Agents vs Automations vs Skills

With the rapid pace of AI, I think many people are overwhelmed right now.

New capabilities. New products. New terminology. Every day feels like another announcement that promises to change how we work.

Personally, I start most mornings reading a few newsletters just to stay current. That helps me get the high-level picture, and then I go deeper into areas where I think real value can be unlocked. The hard part is not learning something new. The hard part is figuring out how to actually implement it into my day-to-day work. The pressure is real, and not optional.

Wayne Gretzky said to skate to where the puck is going. That does not mean chasing hype or collecting buzzwords. What we care about are real use cases that help teams perform better and move faster.

Right now, we’re talking to firms that are not just experimenting with AI. Some are questioning their entire operating models. Others are having real conversations about what their businesses will look like in two or three years.

We can all feel the pace of innovation, but execution is the challenge. My head spins sometimes when I start a task, because suddenly there are five different tools that could help. At Shift, we have committed to keeping up while staying focused on execution. The goal is simple. Use these capabilities to help investment firms work smarter, not just sound smarter.

So let’s slow things down and talk about three terms that keep getting mixed together.

Agents. Automations. Skills. They are related, but they are not the same.

~Ryan Erickson, Founding Executive

What is a Skill?

In the AI context, a skill is a specific capability that a model or system can execute on demand.

It is not a full workflow, nor is it autonomous decision-making. A skill is a reusable AI-powered function that reliably performs a defined task.

Skills are the basic building blocks of AI. They take structured or unstructured inputs, apply reasoning or extraction, and return a consistent output that can be reused elsewhere.

In an investment workflow, AI skills might include:

• Summarizing earnings calls into key themes and risks
• Extracting KPIs from filings or research notes
• Classifying sentiment across internal documents
• Converting long-form analysis into a structured research draft
• Querying macro data and formatting it for downstream models

Most modern AI platforms are quietly built around libraries of skills. When you hear about “tools,” “functions,” or “capabilities,” many of those are simply skills packaged behind an interface.

Why does this matter?

Because skills are composable. They can be chained together inside automations or orchestrated by agents. Without strong skills underneath, agents become unreliable, and automations become brittle.

If AI were a research team, skills would be individual analysts with specific expertise. They do one thing well, consistently, and can be called upon whenever that capability is needed.

Automations vs. Agents

Most teams already understand automations. At their core, automations follow a defined path. When something happens, a set of steps is triggered automatically. They connect pieces of work together so that processes move forward without constant manual input. In investment workflows, that might mean pulling key sections from a new SEC filing and sending a summary to a research team, logging activity after a meeting note is uploaded, or refreshing dashboards when new macro data is released.

Automations are structured and predictable. They reduce friction and eliminate repetitive tasks, but they only do what you explicitly tell them to do. The workflow is predefined, and the outcome is tightly controlled. That reliability is exactly why automations have become widely adopted across enterprise tools.

Agents introduce a different operating model. Instead of executing a fixed workflow, an agent works toward a goal and determines the best way to get there. It can choose which steps to take, which skills to apply, and how to adapt based on new information. Rather than telling a system to run a specific sequence, you define the outcome and allow the agent to orchestrate the work dynamically.

In practice, that might look like:

• Monitoring multiple sources for updates on a portfolio company
• Pulling relevant documents as they appear
• Applying analysis skills where needed
• Generating a draft insight or alert for the team

That flexibility is powerful, but it also introduces complexity. In regulated environments like financial services, agents require strong guardrails, clear context, and thoughtful design to ensure they remain reliable and auditable.

If AI were a research team, automations would represent the operating procedures that keep work moving efficiently. Agents would act more like junior portfolio managers, coordinating tasks and deciding how different capabilities come together to achieve an outcome.

Why Should You Care?

Because understanding the difference changes how you evaluate AI tools.

If you treat everything like an agent, you risk overengineering and losing control. If you only focus on automations, you might miss the adaptive capabilities that are becoming possible. And if you ignore skills, you miss the foundation that makes both agents and automations reliable.

In our work with investment firms, we see the strongest results when teams start small:

• Build high-quality skills around real workflows.

• Connect them through thoughtful automations.

• Introduce agents where flexibility and context actually matter.

The Reality Right Now

We can all feel the pace of innovation. The real challenge is execution.

The firms that will win are not the ones experimenting with everything. They are the ones translating new capabilities into practical workflows that save time, improve insight quality, and help teams make better decisions.

That is where we are spending our energy at Shift.

About Shift
Investment research shouldn't be this hard. Shift turns your firm's scattered knowledge into powerful insights with AI built for how you actually work. We're a team of builders and finance experts based in Charlottesville, VA.