Agentic AI in Solutions Engineering: Worth the Hype?

Let’s face it. If you’ve worked as a solutions engineer (SE), you know how busy you can get. AI agents (which are kind of like ultra-efficient digital assistants) offer the potential to analyze data, make decisions, and even take action on your behalf; all of these use cases promise to reduce repetitive work.

It’s tempting to apply AI agents to as many use cases as you can. 

But should you?

And if you do, when and how could AI agents save you the most time? 

Let’s look at some of the more compelling use cases for agentic AI in SE and sales roles. The TL;DR? We’re nowhere near AI replacing SEs. Your deep knowledge, experiences, and connections to people are the best way to understand prospects’ and customers’ true pain points.

For now, we’ll separate hype from reality and point to some practical, time-saving use cases for AI agents you may not have explored yet. 

Agentic AI vs. Generative AI: What’s the Difference? 

While agentic AI and generative AI are both buzzworthy, there’s a big difference between the two. Generative AI is all about creating content — think text, sample datasets, images, code, videos, and more. It takes your input, or prompt, and relies on pattern-matching to produce something new. 

While generative AI can create content based on learned data, it can’t make decisions or take action on its own. That’s the key difference between generative and agentic AI. 

The Agentic AI described today is built on top of and leverages generative AI — specifically large language models (LLMs) — but only as a part of a greater whole. Without the LLM, AI agents are just workflows, and they’re no better than procedural code. What some people consider agentic AI is just workflow-based task automation. These automated systems typically follow a series of pre-defined steps, and don’t actually involve much decision-making. This is where “agents” have been for a while now, and we are only now seeing real pushes for more capable, independent AI agents.

Unlike generative AI, AI agents don’t rely on human prompts. Instead, they meet certain objectives or goals on their own, such as maximizing sales or efficiency. AI agents can execute on a sequence of complex tasks independently — things like searching a database and then triggering a workflow to finish an activity. 

Most SEs should evaluate using both technologies in their day-to-day, even though generative AI tools are more mature at this point. Let’s take a look at some of the emerging use cases for agentic AI.

Agentic AI: When Should SEs Use (Or Not Use) It?

In general, the SE role is high touch and highly specialized. Humans should still be leading customer-facing interactions, and should check the outputs of AI carefully to avoid unnecessary errors.

When evaluating whether you should use agentic AI, here are some other key considerations to think about: 

• What steps will the AI agent take to help you meet your goal?
• How does the AI agent construct a response?
• Can the agent check the response against the source data it’s using?
• Can the agent make decisions independently? In other words, is it different from a task-based automation or workflow?
• Does the company take into account security (e.g. SOC 2 Type 2) and responsible AI best practices so your data is protected? 

In terms of early use cases for AI agents, below are some of the areas where they’re being tested today. Some may be closer to task-based automation. Today’s agents have varying degrees of capability, so it’s important to think about the goal you’re trying to accomplish and whether an agent can get you there.

Use Case #1: RFPs

Some startups are using AI agents to make the process of gathering information across systems and drafting RFPs much easier. Companies like Arphie, for example, use AI agents for proposal drafting, security questionnaire responses, and due diligence questionnaires (DDQs). AI agents pull in and validate data across a number of sources like Salesforce, Slack, Sharepoint, HighSpot, and Google Drive. 

Use Case #2: Sales CRM Agents

CRMs like Salesforce have introduced AI agents into their platforms to take independent actions and automate certain tasks like providing quotes. While some of the features may be more valuable for sales teams (think automated sales coaching), SEs can provide the inputs and data to make these systems more useful across the entire revenue team. 

Use Case #3: Demos

While there are some agentic AI demo capabilities, most companies will likely not want to take the risk. Having AI agents independently take action on something as customized as demos might be relinquishing too much control for SEs. To contrast, presales teams should investigate more proven use cases for AI in demos, such as using generative AI to create customized demo datasets. Using AI for these types of specialized and repetitive tasks can save hours of time and data engineering resources.

The Bottom Line

SEs should try out technologies like generative AI for proven use cases that will save them the most time. But the majority of agentic AI applications may be too early or untested for such a high-touch role that requires human expertise. Be sure to vet any vendors in this sector, get demos, and see how these applications could be used in your day-to-day — rather than taking lofty promises at their word.

Read more: Demoing AI Software is Different: Why Control is Key