
AI First with Adam and Andy · 2026-06-24 · 13 min
Key moments - from our scoring
Substance score
45 / 100
Five dimensions, 20 points each
Rather than chase efficiency gains, Forum3 built a managed agent that unlocked entirely new workflows for a network of independent sales reps in their forties and fifties who lack AI proficiency. The agent integrates with the client's existing data systems and commerce engine, proactively delivering daily sales priorities, prospect identification, personalized messaging, and strategic coaching via email - channels every user already understands. This is not a chatbot or Slack integration, but a sandboxed, read-only data system that mimics how reps would naturally delegate to a human assistant. Early results show the reps are hitting their sales numbers while spending less time on analytics and CRM management. The key insight: agents become most adoptable when they remove friction from user interaction (email feels natural, not like new software training) and when they generate new business impact rather than automating existing tasks. This pattern is now proliferating across consumer brands, parent management, family scheduling, and beyond.
Deploy the agent through channels they already use daily, like email, so users interact naturally without training or software adoption friction. The agent should mimic how they'd work with a human assistant or colleague, not require new workflows or interface learning.
Instead of making existing processes faster, this agent unlocks entirely new workflows that didn't exist before - daily prospect prioritization, personalized messaging, data analysis - that reps couldn't do themselves due to time constraints and lack of analytical skills. It's a revenue growth play, not an efficiency play.
Use sandboxed, read-only database connections so agents can access all necessary data but cannot accidentally delete or corrupt production systems. Even when agents write data, build guardrails ensuring mistakes cannot hurt the company's core systems.
Independent sales reps lacked their own staff (marketing, business intelligence, CRM management, coaching) and had neither the time nor skills to analyze data, create sales plans, or prioritize prospects themselves - the agent became their virtual staff.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode contains a solid conceptual framework about managed agents unlocking new workflows rather than just optimizing existing ones, but it's undermined by vagueness on specifics. The hosts repeat the same core insight multiple times (agents are easier than software, they enable new capabilities, they should access data safely) without diving deeper into the mechanics, implementation details, or surprising findings. For a 13-minute episode, there's significant repetition and throat-clearing that dilutes insight density.
it's an agent, it's not a software system. they set the agent up with its own email. So it wasn't a chat bot, it wasn't even on Slack or whatever
it didn't take something they were already doing and make them do it faster and cheaper. It actually allowed them to do things they weren't able to do
The core thesis - agents as staff extensions for sales reps - is conceptually fresh, but the execution is derivative. The observation that agents should mimic human interaction patterns and be deployed via familiar channels (email, Slack) is increasingly standard. The idea of applying this to restaurants, brokers, and store managers is positioning, not breakthrough thinking. The episode lacks contrarian claims or first-principles reasoning that would distinguish it from other agent-applications content.
imagine an agent that's plugged into all of the data that sales rep has access to
the beauty of this was it's an agent, it's not a software system
This episode features only the two hosts (Adam Brotman and Andy Sack from Forum3, an AI consulting firm) discussing their own client work. While they are practitioners who have deployed agents, they lack external credibility markers or the perspective of the actual sales reps or business leaders who benefited. The anonymized case study prevents verification of their claims or direct testimony from practitioners. For a B2B operator seeking calibrated advice, this reads more as vendor storytelling than guest expertise.
we have a special episode for you today. This is one of our mini episodes, and what we want to talk about is really our real-world experience
I want to maintain client confidentiality. So I'll be a little bit opaque or whatever or coy about some things
The episode is starved of concrete data. No metrics on time saved, revenue impact, adoption rates, or agent performance are provided. The client and sales reps remain unnamed. Timelines are vague ('30 to 45 days,' 'early results'). The only specific details are behavioral (agents email daily summaries, personalize outreach) but no numbers validate the claimed ROI. The hosts repeatedly promise results will 'have' ROI without showing any actual numbers. This is the weakest dimension of the episode.
deployed recently, like in the last 30 to 45 days, and is already having significant impact
the early results are clearly gonna have an ROI
The conversation is collegial and well-structured, but lacks productive tension. Andy mostly affirms Adam's points ('Yeah, so let me just double click'), adds color commentary, and cues up predetermined talking points (data security, new workflows, ROI). There are no challenging questions, no pushback on vagueness, no probing into failure modes or edge cases. The hosts don't interrogate their own claims - for example, why the sales reps' 40s-50s age is mentioned as proof of non-AI-proficiency, or what actually happened with early adoption. Follow-ups are confirmatory, not investigative.
Yeah, so let me just double click and sort of emphasize a couple of the points that you just made
do you want to talk about just the data security piece? Because I think that was something that you omitted in your story
Computed from the transcript - who did the talking, and the words that came up most.
Most companies approach AI agents as efficiency tools, ways to do the same work faster and cheaper. This episode presents a different model. Forum3 recently deployed a managed agent for a client's independent sales force, and the results challenge that assumption. The agent, delivered entirely through email, acts as a personalized CRM, business analyst, and chief of staff for each sales rep. It prioritizes targets, customizes product recommendations using real data, drafts personalized outreach, and proactively delivers a daily action plan. The sales reps, many of whom are in their 40s and 50s with minimal AI experience, interact with it the way they'd interact with a human assistant. The key insight: this isn't an efficiency play. The workflow didn't exist before the agent. It's a revenue growth tool that gives independent workers the equivalent of a functional team they could never afford to hire, with built-in data security guardrails that protect the business from agent errors.
Transcribed and scored by The B2B Podcast Index.
Adam Brotman I think it's so interesting, Andy, how agents, besides all the things they're gonna be able to do, they're actually incredibly easy to use particularly when they're on email and text and Slack and Telegram and stuff, because everybody knows how to email and message. And everyone knows how to talk to another human. if you can set these agents up to where they're really just mimicking how people already know how to work with other people. They already have an interaction model how to use a human assistant or colleague.
⁓ it's actually easier than using a software tool. It's actually it's not only that, it's easier than using a software program. Andy Sack prompt window. This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI innovation and business.
I'm Andy Sack, and alongside my co-host, Adam Brottman, each episode we bring you candid conversations with business leaders transforming their businesses with AI. No fluff, just real talk, actionable use cases, and insights for you. Greetings everyone. ⁓ we have a special episode for you today.
This is one of our mini episodes, and what we want to talk about is really our ⁓ real-world experience and give a high-level case study without revealing the client's name. But we want to talk about a specific managed agent that that we deployed recently, like in the last 30 to 45 days, and is already having significant impact on the business of the client that we ⁓ deployed it in. Adam is gonna tell us the story. he's the one who's been leading the initiative on our behalf.
And We've just been thrilled both to be able to do it, but really with the early results. So Adam, can you tee up, you know, without revealing the client's name, who the client is and what this business problem that we were hired to address? Adam Brotman Yeah, so I want to maintain client confidentiality. So I'll be a little bit opaque or whatever or coy about some things.
But here's the story. So one of our clients has this really cool business where they have a lot of independent sales reps effectively. And they they basically sell the clients' products, these sales reps do, but ⁓ they're on their own. I they're effectively small business people, right?
I imagine, by the way, in a lot of businesses is the case. like brokers and other things that are out there, like real estate brokers and stuff. And then you think about like even for restaurants and retail, like store managers, which are not independent reps, but they run a business, right? And they are expected to sort of achieve results and run a business.
And yet they don't have like a team of people I'll call it like functional teams. They don't have their own marketing team they don't have their own business intelligence team and supply chain team and all this stuff. And you know, they're just independent salespeople or independent managers, whatever. the cool thing is these independent sales reps have a data system and a commerce engine system that makes sense, that they use for their job.
And this would be true in any of these kind of situations. And And what's fascinating is we've been thinking about agents. We were like, you know, what would be cool is if we could give each one of those independent reps their own staff, basically, the equivalent of the staff through agents. And imagine an agent that's plugged into all of the data that sales rep has access to if they wanted to go in, but they'd have to take the time to go in.
They'd have to then know how to analyze that. And then they'd have to put together their own CRM program, their own sales plan, they'd have to chief of staff themselves in a way that they're probably not qualified to do typically. And so we came up with this idea of like, let's create an agent that effectively unlocks all of that for that sales rep. And so we went to the client, we said, we got this idea, was actually kind of inspired by this idea we have for doing this for.
For restaurants, but we're like, this is a sales organization, but we think it will work for this really well. And it'll help unlock them to be both effective and give them back a lot of time, which helps them be effective. But it's also, you know, a key thing, makes them more effective at their sales goals. and so what we did is we started by just interviewing a couple of the sales reps and actually bringing them in and sort of saying, okay, if you had a staff of people, like what would you have them work on?
And i e what would the unlock be for you? And they worked with us. They said, well, we would do this and I would have it do this, and I wish I could do this, and man, I never have time to do that. And we were able to sort of suss out like where the data was for all those things that they mentioned, and so we created a managed agent and it blew them away.
And it blew us away because we we weren't sure If this would work. And the beauty of it was, and I'm I'm sure you're gonna want to riff about this with me, is the beauty of this was it's an agent, it's not a software system. they set the agent up with its own email. So it wasn't a chat bot, it wasn't even on Slack or whatever, which a lot of people want to do.
everyone knows how to email. ⁓ everyone's comfortable with email, whatever the age is. They may not be comfortable with Slack, but they're definitely comfortable with email. They're young enough, they'll they'll prefer Slack, but the point is they know how to use email.
and what's cool about email and Slack, but in this case email, is that not only does the sales rep know how to use it, it's proactive. So it's interruptive in a good way. It's like it shows up every morning and says, Okay, I've been doing this work yesterday, here's your plan for today. And I went ahead and I checked that, you know, these are the people that we're gonna reach out to today, and I've gone ahead and sort of figured out like what we're gonna send them and both the content as well as personalized the message for them and I've teed it up for you to send.
And so it became like a CRM manager, an analyst, a chief of staff, a coach, right? All of those things all in one and it had access to the data. And it's been cool to see it in action where, you know, we're talking to these sales reps about how's it going. As we want to start to roll this out to all their sales rep and they're like, it's amazing.
Like I'm making my sales numbers, I'm bossing my agent around, it's giving me all these ideas. So, you know, it's just really cool. It's not about efficiency, right? it didn't take something they were already doing and make them do it faster and cheaper.
It actually allowed them to do things they weren't able to do because, how we built the agent. Andy Sack Yeah, so let me just double click and sort of emphasize a couple of the points that you just made. So one is really about is that last point you were just making, which is the unlock. And I think What's amazing is that exactly what you just said is that we have a managed agent working on a workflow that didn't exist before.
So it's not an efficiency play. It's actually a revenue growth play. And it's aimed at making the agents more effective. And just to paint a picture of them being more effective, it's giving each individual agent their individual CRM and these individual Adam Brotman Sorry, each individual sal you say agent, just to be clear, you mean sales rep yeah, just yeah, yeah.
Andy Sack Yeah, sales agent, sales rep. Each individual sales rep in AI agent that's functions as a personalized CRM. and You know, just to paint the picture a little bit more, most of these sales reps are in their forties and fifties. They're not AI proficient people, naturally.
Like they may use AI a little bit, but they're across the country. ⁓ and they're not that proficient with AI. And so suddenly they're now interacting with an AI agent that's helping them prioritize their targets, relying on data to customize which products to sell to whom, and having an immediate impact really by email. And so they don't even know that they're really interacting with an AI system.
They're just responding to email messages and it's having immediate sales benefits. So those pieces make it an unlock for the business immediately, not an efficiency place. and we're just super excited you helped the company come up with the idea. It arose from an idea that we'll talk about in another episode about doing this in the restaurant industry.
But we just thought it was something that we were so excited about and the early results are clearly gonna have an ROI. and lastly, the way in which you implemented it, do you want to talk about just the data security piece? Because I think that was something that you omitted in your story. Adam Brotman Well, I mean when you set an agent up, you wanna make sure they have access to the data, but they can't hurt the data.
and they have access to the data in a way that's you know both sandboxed so that they can't like accidentally delete something and screw up your production database if that's the data they need access to. But also just in general, you know, there's a bunch of safety guardrails around how you need to set these agents up, both in terms of the data it's read only, it's not right. but even if you're gonna allow them to create something. ⁓ do it in such a way where even if they made a mistake, it can't hurt the company, it can't hurt your production system.
So we set these systems up to where, you know, we're allowing the agents to have the power of having all the data they need to be a very effective agent on behalf of the sales rep and do all these things without just, you know, YOLOing it into like, the agents, you know, just got access to the database and in whatever way it wants. Which by the way would have been easier to create the agents if we would have just done that. But I just think it's worth being careful because the agents by definition, literally, have agency and they can make mistakes and you wanna protect the blast radius of what would happen if an agent made a mistake.
And so the other thing I'll point out, just one last point on this topic, is that ⁓ and you kind of mentioned this. I think it's so interesting, Andy, how agents, besides all the things they're gonna be able to do, they're actually incredibly easy to use particularly when they're on email and text and Slack and Telegram and stuff, because everybody knows how to email and message. And everyone knows how to talk to another human. if you can set these agents up to where they're really just mimicking how people already know how to work with other people.
They already have an interaction model how to use a human assistant or colleague. ⁓ it's actually easier than using a software tool. It's actually it's not only that, it's easier than using a software program. I'm a non-technical person.
Andy Sack prompt window. Adam Brotman But particularly when you get like certain demographics of either like experience level, age, whatever, it doesn't matter, where they just are super non-technical. And if that's the case, like not only are they gonna not be AI proficient, you don't wanna like teach them how to use some new SaaS tool either. And we're talking about independent reps and stuff like that, where their job is to go sell, right?
you don't wanna teach them one more thing to have to learn. So if you can just be like, hey, we're giving you this. Assistant that's super powered. It's got access to all the data.
It's actually the equivalent of a small staff. It's gonna be really interesting to see what happens when these agents start to really proliferate. side note, I know we're want end this episode, but I'm seeing this all the time now on X and on Twitter and whatnot and LinkedIn. There's all this like, there's an agent now for Parents, it just came out.
It was going viral this weekend. there's an agent for this, an agent for that. It's starting to happen now. Where the thing that we just said on this episode of like, imagine a family can't afford to hire a family manager, an assistant, or whatever.
Like, get an agent to do that. Plug them into your parent square, plug them into your Instacart, plug them into your calendar and your family calendar, and like. my God, like they can do all this stuff for you and take all this weight off of you. And you don't need to know how to interact with them.
You just got to get them the data access safely and then give them normal communication channels. And so I think we're going to see a lot of this. We we just happen to be on like the cutting edge of it when it comes to, you know, consumer brands and our ability to help out, you know, build these agents. Andy Sack I think we set it, so we'll close, we'll allow me to close, which is I think when looking for an area in your business to identify, don't just look at all the existing workflows to the extent that you're able to see a new workflow and particularly one that impacts sales quickly.
The ROI is readily apparent. And so keep that on the list. And then there's lots of benefits from working with managed agents, which include not having to do any training or changing in the way in which people work. They can work in email and Slack, and there's ways to deploy it.
So we were just really excited to share that with all of you. you'll be hearing more illustrative examples from our work, our agentic work with clients. with that, thank you all for listening to AI First with Adam and Andy. For more resources on how to become AI First, you can visit our website, forum3.
com, download case studies, research briefing, executive summaries, and join our email list. We also invite you to connect with our AI First community, a curated hub and network for leaders, turning AI hype into action. We truly believe you can't overinvest in your AI learning. Onward.