
The Scale Up Show · 2025-06-22 · 7 min
Key moments - from our scoring
Substance score
21 / 100
Five dimensions, 20 points each
Speaker A demonstrates a three-step workflow for connecting ChatGPT's Deep Research capability to HubSpot CRM to automate sales intelligence and outbound strategy. The method involves crafting a detailed prompt by instructing the AI to act as a top prompt designer, then leveraging OpenAI's recently released connectors to link Deep Research directly to HubSpot data - no internet scraping required. The system analyzes contact trends, identifies patterns across industries and job titles, and generates actionable insights (like C-suite targeting recommendations) that would typically require hiring a Rev Ops analyst at $100K - $250K annually. Speaker A notes this approach is currently unique to HubSpot among mainstream CRMs and shares that fewer than 20% of 800+ revenue executives surveyed use Deep Research agents at all. The workflow culminates in automated email and multi-channel ABM sequences tailored to the exact personas discovered in the analysis. Revenue leaders, sales ops teams, and CRO-level executives managing companies from $15M to $1B+ revenue will find practical value in this integration pattern.
Use OpenAI's connectors feature (available in the Tools menu) to authenticate and link your HubSpot account directly to Deep Research, then select HubSpot as the data source when running your research prompt instead of searching the web.
MCP servers are custom connectors that allow you to create your own integrations beyond the built-in OpenAI connectors; they enable linking CRMs and tools not yet natively supported by OpenAI.
Deep Research typically takes 10 to 15 minutes on average to complete analysis and will examine multiple sources within your connected account before generating the report.
OpenAI connectors support Google Drive, Box, Dropbox, Gmail, and Google Calendar, and you can authenticate additional sources through custom MCP servers.
It generates a key metrics snapshot, identifies industries and lead sources, provides visual analysis, includes detailed commentary, and recommends actions like C-suite targeting elevation and persona enrichment.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode's single substantive insight - that OpenAI released a HubSpot connector enabling Deep Research to query CRM data - is real but thin, and is surrounded by credential-dropping, travel anecdotes, and repeated self-promotion. The three 'steps' are described so vaguely (the actual prompt is never shown in full, the tool UI walkthrough replaces any operational depth) that a listener is left with almost no replicable process.
OpenAI recently released connectors a few weeks ago and not a lot of people are talking about it
you would essentially need to hire uh, an analyst or a Rev Ops person and that could be a hundred thousand to $250,000 to hire this person
The core idea - use AI to analyse your CRM contacts and generate outbound sequences - is one of the most recycled AI-sales takes of the past two years. The only marginally fresh element is the specific OpenAI connectors release, but that is presented as a feature announcement, not as original thinking or a contrarian argument.
I need you to act like a top 1% of 1% prompt designer who works at OpenAI
I created a prompt which or wanted to act in the style and voice of Jason Bay who does a really good job at Outbound
This is a solo monologue by a content creator who self-reports large audience numbers and conference appearances but demonstrates no verifiable operating track record at scale. The credentialing is entirely self-asserted and the content reads as a tutorial-influencer format rather than a practitioner sharing hard-won expertise.
I just conducted a, uh, training with a cohort of over 800 different revenue executives ranging from companies at IBM, Amazon, all the way down to solopreneurs
I met with over 1500 executives at the Gartner event, as well as two other ones with the CMO summit and CRO summit
A handful of concrete figures appear (19 sources, 15 minutes, $100K - $250K analyst cost, $15M - $1B+ company range, sub-20% adoption stat) but none are sourced or verifiable, and the core tutorial never shows the actual prompt text, the actual output data, or any real contact-level results - just a screen-share narration of a report the viewer cannot inspect.
it took 15 minutes and looked across 19 sources
companies ranging all the way from 15 million to a billion plus
There is no host-guest dynamic, no interviewer, and no probing questions whatsoever - this is an unstructured solo screen-share monologue. The speaker repeatedly loses his place, apologises for jumping around, and never interrogates a single claim or explores limitations of the approach.
Sorry about jumping around here
Now this will take 10 to 15 minutes on average. So I'm not going to do that right now.
Computed from the transcript - who did the talking, and the words that came up most.
Your competitors are already using AI. Don't get left behind. Weekly strategies used by PE Backed and Publicly Traded Companies → - In this video, I reveal how to enhance CRM systems using the deep research capabilities embedded in AI models like OpenAI. I share a step-by-step guide on creating effective prompts, connecting CRM systems such as HubSpot, and transforming insights into actionable strategies for revenue growth. I emphasize the importance of leveraging AI for data analysis and highlight the potential it holds to significantly improve business outcomes. Chapters 00:00 Unlocking CRM Potential with Deep Research 02:50 Step-by-Step Guide to Implementing Deep Research 06:12 Transforming Insights into Actionable Strategies
Transcribed and scored by The B2B Podcast Index.
Speaker A: I've discovered a way to make your CRM 100 times smarter and it's really simple to actuate. I'm going to show you in three steps how you could do it in less than 10 minutes. Right now, here and today. Thing that shocked me is I just conducted a, uh, training with a cohort of over 800 different revenue executives ranging from companies at IBM, Amazon, all the way down to solopreneurs. The thing that really, really shocked me is that less than 20% of them is use Deep research, which is effectively a research agent that's embedded across Google, Gemini, ChatGPT. There's even a capability in Claude to do that Perplexity Grok that are within all the models. However, what I'm going to show you today in three simple steps is how to actuate that and connect your deep research agent with your CRM to get insane insights. So let's just get rolling on it. I want to ship this fast for you. One of the things that I've been doing and I haven't been able to make as many videos these lately is I've been traveling a lot. I've been going to multiple conferences. I met with over 1500 executives at the Gartner event, as well as two other ones with the CMO summit and CRO summit. So I'm going to get right into this. All right, first of all, step one is I want to show you how you decrease a prompt. Okay, so Deep research, specifically within OpenAI is one of those capabilities where you really need to prompt it good to get amazing results. And so it's very simple. And what you're going to see here is I need you to act like a top 1% of 1% prompt designer who works at OpenAI. And I specifically call out that we need to create a prompt for deep research. And there's two core components to this. Number one is what is the task that I want it to do here? As you can see, I wanted to review my contacts in HubSpot through deep research, analyze trends, patterns and opportunities. Okay. For the last 90 days and then I'm going to identify my outcome. That outcome is reverse engineer who I'm attracting from different industries, titles and companies. Then I say please include no more than one to three customizations. Okay? Now what happens when you do this and you leverage the model to create prompts in that specific way is it comes up with a really hyper detailed example. All right? Now I'm not saying I couldn't create this. There's cases where, if you understand the process at a very deep level, how to manually do this, then that's going to be the best way. However, in this situation specifically, there are a lot of very granular components within HubSpot in terms of how to source data out of there. And the models know do a really good job of understanding that. So that's step one. All right, next thing we're going to do is go to step two. I'm going to go into the contact analysis piece and for purposes of this, actually I think I already executed that. So here we go. This is what I'm looking for. Contact trend analysis. Sorry about jumping around here. So what you're going to see here is this is pretty simple, but it shows me executing the prompt and then it builds out the report. Okay, but before we do that, I want to show and I'm going to create a fake chat so you can see what it looks like within here. All right, so if you go in the chat, you click on tools, the first thing you'd be able to do to run deep research is to click run deep research in the tools. Now what you're going to see here is it has different sources, it could look over the web, which is the most common use of deep research of folks that have used it. However, OpenAI recently released connectors a few weeks ago and not a lot of people are talking about it. You can connect your Google Drive, HubSpot Box, Dropbox, and you can even connect more. All right, I should probably make a whole separate video on this, but as you can see, I already connected HubSpot. If I wanted to connect something, uh, different, let's say I wanted to connect my email account. All I would have to do is go on here and then continue to Google and authenticate. All right, now the benefit of doing this is, and let's go back, let's say, for example, I wanted to connect to Gmail and Google Calendar. I could do a time study analysis on, um, what I'm spending my time on and identifying if it aligns with my goals. That's an example for right now. We're just going to focus on HubSpot. Okay. The other thing I'll tell you as well is you even have the ability to create your own connectors. All right, and so as you can see over here, this is through MCP servers and, and it's explicitly designed and mapped out, um, what you could see and how to access that. So we'll include those in the resources for the video. However, going back, I'm going to show you how to execute this, right, we have HubSpot connected. Then when I clicked on the sources, all I have to do is basically put the source only as HubSpot. Right. So that way it's not surfing the Internet. Uh, if I want to basically enrich it or cross reference other data then I would want to click both of these areas. All right, all right, so that's step two and then I would obviously paste the prompt in here. Now this will take 10 to 15 minutes on average. So I'm not going to do that right now. So you're not just sitting there staring and watching what happens. However, I am going to show you the results of what happened with this trend analysis. And so there's a lot that's in here. What you're going to see is I drop the prompt in and then it's going to say hey can you please confirm should uh, I use the connected HubSpot account? So sometimes it's going to ask that. And then, and this is in beta research, it took 15 minutes and looked across 19 sources. Now what you're going to see here is it goes through and gives me a key metric snapshot. It identifies the industries lead sources, visual analysis of what it found and it even goes on deep dive of commentary. So all this is like very hyper valuable and you would essentially need to hire uh, an analyst or a Rev Ops person and that could be a hundred thousand to $250,000 to hire this person to do all these data and reports. And now you can have an agent that'll do it. It's even recommending, hey Ryan, here's some recommended actions, right? Elevate your C suite targeting, do Persona enrichment and then go with industry specific playbooks. Now what you're going to do is like all right, great Ryan, it's got all the data in here and it's, it gives me the insights. Now the bonus on what I didn't tell you I'm going to show you as well is from there what I did is I created a prompt which or wanted to act in the style and voice of Jason Bay who does a really good job at Outbound. Give me a five step email sequence. All um, right, so what you're going to see here is now it starts to create email sequences specific for the exact type of title, Persona and role that are found across those contacts. Okay, now I want to take this a step further and say okay, just create a multi channel sequence to get more contacts and turn them into clients and customize them for me. So Then what you're going to see is it's going to create that multi step sequence. It's got the different days, emails and exact process to conduct a multichannel ABM strategy. Okay, so what you're going to see is like literally that I showed you how to do that all in five minutes and I know I went fast, you go back and pause it. But long story short, this is going to basically transform what's possible because the biggest barrier I've heard with companies ranging all the way from 15 million to a billion plus is that these AI models are not directly connected to your CRM system. This is going to break down that barrier. Right now though, the only CRM that's available unless you create a custom MCP server that I'm seeing is HubSpot. They were one of the earlier adopters. I assume Salesforce is going to come out too, but wanted to create this because it's a super simple, quick tutorial on how to walk through it step by step, use deep research, connect it with your CRM and get a hundred x the value and then instantly turn that into action. So thanks for joining me today. If you want more details and research on these areas, I drop examples like this week by week. Feel free to join and you'll get a free resource from me. But also stay alert of this and be the first to see these videos as they come out. Anyways, thanks for joining me today. I'll see you on the next video.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.