How a Two-Person Sales Team Punched Above Its Weight With AI Research

At GradRight, our sales team was me and the COO. That's it. We briefly had a third person but he lasted about two weeks, and we eventually brought on an intern who helped with research toward the end. But for most of the time I was there, it was two people running an enterprise B2B sales pipeline — prospecting, discovery calls, demos, negotiations, proposals — selling to university deans and enrollment executives.

Each prospect required serious research before we could have a credible conversation. We needed to understand their university, their specific school or department, their enrollment history, the demographics of their student body, which programs were thriving and which ones needed help. We needed to know what made them attractive to international students — their location, their reputation, their unique selling points — because we were trying to match them with students from India on our platform. If we didn't understand the university well enough, we couldn't assess fit, and we definitely couldn't walk into a meeting sounding like we knew what we were talking about.

This kind of deep prospect research is standard in B2B sales. What's not standard is doing it with two people across a pipeline of dozens of prospects.

The Old Way Would Have Been Impossible

I genuinely don't know how we would have handled our pipeline without AI. I'm not being dramatic — I mean it practically. The volume of research required for each prospect, multiplied by the number of active deals, would have taken a dedicated research analyst or two. We didn't have the headcount or the budget for that.

A lot of the information we needed was publicly available — university websites, published enrollment data, NCES databases, news articles, rankings. But it was scattered across dozens of sources and you had to know where to look. Compiling a useful profile on a single university could easily take hours of manual research.

Perplexity Changed the Math

We started using Perplexity for our prospect research and it transformed what was possible for our team. Perplexity is an AI-powered research tool — you ask it questions and it searches the web, synthesizes information from multiple sources, and gives you an answer with citations you can check.

The key was asking specific, targeted questions rather than broad ones. Not "tell me about University X" but:

• What are the most popular graduate programs at University X's School of Business?
• What is the international student enrollment trend at University X over the past five years?
• What percentage of University X's graduate student body is international?
• What is University X known for academically? What are their strongest departments?
• What is the cost of living in the area around University X for international students?

Each question came back with a sourced answer in seconds. We could check the citations — and we did, because accuracy mattered. We were going to use this information in sales conversations with senior university administrators. Getting something wrong would destroy credibility instantly.

Trust but Verify

Perplexity was good, but it wasn't perfect. We ran into occasional hallucinations — a statistic that didn't quite match the source, a program listed that had been discontinued, enrollment numbers that were slightly off. This was early enough that we didn't fully trust any AI output without checking.

Our process was: get the initial research from Perplexity, then spot-check the most important claims. Sometimes we'd have another LLM cross-reference the same sources to see if there were discrepancies. If something didn't check out, we'd go directly to the source. It wasn't a flawless system, but it was a fast one — and the error rate was low enough that it was a net positive by a huge margin.

I still use Perplexity today, and I've found that running Claude models through it gives me the best results for this kind of targeted research. The combination of Claude's reasoning with Perplexity's web search and source citation is strong.

The Quantic Meeting

One example that sticks with me. We were pitching an acclaimed online MBA program. Before the meeting, Perplexity dug up something I never would have found on my own: this school had published independent research over the years about the efficacy of their programs — student satisfaction, job outcomes, that kind of thing. Not just one report, but multiple studies spanning about five years.

The interesting part was that the older reports weren't prominently featured on their website anymore. They weren't hidden, exactly, but the school had newer research with better numbers, so naturally that's what they put forward in their own marketing. The older reports had kind of faded into the background. A human researcher browsing their site might never find them. Perplexity pulled them up.

So we read the earlier report and the later one, and we could see the delta — how their student satisfaction metrics had improved, how their outcomes data had gotten stronger over time. We walked into that meeting and referenced specific numbers from both reports — the earlier one and the more recent one — and pointed out the strong upward trajectory.

You could see it land. They were impressed, maybe a little surprised, that we had gone that deep. It signaled that we were serious, that we had done real due diligence, and that we understood their institution beyond what was on the homepage. That's the kind of thing that changes the tone of a sales conversation — it goes from "let me tell you why you should work with us" to "we already understand you, let's talk about how we can help."

A two-person team pulled that off. With an AI research tool anyone can sign up for.

Looking Bigger Than We Were

That Quantic meeting was one example, but it happened over and over. A two-person sales team walked into meetings sounding like we had a research department backing us up. We could reference specific enrollment trends, speak knowledgeably about their programs, and ask informed questions that showed we'd done our homework. University administrators noticed. It built credibility and trust before we even got to the pitch.

In direct response marketing, there's a concept that research is the foundation of persuasion — you can't write compelling copy or make a convincing argument if you don't deeply understand your audience. The same thing applies to sales. The more you know about your prospect's situation, the more relevant and credible your pitch becomes. AI didn't change that principle. It just made it possible for a tiny team to actually execute on it at scale.

We were scrappy. We were understaffed. But we showed up prepared, and AI was the reason we could. That's not a futuristic vision of what AI might do for small teams someday. That's what it did for us, last year, with tools anyone can sign up for today.