Build, Buy or Blend? Deploying AI in Your Organization

Implement the best AI strategy for your specific needs — and know that some combination of sources will likely make sense.

Deploying AI requires acknowledging AI will come from everywhere

The idea that AI is solely a responsibility of data science teams and specialized tech groups is a dated one. The evolution of both technology and culture means AI will continue to come from all parts of the business and in a variety of formats. In addition to more traditional AI inputs, business units are integrating their own best-of-breed AI solutions, creating a decentralized AI trend within the organization. 

The most effective AI for today’s organizations will be a combination of existing applications with added AI features, net-new AI-packaged software and enterprise-crafted AI. In this new world, the role of IT and other AI leaders becomes creating a system to safely evolve, coordinate and run all of these AI technologies.

See Gartner research in action at our CIO conferences and events.

Download Guide: 10 Questions to Ask About Your AI Build vs. Buy Decisions

Start with these essential questions to guide your decisions on acquiring and deploying key AI capabilities.

By clicking the "Continue" button, you are agreeing to the Gartner Terms of Use and Privacy Policy.

Contact Information

All fields are required.

    Back

    By clicking the "Continue" button, you are agreeing to the Gartner Terms of Use and Privacy Policy.

    Company/Organization Information

    All fields are required.

    Optional

    Back

    By clicking the "Submit" button, you are agreeing to the Gartner Terms of Use and Privacy Policy.

    Create an effective AI framework for deployment

    When GenAI became popular in 2023, it shifted the typical AI solution from ML-based, built-from-scratch and trained on curated, centralized data to a more complex combination of AI sources: embedded AI in software, bring-your-own-AI (BYOAI) and built and blended AI. 

    AI source 1: Embedded AI in software

    Gartner expects that by 2026, 80% of independent software vendors will embed AI into their applications. In fact, embedded AI is the largest and fastest-growing segment of AI capabilities, according to the 2024 Gartner AI Survey. These AI features will come in the form of upgrades and add-ons to existing applications, including software solutions like ERPs, CRMs and case management tools.

    Review your current application portfolio, and consider how add-on AI might soon impact each application.

    AI source 2: Bring-your-own AI (BYOAI)

    With today’s range of AI options, individual business departments want specialized AI solutions for their specific business needs. Unlike embedded AI, which is added to existing enterprise applications, BYOAI is composed of new, independent, best-of-breed AI. For example, the marketing team may want to implement GenAI software for content creation or legal might ask for AI software that helps write contracts.

    While individual BYOAI solutions aren’t a problem, the cumulative effect of a suite of individual AIs can create challenges and conflicts with existing embedded software. This can result in AI overlap, unnecessary costs and technical debt.

    AI source 3: Built and blended AI

    Enterprise-crafted AI can be either built or blended — both are centrally owned AI capabilities that fall under the purview of in-house software engineering and data and analytics teams.

    Built AI refers to what data science teams build and train from scratch. However, with the advent of GenAI and its massive foundation models — 100B parameters is considered a small foundation model — teams have shifted to a different approach. 

    “Blended” AI refers to how organizations are now “blending in” APIs from foundation models with custom front-ends, integrations and whatever customization is needed to make the models functional for the organization.

    Build AI governance and oversight into the framework

    For all three AI sources, consider how AI governance will factor into safe management. IT and AI leaders are wise to build a trust, risk and security management (TRiSM) layer into the organization. Enterprises looking to scale out 10 or fewer AI initiatives should have:

    • A responsible AI and ethics team to ensure AI safety.

    • A central AI committee to manage demand (who wants AI and where is it coming from?).

    • Communities of practice to share knowledge and resources.

    For organizations looking to scale a greater volume of AI projects, human governance won’t be fast or reliable enough. “TRiSM” technologies, which serve as “guardian agents” to prevent AI from accessing sensitive data, check outputs, or filter out noncompliant or ethically sensitive data, will be needed to mechanize AI policies.

    Build or buy AI FAQs

    What is an AI technology framework?

    An AI technology framework harnesses data and AI coming from everywhere, and incorporates trust, risk and security management (TRiSM) practices. More advanced enterprises should consider using TRiSM technologies to mechanize AI policies.

    How is GenAI shifting AI strategies?

    GenAI is reforging the IT landscape, evolving beyond the centralized approaches built by data scientists into a more complex environment that funnels AI from multiple sources into the organization.

    Drive stronger performance on your mission-critical priorities.