Is Business Process Outsourcing the first target for Service as Software?
Steven Forth is CEO of Ibbaka. Connect on LinkedIn
Service as Software is emerging as a central pattern for the monetization of business AI. Generators are foundational, co-pilots dominated the approaches that many companies brought to market in 2024 and agents are attracting a lot of attention this year. Agents and co-pilots can be woven together to deliver Service as Software (sometimes known as Service as a Software in a play on SaaS). In the long term, there are big opportunities to reconfigure many core business processes with generative AI and deliver them as a service.
Generators power Agents and Co-Pilots and can be woven into Service as Software solutions
What services will be first to fall into this new paradigm?
This question came up in our January 23 webinar “AI Monetization in 2025: Unlocking Revenue Streams for the Future” with Mark Stiving and Michael Mansard.
Watch the On-Demand Webinar Now
Download the AI Monetization in 2025 report here
One possibility discussed in the webinar is Business Process Outsourcing of BPO. Companies are already used to blackboxing and outsourcing these services, have good performance metrics, and are cost-driven. Established pricing models can be migrated to Service as Software.
What is Business Process Outsourcing?
Business Process Outsourcing (BPO) is a business practice where an organization contracts with an external service provider to perform one or more essential business functions or tasks. These outsourced processes can include both back-office operations and front-office functions. BPO allows companies to focus on their core competencies while potentially reducing costs, improving efficiency, and gaining access to specialized expertise and advanced technologies.
There are many, many examples.
Front-Office Processes
Customer Service
Call center services
Technical support
Chat and email support
Customer relationship management
Sales and Marketing
Lead generation
Telemarketing
Content creation
Social media management
Back-Office Processes
Human Resources
Recruitment and hiring
Payroll processing
Benefits administration
Employee training
Finance and Accounting
Bookkeeping
Tax preparation and filing
Financial reporting
Accounts payable/receivable
Information Technology
Software development
IT infrastructure management
Help desk support
Cybersecurity
Data Management
Data entry
Data processing
Document management
Supply Chain and Operations
Supply Chain Management
Inventory management
Logistics and shipping
Procurement
Specialized Services
Contract review
Legal research
Patent applications
Healthcare
Medical billing and coding
Claims processing
Appointment scheduling
Travel and Hospitality
Reservation management
Customer service for airlines and hotels
Some of the main business process vendors are listed below. Among these are companies such as IBM that already have significant AI capabilities. The Service as Software transformation may actually benefit these companies.
Accenture: A leading professional services company offering a wide range of BPO services including business strategy, technology consulting, and digital marketing.
IBM: Provides BPO services focused on creating intelligent and customized workflows, with offerings in procurement, HR, finance transformation, and customer care.
Cognizant; A US multinational known for innovative consulting and IT solutions, specializing in AI, digital strategy, IoT, and intelligent process automation.
Wipro: An Indian multinational offering comprehensive BPO services with a focus on digital transformation, cloud infrastructure, and customer experience management.
Genpact: A global services company providing AI-based digital solutions for enterprises, with expertise in insurance claims processing, supply chain management, and credit risk management.
Concentrix: Specializes in customer engagement and business performance.
Teleperformance: Focuses on customer journey management and digital automation.
ADP: Leader in human capital management solutions, particularly in payroll and HR outsourcing.
EXL Service: Provides operations management and analytics services.
Sutherland Global Services: Offers integrated BPO solutions for front-office and back-office processes.
Why is it a good candidate for Service as Software AI?
The hard work of defining the business process, performance criteria, and measurement has already been done. Most of these processes are routine and undifferentiated, which is what made them suitable for outsourcing to begin with. This is the sort of work that many people think AI is best suited for - automating the routine and well-understood. (We have a different approach at Ibbaka though, where we have found that by using Generators, providing a lot of context, and writing well-orchestrated systems of prompts the AI can create highly differentiated creative work.)
At the same time, these processes are relatively complex to execute and are not the sort of thing that a single agent would be able to do. An AI approach to BPO will require flocks of agents working together often under some human supervision, with the humans being coached and supported by Co-Pilots.
How would AI-based BPO solutions get priced?
At first, AI BPO will be priced in the same way as conventional BPO with two important exceptions.
Currently, the main approaches to BPO pricing are as follows
Two pricing models that will disappear as they do not make sense when a service is provided by AI
Hourly-Based Model: This straightforward model charges a flat rate per hour of work, regardless of the number of tasks completed1. It's commonly used in call centers and customer support services, where providers are paid for the time their agents spend working, rather than the number of calls or tickets handled.
Full-Time Equivalent (FTE) Based Model: This mature pricing model considers the quality of output, number of agents involved, and time invested in the project4. It's often used for more complex, long-term projects.
Pricing models that may persist (for a while)
For all of these approaches, tokens are likely to be one of the main pricing metrics as they are one of the main cost factors. Transactions, time and materials, cost plus, and consumption will all be measured in tokens, at least in part.
Transaction-Based Model: In this model, clients pay per transaction or unit of work completed, such as data entries or customer interactions. It's ideal for processes with clear, measurable outputs and stable service-level requirements.
Fixed Price Model: Clients pay a predetermined amount for a specific project or set of services. This model is popular for small, short-term projects and is often preferred by those new to outsourcing.
Time and Materials (T&M) Model: This flexible model charges based on the actual time spent and materials used for a project. It's suitable for large-scale, long-term projects where requirements may change over time.
Cost-Plus Pricing Model: The service provider charges for their costs plus a fixed percentage of profit. This model allows for adjustments based on expenses and can include various fee structures such as fixed fees, incentives, or performance-based awards.
Consumption-Based Pricing Model: Popular among cloud service providers, this model charges clients based on the actual resources consumed, such as storage space or processing power.
The future pricing models
Jobs-to-be-Done Pricing: Pricing expert Gary Bailey has proposed that Clayton Christensen’s Jobs -to-be-Done framework can be applied to AI agents. Each agent has a clear job, the job has a value, and that value can be used to set prices. The same approach can be taken to pricing AI-enabled BPO services. There may be more jobs filled by AI, but the same principles apply.
Outcomes-Based Pricing: For business processes with a clear outcome the final state will be outcome-based pricing. Four outcomes based pricing to work three conditions must be satisfied:
The outcome can be clearly defined (this is satisfied for most BPO services)
Attribution is clear (when the whole business process is outsourced attribution is also clear)
Consumption is predictable for both buyer and seller (current BPO contracts already do a good job of covering this)
AI-enabled BPO, also known as Service as Software, may be the first place we see the adoption of outcome based pricing on a massive scale.
If you are offering a professional service of any kind, begin by inviting AI in (as Ethan Mollick says in his excellent book Co-Intelligence: Living and Working with AI). Then see where a Co-Pilot helping a human provides the most value and then where an AI agent can be built. Combine those agents together, under the supervision of a human with their co-pilot, and build a Service as Software solution. For pricing, look at how conventional pricing approaches can be mapped to your SaS, but also consider Jobs-to-be-Done pricing and outcomes-based pricing.
