AI Agents: The IT Channel Bets on Its Business Model

The conversation is no longer whether artificial intelligence will touch the IT channel business, but how long it will take to reshape it. AI agents—systems capable of executing tasks autonomously within a workflow—are entering areas that until recently were billed by the hour: pre-sales, support, ticket and incident management, monitoring, and administrative tasks. For an SME that depends on a technology partner, this is not a distant debate: it changes who does what, how much it costs, and how the provider's value is measured.

What is really happening in the channel

According to the Computerworld report, MSPs and partners are already using AI to automate repetitive tasks: IT monitoring, ticket management, technical support, and level 1 and 2 operations. The data handled by the sector points in a clear direction: 90% of MSPs consider that AI will be key to their future growth, and some providers estimate reductions of up to 30% in level 1 and level 2 workloads thanks to automation, according to an Omdia study cited in the news.

Alberto Pascual, executive director of Ingram Micro Iberia, sums it up well: agents are already transforming tasks such as information queries, project coordination, and repetitive actions, and they open opportunities in customer service, incident management, commercial proposals, and support. But he warns that we are in an early stage: automating a specific task is not the same as integrating an agent into complete processes with guarantees of security, supervision, and measurable results.

The real risk: billing hours that the machine does better

Marian Aradillas, from Ayesa Digital, focuses on where it hurts: the danger is not automation, but maintaining business models whose profitability depends exclusively on activities that technology executes increasingly better. Her conclusion is that the future does not involve eliminating the human component, but changing its function: less routine operation and more supervision of intelligent systems, exception resolution, and solution design.

Alejandro Soto, from Arrow Enterprise Computing Solutions Spain, adds useful nuances: the most automatable is the repetitive and transactional—administrative tasks, first-level support, reports, monitoring, provisioning, or initial pre-sales phases. The key is to distinguish between automating tasks and replacing the value that the channel provides. Automation does not necessarily reduce the partner's role: it changes where and how they add value.

Where the business opportunities are

Álex Benito, from TD SYNNEX, identifies five pillars where partners can bill with AI agents:

• Infrastructure for AI: accelerated servers, data centers, and AI Factories, with design and deployment alongside manufacturers.
• Implementation services: integration, workflow customization, connection between channels, and management of the agent's lifecycle.
• Cybersecurity with AI: highlighted by partners themselves as one of the solutions with the greatest impact on their business.
• AI-driven automation: internal and customer processes that previously required manual intervention.
• Managed services and consulting: the area where experience and knowledge of the client's business still make the difference.

The backdrop is the debate on disintermediation: if manufacturers and hyperscalers can deliver agents directly, what is left for the channel? The answer that emerges from the news is that value shifts from selling technology to consulting, integration, and managed services. That is, towards what requires understanding the client's context.

What it means for an SME

For a Spanish SME, this movement has three practical consequences. First: first-level support will tend to be resolved with agents, which should translate into shorter response times and more predictable costs. Second: the provider that only bills hours for repetitive tasks will be increasingly less competitive; it is advisable to review the criteria for hiring and demand outcome metrics, not just activity metrics. Third: cybersecurity and systems integration become the area where a partner provides real value, something we have already discussed in analyses such as AI and geopolitics in cybersecurity investment.

There is also a warning that should not be ignored: integrating agents into business processes requires supervision, security, and measurable results. A poorly integrated agent is not just a cost; it is an operational and security risk. Recent experience with AI tools that find flaws before the attacker, as we analyzed in GPT-6 Cyber, shows that AI is as useful for defending as for attacking if it is not controlled.

Our recommendation from ForgeNEX

1. Audit which tasks are automatable. Before hiring agents, identify in your SME which processes are repetitive and transactional: tickets, reports, monitoring, provisioning. There is the quick return.
2. Don't buy agents, buy results. Ask your partner for clear metrics: time reduction, first-line resolution, cost per incident. If they only sell you hours, ask how they plan to compete with automation.
3. Strengthen human supervision. Agents need a person responsible for reviewing exceptions and validating decisions. Without that control, savings become risk.
4. Prioritize security and integration. An agent connected to your systems is one more attack surface. Demand auditing, access control, and contingency plans.
5. Review your support contract. If your provider automates level 1, that saving should be reflected in the service or the price. If not, you are paying for someone else's efficiency.

The traditional channel does not disappear, but it is redefined. The partners who survive will be those who stop selling hours and start selling judgment, integration, and trust. And SMEs that understand it sooner will have a real advantage: lower operating costs and a provider that helps them decide, not just execute.

Source: Computerworld. Analysis and adaptation: ForgeNEX.