5 key trends shaping the contact centre in 2025.

In 2025, artificial intelligence (AI) will expand its influence in the contact centre space as this prolific technology trend permeates more facets of daily operations, dominating capital expenditures and strategic decision-making.

By Martin Cross, Chief Strategy Officer at Connect.

1. Rollback in GenAI implementations

GenAI offers elegant and natural speech capabilities, which is why many operators rushed to implement solutions.

However, constraints in how GenAI models deal with customers quickly emerged, with some operators experiencing challenges around accuracy, factuality and relevance in the responses provided.

GenAI models trained on open large language models (LLMs) from the internet are not trained for a specific purpose, which means they can sometimes generate generic outcomes, provide incorrect or nonsensical information, and go off track when performing assigned tasks.

Some GenAI engines have also responded in unfair or discriminatory ways to customers based on biases inherent in the data sets used to train the engine, leading to reputational damage to the brand and lost business. The rising incidents of these hallucinations are prompting other businesses that embraced GenAI to roll back their implementations and rethink their approach, with a focus on creating first-for-purpose LLMs because businesses are realising that selling travel insurance is different from supporting retail sales.

2. A move to safe and autonomous AI

The mixed experiences and outcomes from GenAI implementations, combined with more legacy IVR technologies reaching end-of-life will see an increase in the adoption of autonomous rules-based AI and safe AI, especially as AI is set to tackle more frontline engagement in 2025.

Research and strategic advisory firm Metrigy predicts that up to 65.7% of inquiries will be resolved by AI in 2025 and that contact centres without AI will need to invest in 2.3 times more agents.

By implementing guardrails and other safeguards around GenAI, developers can make it more useful in the contact centre to unleash AI-enabled speech agents that can act independently like real people.

Where autonomous rules-based AI, also known as agentic AI, primarily focuses on autonomy and decision-making based on rules, safe AI emphasises safety and risk mitigation, regardless of the underlying technology.

Companies will choose which is most relevant based on the required tasks, as autonomous rules-based AI can make decisions and take actions based on their own understanding of the situation but is generally limited in its ability to handle complex and unforeseen situations, whereas safe AI incorporates techniques like reinforcement learning and human oversight to ensure robustness and adaptability.

By developing and deploying more controlled autonomous AI, rather than relying on open LLMs to train engines, contact centres can create conversational AI capabilities that feel natural and suit particular markets without the risk of hallucinations.

3. AI support for the hybrid workplace

With the hybrid workplace now entrenched as an accepted contact centre model thanks to trends such as remote working, outsourcing and globally distributed workforces, operators will increasingly look to AI-enabled workforce engagement management (WEM) to enhance workforce productivity among work-from-home (WFH) and work-from-office (WFO) agents.

Contact centres that need to respond to periods of high demand or manage large agent numbers, especially across geographies and time zones, need a system that can forecast engagement volumes and determine agent demand and the necessary skills at specific times of the day.

AI can analyse these requirements across an operation and all digital channels to determine the most efficient workforce in terms of cost and customer satisfaction.

For example, AI engines can listen to every recording, analyse every engagement and better manage the workforce by making real-time recommendations on resource allocation.

Moreover, as larger contact centre operators migrate to the cloud, AI-powered WEM solutions will become indispensable in managing workflows..

4. Security and compliance

WFH models, geographically distributed workforces and high agent turnover rates are among the leading factors increasing regulatory complexity and security risks for contact centre operators. 

As agents access different systems and view valuable data and sensitive information, they are high-priority targets of fraudsters and bad actors.

AI-enabled workforce management (WFM) toolsets can help boost security. AI algorithms can analyse vast amounts of data in real time to identify unusual patterns, such as sudden spikes in call times, high error rates, or suspicious login activity. This enables early detection of potential fraud or security breaches.

Advanced WFM solutions can also confirm agent credentials to ensure only authorised persons can log on, or mask credit card details from agents to protect this sensitive financial information and reduce the risk of fraud in an elegant way without making these processes more restrictive for agents and negatively impact the customer experience (CX).

5. Convergence of AI and quality management

As the need to train safe and autonomous AI engines increases, operators will need to turn to internal data sets to provide the insights needed to continually improve AI-enabled engagements.

Quality management (QM) solutions or knowledge management systems will increasingly emerge as valuable data sources to feed AI knowledge engines, enhancing their ability to deliver better engagements and outcomes.

As the quality of conversations improves through AI-powered agent assist solutions, the AI engine can use these quality engagements to update the knowledge base, analysing content in real-time to continually improve and find the best information to answer customer questions. The AI engine can feed these answers to live agents or operators can leverage this data to deliver better frontline digital engagement.