The question of whether to build or to buy is a heavy one to weigh. It’s a dilemma that many business owners and digital leaders face at some point. There is another added level of complexity, the uncharted territory and relatively new area of Natural Language Understanding (NLU) and Conversational AI for Omni channel contact center applications.
There are very clear benefits and drawbacks to each approach and one can weigh in different options over the other depending on the needs and constraints of your business and your positioning on the learning curve of Cloud AI applications. As you decide on the approach, you’ll want to consider the following basic questions, similar to any other digital transformation:
- How small or large is your business? What are your scalability plans? What are your first use cases of a chatbot/ voice bot?
- Do you plan to grow it significantly? Is self-serve part of your growth plan? This will answer your early question about whether the first release looks like a very basic knowledge bot/ FAQ bot or an advanced bot with web fulfillments / orchestrated journeys
- Will your business have the flexibility of structure? Will it evolve over time? If yes, will conversational AI- self serve to make it into this list?
- Does the success of your business potentially hinge on whether it incorporates cutting-edge technology like machine learning/ AI/ NLP-NLU?
- Is time on your side? What are your time constraints for implementing an in-house solution?
- What is your wallet size or are there any constraints?
Keep reading for the 14 criteria to derive a business rationale before designing your conversational AI solution and review the pros, and cons of buying and building.
|#||Key Criteria||Build on a Hyperscaler Platform||Buy an Accelerator/Proprietary product|
|1||Speed and Time to market (STM &TTM)||Relatively fast on cloud native environment with conversational interface services and proper integration architecture planning, right choice of middleware||Accelerator can give you TTM edge, be mindful of integration effort that need to still be planned with existing apps.|
Reusable block model
|Experience Design, any build can be reused and extended at enterprise level with regional nuances.||Each build is tactical and focused on specific business use case at unit level and might not be scalable for future use cases or feature expansion limited to accelerators & its feature roadmap.|
|3||Data use & control||Full control of data model and insights, activated feedback loop feeds back data to build and refine the builds and expand the use case, may have dependency on public cloud app Out of the Box (OOTB) features.||Prepare for surprises/ exit ramps, vendor may have control of data and can reuse your custom data/AI models to sell to other clients & potential competitors or might be you get benefitted from other customers data models.|
|4||Data governance –
Sovereignty, Compliance & Control
|Aligns with Enterprise environment strategy, data security & residency and compliance strategy. Application build may be hosted on your own virtual private cloud.||Vendor typically host the application on their own private public cloud and may not align to your enterprise data privacy/sovereignty.|
|Full control of source code, process map, builds and design that can be expanded to Enterprise/regional level use case leveraging reusable components.||Vendor might have control on propriety source code and might not share with you if you choose an exit ramp.|
|6||Feature Roadmap||Embrace the agility of Hyperscaler’s platform releases.||Wait for limited product version upgrades.|
|No price lock, pay per use, any new
feature/customization can be expanded over the current build.
|Additional charges to enable incremental feature or build sub flow for existing use case.|
|8||Management||Inhouse support and application management, full control of alerting, downtime, and application support. Build your own admin features.||Additional change for support tickets and turnaround time/vendor SLA to get issues fixed and limited admin features, any increment adds cost.|
|9||Pricing/Licenses||Pay as you use pricing, pay for the cloud service that are leveraged to build the application when you use it.||Unit based pricing/License based pricing, fixed price for license whether all functionalities used|
|10||Process alignment||Technology build will be based on your own
business process alignment.
|Technology & process may never be aligned.|
|11||People Enablement||People (business +tech) are engaged from day 1.||People engagement & ownership is low.|
|12||Scalability||Flexibility to scale for global & pivot.||Every build is almost new.|
|13||Learnings learnt||Learnings learnt will be fed into the use case on the platform for optimization, for both incremental innovation and new use case launch.||Learning learnt will feed into proprietary product improvement which is not owned by you and may never be fed back into your enterprise eco-system.|
|Your IP is your own, you have full control on your code.||Locked in vendor solution and architecture which are vendor intellectual property.|
Here’s a checklist summary of the build versus buy discussion.
- The contact center AI stack is not a core part of the business
- You’re on a tight budget and you need a cost-efficient tool
- You’re on a schedule and you’re aiming for fast deployment
- You’re not equipped with the technical knowledge to take on or develop a conversational AI project
- You want to develop a conversational AI capability for omni channel applications
- Your contact center AI is part of critical enterprise applications
- You have a specific set of feature requirements or use cases not available right out of the box
- You can take your time with its development
- You’ve got the right skill set of professionals to take on the project
- You plan to monetize through your technology stack
Arpita Bhowick is the “Digital Change Champion” in her organization’s digital transformation journey, focusing on digital strategy, implementing multi cloud solutions like CRM, AI driven self- serve offers and contact center AI and data analytics, in a fast paced, ambiguous ecosystem. Arpita has experience in pharmaceuticals, telecommunications, healthcare, the public sector, e-Retail, Fin-serv, travel and handling budgets of $20 MM + program portfolios.