Demand Cost and Value Gains from your service providers with AI-powered predictive and autonomous SLAs

Business leaders, your service providers are gaining bottom line and value for AI. Are they sharing these with you?

A recent HFS study showed us that organizations use AI-ML (artificial intelligence and machine learning) to reduce operational costs (see Exhibit 1). Your service provider partners are reducing their costs with AI-ML, too. Are they passing the savings along to you? Maybe, and maybe not. In this POV, we’ll help you uncover and realize these bottom-line impacting benefits of AI-ML.

Exhibit 1: Operational cost reduction is a major driver for AI adoption

Sample: n=153 

Source: HFS Research, “State of Machine Learning, 2018” 

Your service providers’ operating margins are improving thanks to AI; this is the right time for you to ask all of your service partners to showcase and transfer to you the non-linear gains in service quality and cost efficiency that AI has enabled.   

You might be a bit perplexed, though, about what to ask for: 

• How much of the value and cost efficiency gains have you received from your service partner and provider organizations?  
• How much have the quality of service (QoS) and service level (SL) metrics changed?  
• How much of the operational cost benefits have providers transferred to you?  

To answer these questions, it’s most practical that you start with your service level agreements (SLA), which should cover the critical service value levers that monitor and govern your relationships with your providers. The SLA is the first place where your service providers and partners should display and share with you cost, efficiency, and value gains.  

Take action to ensure higher value and lower costs from your service provider relationships  

Demand cost-efficient, value-enhanced SLAs—AI can make them predictive, real-time, and autonomous

Raise the bar for your service provider partners in these three aspects: 

• Predictive and exponentially faster SLA metrics, thanks to autonomous resolvers; 
• Quality of service in terms of reduced manual errors and variance, improved consistency, and service guarantees in availability and reliability; 
• Service costs, showing how efficiently and effectively your providers can leverage AI in terms of scale and coverage of scope of work.  

HFS has crafted three action items to help you achieve an optimum SLA.

Demand predictive, real-time, and autonomous SLAs instead of traditional, reactive, post-facto, incident resolution-focused SLAs  

To achieve this SLA composition:

• Redefine your service level metrics. Retire post-facto metrics like mean time to respond, mean time to resolve or repair, and mean time between failures. Replace them with predictive and autonomous SL parameters.  

• Challenge your internal and external service providers to showcase how they can leverage the intelligent autonomous and predictive capabilities of AI.  

With abundant number of AI use cases available in IT and business operations, not only do the typical response, resolution, fulfillment, and process TATs (turnaround times) become non-linearly faster and near real-time, e.g., from hours to seconds, they also become more predictive—and proactively manageable—wherever precedent patterns are available and learned by machines from historical data and knowledge bases.  

This example shows how proactive, predictive, and autonomous SLA metrics could play out in a ticket resolution scenario. Traditional SLAs have five nines (99.999%) of 2-to-4–hour resolution windows for P1 tickets, 8-to-12–hour resolution for P2 tickets, and so on. When IA agents are assigned service requests or tickets, along with a contextually intelligent and dynamic orchestrator, they can instantly scale out basis work volumes, demand spikes, and infrastructure availability. As a result, they can resolve tickets in near real-time with only seconds or minutes of latency.  

Where the incidents have precedence and resolutions already exist in the autonomous agents’ knowledgebase or patternbase, IA agents can accurately predict issues with fewer false positives and can sometimes resolve incidents even before they happen in a zero-incident service delivery model.  

This scenario would enable these proactive, predictive, prescriptive, and autonomous SLA metrics: 

• The number of P1 and P2 incidents autonomous agents proactively predicted and resolved before they occurred. 
• The percentage of P1 and P2 incidents that already had preventive, proactive resolutions available for autonomous agents and that were resolved in a zero-touch manner, i.e., without any manual intervention. 
• Additional metrics such as mean time to predict, seconds to respond by autonomous agents, mean time to proactively resolve incidents, and mean time to prevent incidents. 

Demand a new service outcome guarantee for consistency, low variance, high availability, and reliability in service quality and CSat 

A unique benefit of AI is that it is consistent. AI-enabled, near-zero-touch service delivery doesn’t suffer from typical human behavioral fluctuations, fatigue, or productivity loss. If an autonomous agent can resolve a certain type of P1 incident in less than 10 seconds, it can continue to do so until the incident pattern changes and requires retraining. This improves the consistency and predictability of outcomes, the quality of service (QoS) parameters for availability and reliability, and, in turn, customer and user experience—ultimately resulting in better CSat and NPS (Net Promoter Score).   

Demand new cost of service models 

The cost of service delivery should decrease significantly. You must demand a partial transfer of those cost benefits back to you in addition to any gains in service quality, efficiency, and experience.  

For example, consider the following scenario: 

• Thirty percent to 50% of service delivery activities happen through an AI-enabled autonomous route. 
• People-costs are at least 40% to 50% of any contact or resolution costs. 
• Zero-touch doesn’t mean that you won’t have any service–delivery teams; your providers will still need to have some human engineers, for example, to train, maintain, and control the autonomous agents. So, team sizes will shrink by about 50% to 60% and extra team members will be reskilled and shifted to other roles with higher demand.  

For this enterprise, these three service delivery characteristics combined could lead to a cost reduction of at least 6% to 12% on existing scope of work of service contracts. 

In general, through effective AI and automation solutions deployments in service operations processes like tickets and requests classification, resolutions, recommendations, service restoration, request fulfillment, and provisioning, enterprises could realistically achieve efficiency gains of 5% to 15%. 

Bottom line: Your service providers must pass the value and cost efficiency gains they are getting from AI to you. 

As end-user enterprise AI leaders and teams, you must demand outcome-based, proactive, and predictive SLAs with near-zero incident frameworks and significant cost efficiencies while negotiating on services contracts.