Latest / Sales Pipeline Radio / Seven AI Agent Archetypes for B2B GTM

Transcript

Seven AI Agent Archetypes for B2B GTM · Sales Pipeline Radio

• 0:16
  All right.
• 0:17
  Welcome everybody to another episode of Sales Pipeline Radio.
• 0:19
  I'm your host, Matt Heinz.
• 0:21
  Excited to have y'all here joining us in the middle of your work
• 0:24
  day, middle of your work week.
• 0:26
  Just a week away, as we record this, from a Memorial Day weekend,
• 0:29
  which is just crazy to think about.
• 0:31
  If you're watching us live on YouTube or on LinkedIn, thank you
• 0:35
  so much for being part of the show.
• 0:37
  We would love to make you part of the show.
• 0:39
  If you have a comment, or question on our topic today and our conversation
• 0:42
  with Jessie, would love to get into that.
• 0:45
  If you're watching on demand or listening on demand, thank you so much, for
• 0:50
  continuing to subscribe and follow.
• 0:51
  Over a quarter million people downloaded and have access to Sales Pipeline Radio
• 0:55
  and every episode of this podcast, past, present and future, always
• 0:58
  available at salespipelineradio.com.
• 1:01
  Very excited to have... We missed each other at Summit a couple
• 1:03
  weeks ago, Jessie, but very excited to have with us Jessie Johnson.
• 1:06
  She's Principal Analyst of Demand Marketing at Forrester.
• 1:10
  Jessie, thanks for joining us.
• 1:12
  Yeah.
• 1:12
  Thanks for having me, Matt.
• 1:13
  Glad that we could connect after the event.
• 1:16
  I'm glad we could as well.
• 1:17
  This is your second year, well, on the Sales Pipeline
• 1:20
  Radio tied to Forrester Summit.
• 1:21
  We had you last year, we have you again this year.
• 1:24
  It's been, what, a week and a half, since we all left Phoenix, and, heads still
• 1:30
  spinning on a lot of what we talked about.
• 1:32
  How about for you?
• 1:33
  Yeah, absolutely.
• 1:34
  I... And I don't know that that's going to slow down anytime soon.
• 1:38
  Right.
• 1:38
  It's like we're, we're really just getting started now.
• 1:41
  Nothing is slowing down in marketing right now, and I know you're spending
• 1:44
  a lot of time, sort of watching the AI space, which seems to be
• 1:47
  going the fastest of all right now.
• 1:49
  What was... Just talking to people, I mean, there's so much work that goes
• 1:52
  into, you know, your presentation and the research you're doing sort of in,
• 1:56
  in your sessions, but as you talk to people at the show, what were some
• 2:00
  of the key things that you heard from people that, like, you think were,
• 2:02
  that were particularly interesting?
• 2:05
  Yeah.
• 2:06
  I would say for me it, it was a lot of the optimism around- kind of what's possible.
• 2:13
  So we introduced the idea of go to market singularity.
• 2:17
  So, you know, things are, are falling apart, but, you know, how do we start to
• 2:20
  put them together in this kind of brave new world that is very much AI augmented?
• 2:27
  I... One thing that, just with my coverage area, one thing that I found, quite
• 2:32
  energizing was that a lot of the fears have started to, you know, turn into
• 2:38
  that, "Well, what does this mean for me?" "What is my kind of, you know, advantage
• 2:44
  that I can bring to this AI driven world based on my own experience and expertise?"
• 2:50
  So I think a lot of our attendees, you know, while it's a little bit, scary
• 2:55
  at times when we think about, about the future ahead, it was great to start to
• 3:02
  see that, that optimism and to see them start to find themselves in that future
• 3:07
  and how they can really work alongside AI.
• 3:11
  Well, you wrote about this a couple months ago, right?
• 3:12
  You talked about the fact that this isn't... The future is not humans versus
• 3:16
  AI, and I think a lot of people as they get into it are realizing how- exciting
• 3:22
  it actually is for the profession, how exciting it's gonna be for their careers.
• 3:26
  And so there's that... I see it all the time still, that switch from like, "Oh my
• 3:30
  God, it's coming for me," to, "Oh my gosh, I get to focus on the things I love."
• 3:35
  Yes.
• 3:36
  Yep, precisely that, and, you know, that, that idea of how do we balance that human
• 3:41
  judgment, creativity, ingenuity with how we're using AI within the organization?
• 3:48
  It's, you know, while we do see in the news, you know, companies that are,
• 3:54
  letting go of their staff, and then they'll, say that it's because of AI, I
• 3:58
  think really what's going on is just a, not necessarily a reorganization, but
• 4:03
  a reshaping of how the work gets done, so making sure that we are calibrating
• 4:08
  humans and AI based on the things that are going to matter most to the business.
• 4:14
  So let's double click on this a little bit.
• 4:16
  Late last year you wrote about what you consider the seven AI
• 4:19
  agent archetypes that... and at the time you said that will reshape.
• 4:24
  It is now... It is definitely reshaping it now.
• 4:27
  I thought that was a really smart way of sort of breaking down how to think
• 4:32
  about functionally how AI's helping.
• 4:33
  Love to have you sort of just quickly walk that through that for people.
• 4:36
  Sure, sure, and thanks for the shout-out and for reading the research.
• 4:41
  You know, it's true, things go so quickly.
• 4:43
  I'm anxious to get this updated to, you know, kind of represent
• 4:47
  that, that current state now.
• 4:49
  But when this came together, you know, we were kind of in this, this place where
• 4:54
  everything was being called an agent, even if it was kind of our traditional
• 4:59
  chatbot rule follower type of interface.
• 5:03
  So a lot of the goal there was to figure out, you know, you've already got a lot
• 5:07
  of AI in your stack, and now we have these really cool and powerful capabilities
• 5:13
  from agentic and generative AI.
• 5:16
  So how do we really rationalize that and figure out, you know, what does
• 5:20
  that look like not only in our tech stack, but how we get work done?
• 5:23
  So the seven archetypes, you know, as we think about kind of the base level,
• 5:29
  so I hesitate to call it old school, but you know, digital process automation.
• 5:34
  So we still have this idea of this rule follower agent that's really
• 5:38
  going to do this if, then, then that.
• 5:41
  If this, then that, kind of approach to its work.
• 5:45
  So, you know, routine, reactive, rules-based.
• 5:49
  Some use cases there, you know, answering scripted questions where we don't want
• 5:53
  anything to necessarily go off script.
• 5:56
  We have a very defined response to a certain question.
• 6:00
  Things like, you know, enabling very sequential tactic automation.
• 6:05
  So, even from a marketing perspective, as much as we say that, you know, the,
• 6:09
  the buyer's journey, customer experience, it's nonlinear, we need to meet our
• 6:14
  buyers where they are, that is true, but there are times where we're going
• 6:18
  to want to, you know, kind of distill those tactics in a certain order.
• 6:23
  Above that, so I'm starting at the bottom of this hierarchy.
• 6:27
  Above that we have kind of the producer archetype.
• 6:30
  So this is all about leveraging things like retrieval-augmented generation
• 6:35
  to figure out how can we embed both ideation and co-creation of content
• 6:41
  or whatever it is that you're working on in your interfaces and workflows.
• 6:46
  So this idea of kind of, you know, working side by side with
• 6:50
  AI as you get your work done.
• 6:52
  So a lot of, you know, generating content that's everything from a social
• 6:56
  post to now as things have evolved, being able to, you know, generate
• 7:01
  a custom pitch deck for a client.
• 7:04
  Above that we have our savant.
• 7:07
  So as you think about you know, kind of a specialist with
• 7:10
  deep expertise and mastery.
• 7:12
  So this is the agent that's really learning a lot about your
• 7:16
  business, about your customers.
• 7:18
  They have that domain expertise wherever that may sit, whether it's, you know,
• 7:22
  listening in on those intent signals and figuring out what to do with them or, you
• 7:27
  know, how do we respond when we're getting this type of feedback from a community.
• 7:32
  So these are, you know, agents that are able to conduct deep research
• 7:36
  for us to make the connections that we need to have made.
• 7:40
  A good example there as we think about buying signals, understanding
• 7:44
  not only what is that signal of interest, but who else might be
• 7:47
  involved, in that purchase process.
• 7:51
  Above that, we have our influencer agents.
• 7:53
  So this one, I got a lot of love for this particular category.
• 7:58
  We've seen these agents, you know, in the past within
• 8:01
  conversation automation solutions.
• 8:04
  So, a big, kind of differentiator there as we, run up the ladder here,
• 8:09
  the influencer agent is actually ready to have external conversations.
• 8:14
  This is the agent that knows how to talk to your buyers and customers, what to
• 8:18
  say, when to bring in that human into the loop, and not always, you know,
• 8:23
  to prevent failure, but to deliver additional value at that point of handoff.
• 8:30
  And then as we, you know, start to look a little bit ahead, we have the
• 8:35
  choreographer agents, so kind of a middle manager that's going to be
• 8:39
  coordinating the work of, you know, your, say, junior level agents, and, you
• 8:45
  know, really serving as that go-between between the human and the team or the
• 8:50
  crew of agents, that we're working with.
• 8:53
  Looking ahead again, you know, the planner and the innovator, the top
• 8:57
  two archetypes, were really considered long-term when this report was written.
• 9:02
  I think we're getting even closer to that planner archetype.
• 9:06
  So, you know, kind of an agent that can, you know, really
• 9:10
  practically develop and enact.
• 9:13
  So I think that's kind of the key there to not just developing that plan, but
• 9:16
  actually being able to action that plan, likely through that choreographer
• 9:21
  that's going to be, you know, kind of reigning in multiple agents with
• 9:25
  different areas of domain expertise.
• 9:28
  So they can put together, you know, your account-based engagement plan,
• 9:32
  and then also action that plan with that human, approval and oversight.
• 9:38
  And then the innovator bucket, I think this one is going to be, you
• 9:41
  know, really interesting to watch.
• 9:43
  So these are, these are the agents that can actually come up with
• 9:47
  new ways of doing things, novel approaches, maybe even invigorating,
• 9:52
  you know, established demand products.
• 9:55
  So this is gonna be a really interesting, kind of agent
• 9:59
  archetype to watch in the future.
• 10:00
  They can help us to create tools, create tools on their own, maybe even create,
• 10:06
  you know, a team of agents to execute on that account-based engagement plan that
• 10:11
  that planner archetype came up with.
• 10:14
  So yeah, the future is certainly bright, as we think about how we
• 10:19
  can augment our go-to-market with these various agent archetypes
• 10:24
  Yeah.
• 10:25
  I really like the way you've broken those down, and if you're a Forrester
• 10:28
  client, definitely go make sure you download that and get that to your team.
• 10:31
  If you are not, just go listen to that again and write it down.
• 10:36
  'Cause I think it's a really good way of thinking about, sort of structurally
• 10:40
  how to sort of get the jobs to be done that are possible with AI are
• 10:43
  really rolling the organization.
• 10:44
  Talking here on Sales Pipeline Radio today with Jessie Johnson.
• 10:48
  She's Principal Analyst of Demand Marketing at Forrester.
• 10:52
  I mean, I think so many people are starting to rely more and more on
• 10:56
  these models, on agentic workflows.
• 10:58
  We're still, though, in sort of a trust but verify mode.
• 11:01
  Key example, just as we got started, before we got started here today- ... I
• 11:05
  mean, I've been using AI for everything, and including sort of show prep here.
• 11:09
  To say, like, what are some things that I don't know?
• 11:10
  What are some things I should know?
• 11:12
  Claude told me you were retiring this year.
• 11:14
  That this was your 10th and final summit.
• 11:18
  Neither of us can figure out where it came up with that.
• 11:22
  I'm glad... I'm glad I caught that before I sort of, like, leaned into it.
• 11:27
  But, like, it's sort of like I've seen a lot of people say, "Well, like, I
• 11:31
  can't really trust this. I can't just let this set it and forget it." How
• 11:35
  do you tell, to help teams balance sort of being proactive at, and sort
• 11:40
  of aggressive even at using AI, but knowing that it's still not perfect?
• 11:45
  Yeah.
• 11:45
  It's the... A lot of that comes down to, you know, again, what is
• 11:50
  the human bringing to the table?
• 11:51
  And so very much, a lot of that comes down to that human experience
• 11:55
  and expertise that's going to, you know, help that, seems so weird to
• 12:01
  say human in the mix, but will really help your frontline marketer to know
• 12:06
  when do we question that output of AI?
• 12:10
  How often, if we're, you know, kind of getting into to regular, I don't wanna say
• 12:15
  rinse and repeat work, but as they become more ingrained into those workflows, how
• 12:20
  often do we need to have kind of that check of that AI output until that trust
• 12:26
  is really, really in place and, you know, we're able to observe, here's the sources
• 12:31
  that that agent used for information.
• 12:34
  So we're not just looking at, you know, the output of that agent, but what is the
• 12:38
  source that they found that information.
• 12:40
  So human judgment is, you know, certainly going to be required there.
• 12:45
  And what we're finding in our data too is, you know, where
• 12:48
  does that split happen right now?
• 12:52
  You know, what is AI able to fully automate on its own?
• 12:56
  So tasks that, you know, have a lot of data complexity and
• 13:00
  speed is really what matters.
• 13:02
  But those, you know, kind of tasks, engagement, the interactions, the-
• 13:08
  workflows that really impact the brand, the customer experience,
• 13:13
  or, you know, are a little bit more risk-intensive- ... that's where we
• 13:17
  tend to see that kind of partnership.
• 13:19
  So, and I think that's gonna vary with those individual, you know,
• 13:23
  human employees that are using AI.
• 13:26
  Some of them may be a little bit more trusting of AI out of the gate, for
• 13:30
  better or worse, and others are always going to kind of want to be able to
• 13:35
  collaborate versus just kind of receiving and acting on whatever the AI kicks out.
• 13:41
  Like my retirement plan.
• 13:43
  Like your retirement plan.
• 13:44
  Well, you know, eventually it's gonna be right, just not right right now.
• 13:48
  We've got just a few- ... minutes left with... yeah, we got just
• 13:50
  a few minutes left with Jessie before we need to wrap up.
• 13:52
  Where's the ceiling on this?
• 13:54
  I mean, I lliterally as we record this today, there's- a conference... As we
• 13:58
  record this, there's a conference that has a 45-minute session that says, "We will
• 14:02
  build you an AI VP of Marketing," that it's just gonna do it all for you, right?
• 14:06
  And some people look at that and say, "Oh, crap, maybe I am in trouble."
• 14:09
  Like, where do you... how far does this go, and where do you think the future
• 14:13
  of AI and human interaction is gonna be, especially for go-to-market teams?
• 14:18
  Yeah.
• 14:18
  Yep.
• 14:18
  I think, you know, part of that is a lot of this is still kind of shiny object.
• 14:26
  You know, it's cool to play around.
• 14:28
  It's, you know, to think the art of the possible.
• 14:31
  What we find when we talk to our clients, they're looking for, you know, kind
• 14:36
  of what does this look like as we, we continue to look into the future?
• 14:39
  But also, what does that mean for me today?
• 14:42
  What is pragmatic and practical that I can integrate into my business, my
• 14:47
  workflows, my go-to-market, the way that we think about the world even?
• 14:51
  And you know, that idea of let's build you a VP agent, that's certainly compelling.
• 14:56
  That's absolutely that innovator archetype.
• 14:59
  But I think what gets lost in kind of the rush to build and design is the idea
• 15:05
  that we still need to integrate those agents into our systems, our data, and
• 15:11
  also other agents that are in the mix, and we need to onboard that agent as though
• 15:16
  it... I won't say as though it were a human, but onboard that agent in context
• 15:22
  of the role that it is going to play.
• 15:24
  So having a very human-first approach to onboarding those agents, I think
• 15:29
  that integration and onboarding piece are really what, what brings the value.
• 15:35
  We can create agents all day long, but we need to make sure that they
• 15:38
  can actually deliver that value.
• 15:41
  Yes, we do.
• 15:41
  Hey, before I let you go, so I, we already established that,
• 15:46
  that you are not retiring.
• 15:47
  And so, like, I do have this section where I ask Claude, I say, "Okay," like, "tell
• 15:51
  me things, a couple things that I, that people may not know about this person."
• 15:55
  So this could be two truths and a lie.
• 15:56
  Here's the three things it said.
• 15:57
  First, it says you're retiring.
• 16:00
  Then it said, then it said you have a Master of Fine Arts in Creative Writing.
• 16:05
  Is that true?
• 16:07
  Correct
• 16:08
  Okay, so that one's right, yeah And it also says that you do the Michael
• 16:11
  J. Fox Ride for a Cure every year to raise money for Parkinson's disease.
• 16:17
  That is not correct.
• 16:19
  Yeah, yeah.
• 16:20
  I wish it were.
• 16:21
  That, it sound- that was a really great set of things, so it was two
• 16:25
  lies and a truth, thanks to Claude.
• 16:27
  So,
• 16:27
  Two lies...
• 16:28
  Yeah, so always a, it's a work in progress.
• 16:31
  But it's pretty amazing.
• 16:32
  It is.
• 16:32
  Well, Jessie, thank you for joining us.
• 16:35
  If people wanna read more of your stuff, especially new research and
• 16:38
  content you have coming out, what's the best place for them to go?
• 16:42
  I would say, you know, if you're not a client, follow the Forrester
• 16:45
  blog in addition to what we put out, on LinkedIn as well.
• 16:49
  So yeah.
• 16:51
  Awesome.
• 16:51
  A few different, different places.
• 16:53
  Awesome.
• 16:53
  Jessie Johnson,
• 16:54
  Forrester, thank you again so much for joining us.
• 16:56
  I'm really glad you're not retiring, 'cause we need more content and
• 16:58
  more insight from you on this.
• 16:59
  So, thank you so much.
• 17:02
  Yes, thank you Matt.
• 17:03
  Take care.
• 17:04
  Awesome.
• 17:04
  And thank you everyone for attending and for joining us for another
• 17:07
  episode of Sales Pipeline Radio.
• 17:08
  We'll see you next time.
• 17:09
  See ya.