI Use Cases’ Timeline for Customer Service

Diagram illustrating AI use cases' timeline for customer service. It shows progression from internal use cases like experimentation, case summarization, knowledge article generation, agent assistant, and sentiment analysis to customer-facing use cases such as customer routing, virtual assistant, and correspondence.

Decide Whether to Make or Buy

One important step when selecting the technology, if not yet available in the company, is the decision to build or buy. CSS leaders must evaluate the following factors when approaching this decision with AI initiatives:

AI skills’ availability
Technology costs
The company’s AI ambition
The company’s ability to keep up with technological advancements
Time to market

Gartner recommends that most customer service organizations buy AI technology platforms along with planning for the customization, integration and/or additional build activities, depending on the out-of-the-box capabilities of the vendor platform required to meet their goals.

As technology evolves rapidly, partnering with a platform and vendor that naturally adapts to these changes ensures a more favorable time to value while helping to ensure compliance with security policies and other standards.

Many technology vendors for customer service platforms, including CRM customer engagement center (CEC) and contact center as a service (CCaaS) solutions, have developed or incorporated AI capabilities to support various customer service use cases (for more, see Magic Quadrant for the CRM Customer Engagement Center and Magic Quadrant for Contact Center as a Service).

The advantage of using these incumbent systems is that the data and processes supporting customer service workflows and use cases are already housed within them. Organizations that already own these CCaaS and CRM CEC platforms should test them against their prioritized use cases.

Furthermore, leaders need to identify and incorporate security requirements when choosing AI technology by carefully evaluating potential risks and establishing strong safeguards. This task becomes even more challenging when developing the technology in-house rather than purchasing it (see Figure 6).

For companies deploying an AI system which is used in the EU — whether its use in the EU was intended or not, and regardless of the company’s headquarters location — the EU AI Act will apply (see Why Service Leaders Everywhere Must Address the EU AI Act Now).

Figure 6: Second Pillar: Define the Technology Strategy

This diagram outlines key steps for AI implementation. It includes identifying use cases aligned with business objectives, creating a roadmap starting with "likely wins," and deciding between building or buying AI technology for faster adaptation.

Ensure Data Readiness

— Dr. Kate Crawford, AI Scholar and Research Professor, USC Annenberg in Los Angeles
“Artificial intelligence is only as good as the data it learns from. It is our responsibility to ensure that the data is diverse, unbiased and representative of the world we want to create.”

When it comes to powering AI for customer service, the richest source of data lies within customer service itself.

Assess Data Hygiene and Architecture

In a survey of senior enterprise AI leaders, insufficient data quality was identified as the primary cause of failure for AI initiatives (both GenAI and non-GenAI).⁵ AI initiatives within service and support are no different, therefore, CSS leaders should actively follow these activities:

Ensure data hygiene by establishing processes for data cleansing during implementation and applying them continuously within the CS team, eliminating errors and inconsistencies.
Organize data hierarchically, including the data collected within the team, with clear categories and metadata for easy navigation and AI model training.
Ensure data accessibility and contextualize its quality for each use case. Use the CSS team’s expertise to assess and improve data correctness and comprehensiveness, and implement strong data governance practices.

Create and Enrich the Knowledge Base

Success with AI use cases such as AI assistants, agent advisors, AI chatbots, and AI training within customer service depends upon accurate data. The quality of the information that makes up the knowledge base is key to ensuring not only that the use cases function, but that their operation yields quality results.

This knowledge may be scattered across many repositories. Before launching these use cases, it’s essential to:

Identify where this knowledge resides
Analyze it to identify potential gaps and fill them
Make it available for the AI model

When building this knowledge base, it’s essential to create a comprehensive taxonomy and ontology to organize and structure information, ensuring easy access and interpretation by AI. Modern tools now incorporate emerging AI techniques to automate the process, significantly reducing or even eliminating the manual effort required to build a taxonomy and ontology for certain use cases.

Establish a Knowledge Management Program

Without an effective knowledge management (KM) program, even the best knowledge base can quickly deteriorate. In a recent survey of service and support leaders, only 59% said their organization has a process for revising outdated content.⁶

CSS leaders should collaborate to develop a KM program that is simple, intuitive and structured to ensure its adoption. Also, it should keep the knowledge essential for the success of AI systems up to date (see Figure 7). The program should also:

Define roles for managing the program itself
Establish operating rules
Use collaborative platforms for exchanging opinions on knowledge base elements
Conduct regular assessments

Additionally, it should include the creation of feedback loops, which are essential for refining AI systems, such as using insights from chatbot interactions to improve AI models. See 2024 Strategic Roadmap for Knowledge Management to plan the KM roadmap.

Figure 7: Third Pillar: Ensure Data Readiness

Prepare People for Change

In many projects, change management programs are often overlooked because leaders believe their benefits are not directly measurable. With AI projects, especially in customer service, the effects are always significant and noticeable.

Create an AI-Specific Customer Service Change Management Program

When corporate functional leaders were asked in Gartner’s 2024 Generative AI Planning Survey about their challenges with implementing GenAI in their organization, a third of respondents indicated the resistance from employees who fear role reduction/elimination (34%).⁵

This concern is particularly relevant in customer service roles where AI can automate routine tasks.

To address this issue, it is essential to foster a culture of change. This can be achieved by implementing a change management program, which will also aim to demonstrate how AI can enhance agents’ roles rather than replace them.

Moreover, Gartner’s 2023 AI in the Enterprise Survey³ shows that organizations implementing change management activities regularly tend to see a greater impact on business outcomes of AI projects (see Figure 8).

Figure 8: Impact of AI on Business Outcomes by Frequency of Change Management Activities

As per the 2023 Gartner AI in the Enterprise Survey, 95% of respondents who perform change management activities always/usually experienced revenue growth. Organizations that perform change management activities frequently have a greater impact on business outcomes and gain more trust from BUs.

The following are the most impactful change management activities for an AI initiative in CSS:

Assess the impact of the AI initiative on customer service roles.
Create a communication plan tailored to the CSS team by understanding their tasks and challenges. Clearly explain how AI will enhance their roles, improve efficiency, and boost service quality.
Manage employees’ transition via training and resources.
Identify and secure buy-in from key stakeholders in the CSS department (supervisors and CSS team leaders).
Establish mechanism for CSS agents to provide immediate feedback on the quality of the AI tool and the impact on service quality.

Involve Key Personnel in Planning

CSS leaders should select individuals with leadership qualities who embrace change to participate in AI project implementation. Involving them in decision-making leverages their insights to enhance AI solutions and positions them as change agents, keeping the team informed. Key involvement stages include:

Planning
Vendor selection
Implementation
Piloting
Milestone meetings

User involvement can also foster new specialized roles such as AI trainers, data analysts, knowledge managers and AI system supervisors. While previously limited to a few IT departments, these roles will become critical for customer service and support.

Develop a Training Program

From Gartner 2024 Generative AI Planning Survey, 31% of leaders identified training, development, upskilling and knowledge sharing as challenges in AI implementation.⁵

Meanwhile, as per the Gartner 2023 HR Technology Employee Experience Survey, 87% of employees are interested in building at least one AI-related skill.⁷ Training is both fundamental and in demand, making it crucial to pinpoint the specific AI-related skills needed for the project. These may include basic AI literacy, proficiency with AI tools, data analysis, feedback loop creation and AI ethics. It is essential for CSS leaders to develop targeted training programs to enhance their team’s skills in these areas.

Training emphasizes how delegating lower-level routine tasks to AI can enhance skills. It also highlights the importance of developing new capabilities and focusing on high-touch human skills to improve the customer service experience.

A continuous training program enables the team to provide feedback on potential improvements, ensuring their engagement and involvement as they adopt and utilize the AI system. This strategy can be the key factor that differentiates a successful and widely embraced initiative from a brilliant yet ineffective concept that goes unused (see Figure 9).

Figure 9: Fourth Pillar: Prepare People for Change

Figure showing preparation for change through early change management programs with AI-specific initiatives to address job security and boost adoption. It suggests selecting adaptable leaders to drive change and developing training programs on AI literacy and tool proficiency to enhance customer service experience.

Evidence

¹ 2025 Gartner Business Buyer Survey: This survey sought to understand how functional business units (customer service, finance, human resources, marketing, sales and supply chain management) within organizations approach large-scale software purchases to support their business function. The survey was conducted online from October through December 2024 among 3,068 respondents from organizations with annual revenue of at least $50 million or equivalent from North America (36%), Western Europe (32%), Asia/Pacific (19%) and Southern Europe (13%). Industries surveyed include education providers, energy, financial services, government, healthcare, health payer, technology, telecom, insurance, manufacturing, natural resources, retail, transportation and utilities. Qualified respondents were at manager level or higher, and had been actively involved in the purchasing process for the most impactful software capabilities for their respective functional business units during 2023 or 2024. Software purchases were either new, replacement or expansion purchases. Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.

² 2025 Gartner Maximizing Value from Customer Service Technology Survey: This study was conducted to understand how service and support leaders get value from their technology investments. The research was conducted online during December 2024 and January 2025 among service and support leaders from a wide range of industries. In total, 224 respondents participated from North America (n = 169), Europe (n = 44), the Asia/Pacific (n = 10) and Africa (n = 1). Respondents are all from companies with more than $250 million USD in annual revenue. Disclaimer: Results of this study do not represent global findings or the market as a whole but reflect sentiment of the respondents and companies surveyed.

³ 2023 Gartner AI in the Enterprise Survey: This study was conducted to understand the keys to successful AI implementations and their impact on the broader AI that has been brought by generative AI. The research was conducted online from 19 October through 21 December 2023 among 703 respondents from organizations in the U.S., Germany and the U.K. The main sample consisted of 645 out of the 703. Organizations were required to have developed or intended to deploy at least two AI initiatives within the next three years. Respondents were required to be part of the organization’s corporate leadership or report to corporate leadership roles. Fifty-eight out of 703 are the business intelligence (BI) sample. Organizations were required to have developed or intended to deploy at least one AI initiative within the next three years. Respondents were required to be part of the organization’s corporate leadership or report to corporate leadership roles or below (senior manager and above) and to be primarily responsible for BI in their organizations. Both the main sample and the BI sample respondents were required to have a high level of involvement with at least one AI initiative, and they were required to have one of the following roles when related to AI in their organizations: determine AI business objectives, measure the value derived from AI initiatives, or manage AI initiatives development and implementation. Quotas among the main sample were established for company size and for industries to ensure a good representation across the sample. No quotas were established for the BI sample. Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.

⁴ 2024 Gartner AI Mandates for the Enterprise Survey: This study was conducted to understand how AI and generative AI (GenAI) are being adopted by enterprises, focusing on areas such as AI strategy, data, governance, literacy, engineering, organization, portfolio and value, to assist clients in keeping pace with AI’s rapid evolution. The research was conducted online from October through December 2024 among 432 respondents from the U.S. (n = 181), the U.K. (n = 70), France (n = 50), Germany (n = 50), India (n = 51) and Japan (n = 30). Quotas were established for company sizes and for industries to ensure a good representation across the sample. Organizations were required to have deployed at least one AI use case in production. Respondents were screened for C-level executives (e.g., chief AI officer, chief data officer, chief data scientist, chief digital officer, chief information officer, chief operating officer, chief technology officer or equivalent) or roles above vice presidents. All respondents were required to have high involvement in at least one AI initiative. Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.

⁵ 2024 Gartner Generative AI Planning Survey: This survey was conducted to examine generative AI’s use case implementation and impact by business function. The survey was conducted from September through November 2023. In total, 822 business executives who lead corporate functions outside IT and who indicated will begin or continue to implement Generative AI across the next 12 months qualified and participated. The research was collected via online surveys in English. The sample was equally split across the following eight corporate functions: finance; HR; marketing; sales; customer service; supply chain; procurement; and legal, risk and compliance. The sample mix by location was North America (n = 536), Europe (n = 176) and Asia/Pacific (n = 110). The sample mix by size was $50 million to less than $500 million (n = 119), $500 million to less than $1 billion (n = 129), $1 billion to less than $10 billion (n = 374) and $10 billion or more (n = 200). Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.

⁶ Gartner Customer Service and Support Priorities for 2025 Survey: This survey was conducted online from 11 July through 22 August 2024 to understand the most pressing priorities and key challenges for service and support leaders in the upcoming year. This year, it also measured service and support leaders’ investments in generative AI. In total, 187 people participated. Survey participants were from North America (n = 137), EMEA (n = 34) and Asia/Pacific (n = 11). Disclaimer: The results of this survey do not represent global findings or the market as a whole, but reflect the sentiments of the respondents and companies surveyed.

⁷ 2023 Gartner HR Technology Employee Experience Survey: This survey was conducted to understand employees’ ratings of 75 technologies and innovations across seven HR subfunctions based on the level of adoption in their organizations, the impact on current performance and the future importance for employee performance. The research was conducted online from 10 October through 7 November 2023 among 3,477 respondents from various geographies, industries and functions. The survey was designed and developed by Gartner’s HR Practice research team.