Prebuilt industry AI agents: buy or build?
On November 10, 2026, IBM and Microsoft introduced Imagine Enterprise, a platform for building, running and governing AI agents designed for industry-specific processes. It comes with more than 6,000 agents and assets already prepared by IBM Consulting (IBM, 2026).
It is not the only catalog of ready-to-use agents. The question for any company is no longer whether it will use agents, but which ones to buy ready-made and which ones to build to measure.
What IBM and Microsoft announced
| Aspect | What we know |
|---|---|
| What it is | A platform for building, running and governing AI agents in each industry’s processes. It is not just a catalog: it includes IBM Consulting services. |
| What it includes | More than 6,000 agents and other assets from IBM Consulting. |
| First industries | Financial services, insurance and retail. Manufacturing, energy, the public sector and mobility will follow. |
| Technology | From Microsoft: Microsoft Foundry, Copilot Studio, GitHub Copilot, Fabric, Azure and security from Entra, Purview and Defender. From IBM: Process Studio, Context Studio and Advantage Core. |
| Governance and security | Controls, observability and compliance support, with security, identity and AI governance built in. |
| How it is contracted | Through IBM Consulting. IBM can deploy and run it, co-develop it with the client or let the client manage it. |
Sources: IBM, 2026 and Microsoft, 2026.
The announcement does not mention pricing, general availability dates or regions. The first named client, investment firm SEI, is exploring the technology, not running it in production.
Not the only catalog
- Google launched Gemini Enterprise in October 2025, with its own agents for tasks such as deep research and data analysis. It also includes a finder with thousands of third-party agents that can be filtered and deployed (Google Cloud, 2025).
- Salesforce opened AgentExchange in March 2025, a marketplace with more than 200 initial partners and hundreds of ready-made actions and templates for its agents (Salesforce, 2025).
The big vendors are betting on the same idea: that companies choose agents from a catalog instead of building them from scratch.
Buy or build: criteria for deciding
| Criterion | Buy (catalog agent) | Build (custom agent) |
|---|---|---|
| Type of process | Standard in your industry: complaint handling, reconciliations, frequent questions | Sets you apart from competitors: how you sell, quote or serve customers |
| Data | Structured data in well-known systems (ERP, CRM, email) | Scattered in-house knowledge: manuals, contracts, case history |
| Integration | You already use the vendor’s platform | Your systems are in-house or local and the catalog does not connect to them |
| Language and context | The process does not depend on local rules | You need Ecuadorian regulations, local Spanish or your business rules |
| Control | The vendor’s controls are enough | You need to decide where data is stored, what the agent can do and how it is audited |
| Time | You need results in weeks | You can invest a few months in a lasting solution |
| Dependency | You accept being tied to one ecosystem | You want to be able to switch models or vendors |
In practice, the answer is usually mixed: a purchased base for the standard work and an in-house layer for what sets you apart, meaning your data, your rules and your controls.
Questions for any agent vendor
Before signing, ask for written answers to these questions:
- What can the agent do on its own, and what requires human approval?
- Where is my data processed and stored, and is it used to train models?
- How is each agent action logged, and who can review it?
- Is there a spending cap and a way to stop the agent instantly?
- Does it connect to my current systems, or does it require migrating them?
- Can I switch models or take my configurations with me if I leave the platform?
- What is the total cost: licenses, usage, integration and maintenance?
We covered questions 3 and 4 in depth when reviewing the controls any AI agent should have (Wasivia, 2026).
What it means for companies in Ecuador
Imagine Enterprise is contracted through IBM Consulting and is designed for large banking, insurance and retail organizations. For a medium-sized or small company, the shortest path is usually different. It can start with the agents already included in the tools it pays for, such as its office suite or CRM, and add one or two custom agents in the process where the most is at stake.
The decision does not depend on company size, but on the process. A catalog agent can handle a generic task well. The process that makes customers choose you deserves its own solution.
What to do now
- List candidate processes. Mark which ones are standard and which ones set you apart.
- Review what you already pay for. Many platforms include agents you are not using yet.
- Pick one process for a measurable pilot. Define beforehand how you will measure the result: time, cost or errors.
- Ask each vendor the seven questions and compare the answers.
- Build custom only where it adds an advantage, with controls from the start.
At Wasivia, we help you decide what to buy and what to build, and we develop AI agents and multi-agent systems integrated with your systems and with controls built in from the design stage.
References
- IBM Newsroom — IBM and Microsoft Announce a New AI Agent Platform Designed for Industry-Specific Transformation (Nov 10, 2026)
- Microsoft Source — IBM and Microsoft announce a new AI agent platform designed for industry-specific transformation (Nov 10, 2026)
- Google Cloud — Introducing Gemini Enterprise (Oct 9, 2025)
- Salesforce — AgentExchange announcement (Mar 4, 2025)
- Wasivia — Gemini agent for business: the controls you should require from any AI agent
