AI for B2B marketing without automating trust

Tihon (00:00)
Hi Daniela, it's amazing to have you here on the podcast.

Danijela (00:03)
Hi Tihon. It's my pleasure to be here and share my thoughts with you and your audience, of course.

Tihon (00:09)
Amazing, thank you so much. So I've invited you because you have pretty interesting experience. First, you've been working with the Megadeals who are pioneers in finite market methodology and huge and very complex sales. Then you've shifted to energy tech market specifically, right? working first with software.

Tihon (00:33)
B2B company and now you shifted to a deep tech company, right? So I'm wondering what transition you're going through in your career currently, right now, if you could elaborate on that.

Danijela (00:48)
So the biggest shift I made first going from working on an agency side, which Megadils is, although although it's not a classic marketing or classic ABM agency, it was a much more complex than that. But still we had clients that we helped with consultancy, with the ABM, with sales orchestration, etc. And then I shifted to being in-house marketer first at utility cloud.

Danijela (01:16)
As you mentioned, SAS company leading their marketing and communication department and now working at Exager in the same role, but different industry, still energy. But Exager works with a product that they innovated and developed powerfoil, so solar cells on a very thin material called powerfoil that can be integrated in different kinds of

Danijela (01:40)
Consumer electronics, retail products, I mean the application area is very broad. And what is different is also first shifting from working from the outside with the companies and helping companies like Utility Cloud and Exager, for example, because Megadils worked with back then sustainability companies, companies from the energy space, and then seeing how the things are when you work from the inside.

Danijela (02:09)
It's a bit different, I would say, being a consultant and helping companies and leading the initiative from the inside and working directly with the sales department, with the founders and the rest of the team. So that was I would say the biggest shift in my career so far. And now it's a new challenge going into deep tech, going into marketing, which

Tihon (02:32)
Mm.

Danijela (02:33)
is more of an integrated marketing part.

Danijela (02:36)
than selling a product. So it's not only marketing a product to the end consumer, whether it's a

Tihon (02:43)
Mm-hmm, mm-hmm.

Danijela (02:45)
end consumer in B2C or end consumer in B2B, but now going through partners and going through clients who work with the end consumer, which is integrated marketing in a nutshell.

Tihon (02:58)
Another layer of complexity, right?

Danijela (03:00)
Exactly.

Tihon (03:00)
Nice. Wow, this sounds very interesting and exciting, and I'm pretty sure lots of marketers in the space can be jealous of this situation.

Danijela (03:10)
Jealous or not, it's always a good challenge to experience new things, to work with different products. I think when we are in a challenge, it can be tough. I mean it's always easier to be in a comfort zone and work with more or less the same product or industry or way of working. But every time when I experience something new, I try to see it as a challenge.

Danijela (03:34)
And in that

Tihon (03:34)
Mm-hmm.

Danijela (03:35)
sense it's probably something to be envious of to have opportunity to grow through challenges and

Tihon (03:41)
Right. Right.

Danijela (03:43)
new situations. Because what is the problem generally in business? Now we are a bit deviating from the topic, but it's maybe worth mentioning and also in marketing, you are often put in a box and you work

Tihon (03:54)
Mm-hmm.

Danijela (03:55)
in a certain industry, and then your chances to get a job outside of that industry are very small.

Danijela (04:01)
Because people who

Tihon (04:01)
Uh-huh. Great.

Danijela (04:02)
are hiring are always looking into your track record of where you've been before.

Tihon (04:07)
Mm-hmm. Right, right.

Danijela (04:10)
while I think that type of thinking is very old because what you always

Tihon (04:14)
Mm-hmm.

Danijela (04:14)
should look for today are people who are adaptive, who are eager to change and to jump into new things. Because if you don't have that ability today, it's very difficult to survive

Danijela (04:28)
on the job market, regardless what job you do, and especially in our field, which is so fast changing today.

Tihon (04:36)
Right,

Tihon (04:36)
right, right, right. Love your mindset around challenges. I want yeah, actually you're right. We are deviating a bit from the topic because we decided to have this episode around AI agents and the help AI can provide for specifically

Danijela (04:50)
Mm.

Tihon (04:51)
your position, which we will come

Tihon (04:53)
to but before we go there I would love to ask you about like general patterns you've seen so far in the industry. I mean first of all the mistakes you've noticed founders do in finite markets and we can talk both about B2B software or deep tech, whatever has these finite market dynamics.

Danijela (05:17)
So founders and salespeople that work close to founder and founders are often at the beginning of those startups scale ups salespeople as well because sales at the beginning are founder led sales,

Tihon (05:31)
Absolutely.

Danijela (05:33)
they they often sell too much and market too little in a way that they don't

Danijela (05:40)
under s probably they understand, but they don't act according to the principle that you probably heard the statistics: 90% of the people are not in the buying mode at the moment you speak to them.

Tihon (05:51)
Right, right. Uh-huh.

Danijela (05:54)
So those five percent, let's say, are, and then there is a specific time when they are in the buying mode. So, what do you do all that time when they are not in the buying mode?

Danijela (06:05)
and when you are creating the awareness and selling. so you should exactly do that merge selling and creating the awareness because you should be on top of mind for those five percent when they are in the buying mode. So to

Tihon (06:21)
Absolutely.

Danijela (06:21)
come to the point, what founders and salespeople often do as a mistake is they don't create the awareness

Danijela (06:29)
l let's say those 300 days per year and don't create thought leadership content, they are not active on LinkedIn, they are not on top of mind for those people who are making decisions when they are making the decision. So I think it's a missed opportunity and I understand everyone has a lot to do. So it's

Tihon (06:49)
Mm-hmm.

Danijela (06:49)
very easy that those activities like

Danijela (06:52)
Creating content, writing posts, writing blogs, newsletters, always jump down on the the to-do list, so become somehow misprioritized. but

Tihon (07:02)
Mm-hmm, huh, huh.

Danijela (07:05)
I think that we have a great opportunity today with help of AI,

Tihon (07:09)
Mm-hmm.

Danijela (07:09)
with help of freelancers at the end of the day to maybe fill that gap.

Danijela (07:14)
And what is good with AI, and we'll c talk about it later, is that you can amplify even one person. Let's say you have one one marketer in your team, that that

Tihon (07:23)
Right, right.

Danijela (07:24)
person can feed the today, even the whole team with the material that was not possible before, because AI helps us multiply the effect of our time and efforts.

Tihon (07:38)
Got it. sounds great. I would also love to ask specifically about the energy market, because it it's also has its own bubble and specifics. So energy market decision making, what have you noticed there different compared to other markets? Like how the logic is different, how this

Tihon (08:02)
niche operate if you can, you know,

Danijela (08:04)
Mm.

Tihon (08:05)
paint paint a quick picture it in a couple of minutes.

Danijela (08:09)
I I don't think if I say that long sales cycles are common in this market, no one will be surprised if I say that. I mean it's almost like a well

Tihon (08:20)
Yeah, absolutely.

Danijela (08:21)
known fact, but it is a fact, so that should be

Tihon (08:26)
Uh-huh.

Danijela (08:26)
taking in consideration. Sil sales cycles are long, and that

Tihon (08:30)
Mm-hmm.

Danijela (08:32)
taking that in consideration.

Danijela (08:34)
That implies a completely different logic how you work, how you market, how you sell. But what surprised so that's not something that surprised me, what I'm trying to say.

Tihon (08:44)
Right.

Danijela (08:45)
what surprised me, on the other hand, was that not always the best product win wins the wins a deal.

Tihon (08:54)
Mm. Mm-hmm.

Danijela (08:55)
So

Danijela (08:56)
The logic that is also common in other industries is common here, and that's that most convenient, most familiar product is maybe tend to be chosen instead of the best. Because taking going after the best project means taking the risk. and you probably heard about

Tihon (09:14)
Mm-hmm.

Danijela (09:16)
this saying no one got fired by choosing IBM. Yeah,

Tihon (09:20)
Right, right, right, right.

Danijela (09:22)
it is that logic in a nutshell.

Danijela (09:24)
So

Tihon (09:24)
Risk aversion.

Danijela (09:25)
risk version and energy space is a very regulated space. It's

Tihon (09:31)
Mm-hmm.

Danijela (09:31)
at the end of the day, it's there are niches, people know each other, so trust

Tihon (09:35)
Uh-huh.

Danijela (09:36)
is a very important factor and hence what we were talking about before, how important it is to be build awareness, to be on top of mind, because all that leads to building trust. So

Danijela (09:48)
Trust and familiarity and going safe, going after compliance regulations,

Tihon (09:53)
Mm-hmm.

Danijela (09:54)
then taking maybe product again which is better but new. you don't have proof proof from previous big customers that tested the product. So risk taking is less common than I thought because

Danijela (10:10)
Energy space, especially here in Scandinavia, green energy space is

Tihon (10:14)
Mm-hmm.

Danijela (10:14)
fast moving, is modern, is innovative. So you would expect that people are going

Tihon (10:18)
Right.

Danijela (10:19)
after the newest, best, latest, but it's not often the case. It's still a pretty traditional market, I would say. And that was what

Tihon (10:24)
Right. Mm-hmm. Mm-hmm.

Danijela (10:28)
surprised me the most.

Tihon (10:29)
Nice, cool. Thanks a lot for sharing this. I think it's now a perfect time to go more towards AI, the meaty part. So Daniela, you've mentioned before in in our conversations that you had some interesting experience building AI agents, working with Copilot on some different projects. Could you please tell a bit more?

Tihon (10:53)
from your position like first of all what is the most helpful like use cases for marketers currently what we can create with AI agents and maybe name in couple cases you've personally created and how you did it because I guess for our listeners this would be the most valuable part the use cases what they

Danijela (11:17)
Mm-hmm.

Tihon (11:18)
brought

Tihon (11:18)
As a result and how you did that.

Danijela (11:21)
So I would like to mention two use cases that I've been working the most with. one is

Tihon (11:27)
Mm-hmm.

Danijela (11:29)
I first of all I tested different LLM models. Privately

Tihon (11:33)
Uh-huh.

Danijela (11:33)
I prefer Claude after ChatGPT. ChatGPT was our first baby for all of us, then we all

Tihon (11:40)
Right.

Danijela (11:41)
transitioned to Cloud when we discovered all possibilities that Cloud allows.

Danijela (11:45)
now

Tihon (11:45)
Mm-hmm.

Danijela (11:46)
in my job we work in a Microsoft three six five space where I'm only allowed

Tihon (11:52)
Uh-huh.

Danijela (11:52)
to use copilot for the security reasons. So I found

Tihon (11:55)
Right.

Danijela (11:56)
a way how to use copilot for the the second use case that I'm gonna mention here. But the first one is the one one workflow I built with Claude, which can help and inspire, let's say, founders. We mentioned founders and their

Danijela (12:12)
shortage with the time and energy to put into developing content and amplifying the content and their presence on social media or LinkedIn, let's say. So

Tihon (12:24)
Mm-hmm.

Danijela (12:25)
what I recommend the small teams to do often is to start with meeting transcriptions. And most of the companies and teams today use

Danijela (12:36)
Some kind of meetings transcriptor, either it's a Firefly, Phatom, Teams, no doesn't matter, all of them more or less work in the same principle, and the main output which is transcript is the same for all of them. So start with

Tihon (12:53)
Mm-hmm.

Danijela (12:53)
the meeting transcriptors, transcribe your meetings, and then use those transcriptions.

Danijela (13:00)
As a source for your content creation with AI. So after you collect all your meeting transcriptions, let's say in a folder, you can even automate with tools like Zapier, for example. I think the right pronunciation is Zapier, I'm not sure, but I'm

Tihon (13:18)
Right.

Danijela (13:18)
sure your audience knows. So you can make a connection from tool like that from a folder.

Tihon (13:23)
Mm-hmm.

Danijela (13:24)
To Claude and

Tihon (13:25)
Uh-huh.

Danijela (13:26)
directly send those transcriptions unless you want to pull them to an LLM manually. In Claude, they can create folders and connect folders to Claude system, Claude cohort.

Danijela (13:40)
Which is a better and smarter way to work with. And either again you send transcriptions directly to Claude to certain folder or you put them on a folder on your computer, and then that's your kind of knowledge source for content creation.

Tihon (13:55)
Mm-hmm, mm-hmm.

Danijela (13:56)
When you create that kind of a system, then next step is that you create a proper tone of voice.

Danijela (14:04)
In Claude, so that Claude knows how you speak. I use a prompt, we can share it in a link or whatever you want with your audience. It's an

Tihon (14:11)
Amazing. Mm-hmm.

Danijela (14:14)
interview, 100 questions that you copy-paste in Claude or ChatGPT if you prefer ChatGPT or any other LLM, and then AI asks you those hundred questions in a very intuitive way, I would say. So

Tihon (14:30)
Mm-hmm.

Danijela (14:30)
Those hundred questions are categorized in different segments, but it doesn't mean that my interview flow would be the same as yours because the

Tihon (14:40)
Mm-hmm. Right.

Danijela (14:41)
prompt adapts your answers. So the follow-up question is always quite natural depending on your reply. so after you conduct those hundred questions with Claude, let's say, I also

Danijela (14:56)
Just a little disclaimer, I recommend people using some kind of voice to text tool. My favorite is Whisper, Whisper Flow. That one is the best I tested so far,

Tihon (15:02)
Absolutely. Mm-hmm. Mm-hmm.

Danijela (15:07)
because it will save you tons of time. Otherwise, if you answer to or type your replies, it will

Tihon (15:10)
Right. Right. Right. Right. Mm-hmm.

Danijela (15:13)
take probably three hours. For me, it took me maybe one hour to conduct this interview. The result of the interview will be a voice profile document.

Danijela (15:23)
That you also upload in this folder that we mentioned you should have on your computer, which will be your kind of content creation folder. I can also share a document with your audience later on to see how I to show how I structured these folders and

Tihon (15:39)
Sounds great.

Danijela (15:41)
documentation that helps me create content as a final result. And then

Tihon (15:46)
Mm-hmm.

Danijela (15:47)
after that, you have on the one hand you have knowledge source with your meeting transcriptions. Why meeting transcriptions are important? Because they they are the source of the latest conversations, both with your clients, internal conversations, so there you get the freshest and latest information. And on the other hand, you have this voice profile, so meaning that AI will create the content which sounds like you.

Danijela (16:12)
And pull the right information from your knowledge source. In that way, with a simple prompt, create me 10 LinkedIn posts that are referring to the latest 10 meetings, client meetings, let's say,

Tihon (16:25)
Mm-hmm. Mm-hmm.

Danijela (16:27)
you will get pretty good material, I would say. 80%, my

Tihon (16:30)
Yeah.

Danijela (16:31)
experience is like 80% ready to

Danijela (16:33)
ready to post. Maybe we'll do some small adjustments at the end, which

Tihon (16:34)
Mm-hmm. Mm-hmm.

Danijela (16:37)
is quite normal. I never I would never outsource and automate the whole flow, meaning from

Tihon (16:44)
Right.

Danijela (16:44)
the meeting transcription through AI to LinkedIn. So I think

Tihon (16:48)
Yeah.

Danijela (16:49)
before it's published it should always go through some human check, because

Tihon (16:53)
Yeah.

Danijela (16:54)
otherwise you don't have control of what is written, how it's written and you minimize

Danijela (17:00)
This generic output that we are all allergic to nowadays. So that's one inspiration or way how I recommend people to create content on scale for small teams, for teams where, like me, one person is responsible for many marketing touch points. Another way how I use AI and in this case, agent is

Danijela (17:25)
Through co-pilot here at the company where I work at at the moment. We have a lot of material that was created before I came into the picture.

Tihon (17:34)
Mm-hmm.

Danijela (17:35)
So old marketing teams created a lot of marketing material that was stored on our SharePoint database. And also we have a lot of policies that we work around and need to apply and be aware of when we create any kind of material. So now I talk about

Danijela (17:53)
hundreds of documents I need to take in consideration and compare my current material to in terms of quality, in terms of regulations, compliance, etc. If you work in a company where you handle a lot of documents, then it's very good to to build specialized agents for each of these fields. So what I've done in Copilot, I built agents

Danijela (18:20)
six agents. One is investor communication agent, board communication agent, governance, taking in consideration all these policy documents, agent that is referring to old campaigns, media intelligence agent that is referring to old media coverage, where we were mentioned in different kinds of publications, and also I'm planning to expand this to media intelligence agent that will monitor the

Danijela (18:47)
company mention in different different publications or online in general, but also monitor the competitors. And then I on the top of that pyramid, I build one chief agent, which is talking to all these specialist agents when I'm giving it, for example, a very high-level task. Otherwise I can go to specialized agents for very specific questions. So in this way I don't need to

Danijela (19:13)
Sit down and think, okay, what did we say in the previous campaign with this client and where can I find the document? Because those agents go directly after I ask such question question, like literally

Tihon (19:25)
Mm-hmm, mm-hmm.

Danijela (19:26)
this question that I mentioned here. the agents can find the document, quote the relevant parts from the documents, and also give me a link to the document. So

Tihon (19:37)
Right.

Danijela (19:37)
this is something I've just built. I'm still finding the right use cases, but I see enormous potential in this infrastructure that helps again companies handling a lot of data to quickly find the right answer and do a lot of comparative analysis between those documents and data, which

Tihon (19:56)
Mm-hmm. Mm-hmm.

Danijela (19:58)
would take plenty of time. And you would maybe even

Danijela (20:01)
I don't know, hire a consultant doing that kind of analysis in in the old ways. Now agents can do that very quickly and cheap once you have them in place. But we are we are planning to take

Tihon (20:12)
Absolutely, absolutely.

Danijela (20:14)
it even higher, like on the higher company level, not only marketing where you would have

Tihon (20:19)
Mm-hmm.

Danijela (20:19)
agents checking in maybe even in real time what is what is correct, what is not correct, if we are compliant, if we are not compliant, etcetera.

Danijela (20:27)
So I think

Tihon (20:28)
Mm-hmm, mm-hmm.

Danijela (20:28)
it's a very very in interesting and exciting future ahead of us.

Tihon (20:33)
yeah, that's true. So yeah, couple notions here. So first of all, about the first workflow, content creation for founders, executives. What I love there is that it's plugged to actual intelligence from the calls, because where exact where the main part happens is usually on exactly calls. And here we should be careful with so you know, like private.

Tihon (20:59)
privacy data and policies, but this is very easy to handle inside L LM and I'm sure everyone who is doing that will be doing that, not you know, flashing names of clients they shouldn't be mentioning and stuff like that, obviously. But yeah, I think talking about like degree of choice, the founder or executive has to create content. This is one of the sweet sports in terms of value and

Danijela (21:23)
Yeah.

Tihon (21:24)
Quality and price what's important. It it costs not so much. You build it inside, you own it, and then you can implement. So it's it's all super super good compared to agency, hiring ghostwriter, stuff like that, much more expensive. Might be better results, right? Might be better outcome in terms of wow, super cool. Also, like when founder is personally writing, do the actual writing.

Tihon (21:51)
and do it with a coach, it can be like it it can be just like amazing, uncomparable in terms of quality, but it requires so much time and so much money, energy and so on so forth. So that's why this category of choice right now is is a super good one. that's my thought.

Danijela (22:07)
Exactly.

Danijela (22:08)
just to comment on that, that's often the reason why people don't post, because they are after quad they are pixel perfect they're after pixel perfect content. And when you

Tihon (22:17)
yeah. I've seen that a lot.

Danijela (22:19)
can't achieve that, then you you don't compromise with what most of people recommend.

Tihon (22:24)
Yeah.

Danijela (22:25)
Eighty percent is good enough. and

Tihon (22:28)
Yeah.

Danijela (22:28)
today's landscape, digital landscape don't allow you to chase pixel perfect because it slows you down.

Tihon (22:36)
Mm-hmm.

Danijela (22:36)
and you probably s you you've probably seen these latest reports on latest trends in social media, like Instagram, which is one of the social media that is most most polished. Even there today it's favor favorable content which is the the the least edited. you know,

Tihon (22:56)
Uh-huh.

Danijela (22:56)
those reels that were very trendy for a while.

Danijela (22:59)
quick editing, a lot of elements popping up on the screen. Now it's more like conversations as we are having now. Pretty raw,

Tihon (23:05)
Yeah yeah. Yeah yeah yeah.

Danijela (23:08)
unedited, you know, genuine. So I think people got tired of overly polished material and offering

Tihon (23:12)
Mm. Right. Right. Five cents.

Danijela (23:17)
genuine organic and even little mistakes. I listened to your previous podcast where you were discussing

Tihon (23:23)
Yeah.

Danijela (23:24)
people making mistakes like a grammar no

Danijela (23:26)
grammar or spelling mistakes, probably more spelling mistakes by purpose to sound more like a human. Because we humans we are not perfect. and that's a

Tihon (23:30)
Yeah, yeah, yeah, yeah. Yeah, yeah, yeah, absolutely. We are making mistakes.

Danijela (23:38)
exactly, exactly. And one more comment

Tihon (23:40)
And we love them sometimes.

Danijela (23:43)
makes us more human. And it's

Tihon (23:45)
Mm.

Danijela (23:45)
you become more relatable as a person, I would say.

Danijela (23:49)
So just a just a quick comment on what you mentioned about security in terms of the first workload workflow that I described.

Tihon (23:54)
Yeah, yeah.

Danijela (23:56)
that's a comment and question I hear a lot. And I've been thinking about it too, because yes, it might be a problem. now I work in a company that is very aware of security, so I think always I have always second thoughts. if this secure enough, I don't wanna leak the information out that is

Danijela (24:15)
Not supposed to be public. So I heard tips recently. There are apparently the tools, or simply you can prompt it in your AI. create a workflow where you write in a prompt each time I mention company X, write this instead.

Tihon (24:34)
Mm-hmm.

Danijela (24:35)
So exchange all company names. Let's say company is called

Tihon (24:38)
Mm-hmm. Right, right,

Tihon (24:39)
right, right.

Danijela (24:40)
whatever, puncture. Every time we say puncture exactly every s every

Tihon (24:41)
Mm-hmm. Mm-hmm. Right. A N D A. Uh-huh.

Danijela (24:46)
for every time you see the word puncture, exchange it to this. So you never actually even if you would put a very confidential text with this kind of w word exchange, you would not actually

Danijela (25:00)
send the signal out there that you're talking about to this specific company. So this c this is a very simple way to simple

Tihon (25:05)
Mm-hmm. Yeah.

Danijela (25:07)
workaround to to make it make it more secure.

Tihon (25:08)
100% 100%. Yeah.

Tihon (25:10)
Yeah, yeah. As far as you are aware about that, it's like super fixable with one prompt. 100%

Danijela (25:16)
Yes.

Tihon (25:17)
on the same page here. About the second use case you've mentioned, creating this database library and agents helping navigate that. I think also super super useful case.

Tihon (25:29)
in you know like loan cycle deals we should have proper memory memory of all details, all touch points ideally. so if we have this in place helps a lot because a a lot of decision fatigue and keeping this in in in one place in one's mind it's it's like it's unhuman even so humans shouldn't be you know like

Danijela (25:52)
Indeed.

Tihon (25:53)
managing that.

Tihon (25:54)
I I find this very useful case. Everyone should do this. And talking about things, Daniela, you would not recommend to do with the eye. What what is hard? No, what you would never do, for example.

Danijela (26:06)
Hmm.

Danijela (26:07)
So I mentioned one of those things earlier. So never publish anything before it goes through a human check. Whoever needs

Tihon (26:16)
Mm-hmm.

Danijela (26:17)
to if you need to check the style, so it sounds like you, if you need to check the data, if it's really accurate data, or AI, of course we all know AI can hallucinate, so no one guarantee you 100% accura accuracy.

Danijela (26:32)
so those things need to of course be checked and always always give a little touch of you in your content. so the workflow I I

Tihon (26:42)
Mm-hmm, mm-hmm.

Danijela (26:42)
described gives you eighty percent, probably eighty-ninety percent good enough content. But

Tihon (26:47)
Mm-hmm. Mm-hmm.

Danijela (26:49)
I'm sure that you will find a way to make it more yours. So put

Tihon (26:53)
Mm-hmm.

Danijela (26:53)
those five percent of your energy.

Danijela (26:56)
instead of ninety five percent writing it from scratch. So instead put your five percent of energy into it to make it more sounding as you. what I've heard recently is that in the future and even now we are moving from being writers to editors. So you

Tihon (27:12)
Mm-hmm.

Danijela (27:12)
don't need to do even create different kind of content, video, photo so much from scratch now. You you are more like a quality controller.

Danijela (27:22)
what was what were editors before. So I guess we are all becoming editors today. And this editor

Tihon (27:24)
Mm-hmm. Mm-hmm.

Danijela (27:28)
part keep, never outsource it to AI and trust it 100%. So that's one thing. Another thing is all activities that involve building trust and communication. So I know that many

Tihon (27:41)
Mm-hmm.

Danijela (27:42)
people are using different kind of agents and workflows to send direct messages.

Danijela (27:48)
on any kind of social platforms or even emails and outreach

Tihon (27:50)
Mm-hmm. Outreach. Mm-hmm.

Danijela (27:53)
is a big no-no when it comes to AI automation, I would say. Maybe

Tihon (27:59)
Mm-hmm.

Danijela (27:59)
maybe if you work in infinite market and you have something really interesting to share. But now we talk about specifically about about finite market where you have to nurture and be careful and really take care of

Danijela (28:16)
each relationship you are building because you don't want to be burn bridges with these twenty, fifty potential customers you have in the market. What you need to do instead and how you smartly need to to use AI in this context is to make it more personalized by doing

Tihon (28:33)
Mm-hmm.

Danijela (28:34)
research with help of AI, by monitoring their activity online or their whatever kind of

Danijela (28:41)
That can help you understand your customer, potential customer better. But never in the outreach, never for

Tihon (28:48)
Mm-hmm.

Danijela (28:49)
spamming them. Be very careful how much and to to to show up with reason. You know, not I think it's an international expression, jump out of the fridge. Is it or

Danijela (29:03)
Is it from my local from my local

Tihon (29:04)
Ha ha ha.

Danijela (29:06)
language, but meaning that you show up everywhere

Danijela (29:09)
just for the sake of showing up?

Danijela (29:11)
And AI can allow you s such s s such frequency.

Tihon (29:11)
Uh-huh, right, right, right. I love jump out of the fridge. Uhhuh.

Danijela (29:16)
Yeah, yeah, yeah. AI can allow you such frequency today, but I think you should not misuse it for that purpose.

Tihon (29:24)
Yeah, yeah, a hundred percent. I think outreach is especially dangerous as soon as as you automate

Danijela (29:29)
Mm.

Tihon (29:30)
and try to make a funnel in relationship based market, like heavily dependent on relations, and finite market is the ultimate, right? It's completely like fully relationship based. Just leads to a disaster.

Danijela (29:44)
Absolutely. So there I see the biggest biggest possibility to ruin the relationships. We all have

Tihon (29:51)
Mm-hmm.

Danijela (29:51)
experienced those automated messages. On LinkedIn I get them all the time. I

Tihon (29:54)
Mm-hmm. Yeah.

Danijela (29:57)
never yeah, exactly. I I don't even answer reply to the first message and then in two days I receive the second one, which

Tihon (30:06)
Mm-hmm.

Danijela (30:06)
you realize is really not

Danijela (30:08)
i it's automated, it's not related to the real conversation. And then you you already lost them on the second message, I would say.

Tihon (30:18)
Yeah, we we all been there.

Danijela (30:20)
Yeah, exactly.

Tihon (30:21)
Right, right, right. I'm also curious, like the marketer's job is now shifting, right? So

Danijela (30:29)
Mm-hmm.

Tihon (30:29)
okay, we have we will have an army of agents soon. So they would be doing each different function, they will get enhanced, super skillful. Right now they are like we are halfway there.

Danijela (30:43)
Mm-hmm.

Tihon (30:44)
In your opinion, what

Tihon (30:46)
marketer's job will look like? What would you focus on like when you when many parts taken care of? What are your thoughts on that?

Danijela (30:57)
Yeah, I speaking of marketer and army of agents, I'm already there. I even called this machinery that I explained to you earlier with the specialist AI agents. I call them my marketing and communication AI agent team. So that's my team at the moment. what

Tihon (31:12)
Yeah, right. Amazing.

Danijela (31:16)
what I've s what I see changing already now, marketing teams are getting smaller and smaller.

Tihon (31:22)
Yeah.

Danijela (31:23)
With that being said, those who are left in the company are supposed to do more work. So we are moving more toward being generalists than specialists.

Tihon (31:33)
Mm-hmm. Right.

Danijela (31:34)
And I think it's seen as a quality on the market today because most employers are expecting to you to do more things than people were used to do before when they were working

Tihon (31:43)
Mm-hmm. Mm-hmm.

Danijela (31:46)
in big marketing teams and everyone had specialized function.

Danijela (31:49)
Now they want to make to shrink marketing teams because they believe okay, AI can help us with a lot of different things. So you will mostly manage AI as a marketing in the future and AI agents and AI agencies. There are more and more

Tihon (32:06)
Mm-hmm.

Danijela (32:06)
AI agencies offering different services. So I think that will be the work.

Danijela (32:12)
future work and then you will be this kind of guy guardrail or quality controller or editor or whatever we call it that is saying what is the criteria and the the the way the path we we should follow and then s then work on the approval approval process.

Tihon (32:34)
Mm-hmm, right, right, right. Also, huh. sure, sure.

Danijela (32:35)
I don't think we will create sorry for interrupting, but I don't think we

Danijela (32:39)
will create a lot from scratch in the future.

Tihon (32:43)
Uh-huh.

Danijela (32:43)
So it will be more like a qu quality control.

Tihon (32:46)
Yeah, probably. Also, I see a lot of pushback on AI recently

Danijela (32:52)
Mm-hmm.

Tihon (32:53)
in the feed, what like around content creation. I mean like a lot of AI slop, and you've mentioned already about like quality control, you should add it properly, and so on so forth. Also, I see a lot of not

Tihon (33:10)
Proper confidence about the place of AI, you know, like in long cycle deals, finite markets. because yeah, managing relations, it's super hard. Then building the full funnel also kind of super hard. So small, small parts you can delegate, but still thinking and judging seems like

Tihon (33:32)
should be on human side because as soon as you try to delegate strategy fully tactics, you cannot fill fit all the context, all the details, all the information about the relations and stuff like that. And even I think right now, currently, if you will fit, if you would would manage to create like

Tihon (33:53)
perfect, you know, directories architecture with a lot of bunch of information,

Danijela (33:57)
Mm-hmm.

Tihon (33:59)
it still won't be as good as best people in the space. So yeah probably we won't be able to answer this question. The future will show where technology will lead us. But yeah what I'm saying right now I see like in the last especially three months big pushback

Tihon (34:18)
in in LinkedIn,

Danijela (34:19)
Yeah.

Tihon (34:19)
right? People just like shit, I I I cannot stand this AI stuff anymore.

Danijela (34:23)
Yeah.

Tihon (34:24)
I cannot do this, that, that. How do you think w what what it is about?

Danijela (34:28)
I noticed that too, and you we talked about it LinkedIn even added this special

Tihon (34:34)
Right.

Danijela (34:35)
feature where you can mark other people's content as AI. according to your judgment, AI generated and not of certain quality.

Tihon (34:43)
Mm-hmm, mm-hmm, mm-hmm.

Danijela (34:44)
I think we reached this point for a simple reason because people misused the tool, didn't put enough energy

Tihon (34:52)
Mm. Just pushing the same button,

Tihon (34:54)
right? That's it.

Danijela (34:55)
Yeah, exactly. And

Danijela (34:57)
I see it this way. Let's say you talk to the world famous expert in

Tihon (35:04)
Yeah.

Danijela (35:04)
the f in your field and you are seeking

Tihon (35:06)
Mm-hmm.

Danijela (35:07)
for an advice, but you are not giving that person context. so you just

Tihon (35:13)
Mm-hmm.

Danijela (35:14)
say, give me the best advice you can for me, but he or she exactly

Tihon (35:18)
Yeah, then it's a bad question, right?

Danijela (35:20)
He or she doesn't know who you are or for which context

Tihon (35:23)
Mm. Mm-hmm.

Danijela (35:25)
you need his or her advice. So

Tihon (35:27)
Uh-huh, right.

Danijela (35:28)
it's the same way how you approach AI. AI has enormous possibilities to create great output, but you need to give

Tihon (35:37)
Mm-hmm.

Danijela (35:38)
it a context. And to give it a context it means that you need to invest certain time

Danijela (35:43)
To give it a context. And that's why if you remember this first work while we were talking about, you you give it a context via meeting transcriptions so it understands what is your business and industry about. You give it a context about you by doing this interview that takes around two hours, and not all people are willing to invest two hours to

Tihon (36:03)
Yeah.

Danijela (36:04)
teach AI how you speak, think,

Danijela (36:07)
resonate about certain things. And

Tihon (36:10)
Mm-hmm.

Danijela (36:11)
that's what I mean. You need to give it the right context, you need to create the right prompt. Now it's not even so much about prompt. If you create these folders and connect it to right no knowledge source, it will understand what you or your business are about. So I believe this kind of generic approach and superficial approach to AI

Danijela (36:31)
Created too much generic outputs that we see

Tihon (36:35)
Mm-hmm.

Danijela (36:35)
everywhere, and that's what people are tired of. So

Tihon (36:39)
Mm. Right.

Danijela (36:41)
yeah. So I think the shift will happen when people learn how to use AI in the right way. But

Tihon (36:48)
Uh-huh.

Danijela (36:49)
again, we also mentioned this doesn't mean that you should outsource everything to AI, but certain things can be

Tihon (36:55)
Harry.

Danijela (36:56)
outsourced and help you save time and energy.

Tihon (36:58)
Totally,

Tihon (36:59)
totally.

Danijela (36:59)
on the

Danijela (36:59)
very very repetitive and boring things like reading again documents that no one is excited about.

Tihon (37:06)
Right,

Tihon (37:06)
right, right. I also cannot say that I belong to this camp that like that I'm pushing back on it. I say no way actually I use myself a lot and one use case I wanted to add to

Tihon (37:19)
Both we've mentioned earlier. So what also I do for my clients is social monitoring using AI when we collect the data from custom feed on LinkedIn to

Danijela (37:29)
Mm-hmm.

Tihon (37:30)
analyze what the best practitioners in our niche are talking about, what the narratives are about.

Tihon (37:36)
And this is done to understand the gaps in the narratives, what people are not talking about, but what where is the demand on side of

Danijela (37:46)
Yeah.

Tihon (37:46)
information, on side of knowledge? Because somebody would be writing in the comments like, hey guys, what's your take on that? And there's gonna be silence. So we we should be looking for this silence also from the overall picture in the industry. let's say it's about sustainability fashion or it's it's

Tihon (38:05)
about energy tech so we can make this list of people practitioners and companies who are

Tihon (38:11)
active and who are let's say main players in the industry and monitor them monthly and then understand what the market is talking about where we should go and so on so forth so AI is extremely helpful and if you get enhanced in prompting and collecting data and by the way yeah we've mentioned this before collecting data should be not scraping with agents because it's harmful for LinkedIn then link in LinkedIn can ban you ban accounts but it's it's possible to outsource

Tihon (38:41)
To people or to do it yourself manually, just to collect control C, Ctrl V. Yeah, just cross through custom field.

Danijela (38:44)
paste, yeah. But it's easy but this is a good example.

Danijela (38:49)
You still do some things manual like copy pasting because LinkedIn doesn't allow scraping

Tihon (38:52)
Right, right, right.

Danijela (38:54)
the the data, but then you use AI for the analysis, which would also

Tihon (38:59)
Yeah, exactly.

Danijela (39:01)
take you time to read through all these posts and comments. Yes. It's an excellent example.

Tihon (39:04)
Weeks. No jokes. It would be weeks to yeah,

Tihon (39:08)
yeah, yeah. And and then you have all this knowledge, all this analytics, and just okay, this is what happens. This is where we should go. Because like we

Danijela (39:16)
Mm. Right.

Tihon (39:18)
can base content on ICP pains. That's great. It should be around that. But what is happening currently in the industry, where the wind blows right now? What is the hottest take? Where's the zone of silence?

Danijela (39:30)
Exactly.

Tihon (39:30)
It should be this month.

Danijela (39:32)
Yeah, yeah, agree. Totally agree. This is a great example of how you can use AI to help to get information

Tihon (39:38)
Uh-huh.

Danijela (39:39)
that will help you resonate about and make decision about certain things. So it's it's a it's a mutual it's a mutual play nowadays. and I I'm happy while it's mutual play because I'm a exact that we are still part of the game. That's a good point.

Tihon (39:52)
That we are still part of the game. Right?

Danijela (39:58)
Because I'm afraid with this speed of development that soon. and both speed of development, we you have on the on one hand technology that is developing the fastest ever, and on another hand is human nature that is per se lazy. So that's another good combination. That leads to

Tihon (40:16)
Yeah.

Danijela (40:19)
the future where we will be eager to outsource and then there will b

Danijela (40:23)
There will no be way back.

Tihon (40:26)
Yes, yes, yes. I hope yeah, this this time will happen in some future future. Elon

Danijela (40:33)
Yes.

Tihon (40:34)
Musk in his recent interview to Economist, by the way, said like in turn in ten years you would not need any money because you'll live in abundance. The robots will create everything. But I'm like, okay, but who's gonna pay me this money? Yeah, everyone is

Danijela (40:46)
I wish exactly. There are some

Danijela (40:51)
science fiction movies that show a different scenario or different outcome. What happens in those

Tihon (40:55)
Yeah, yeah, yeah.

Danijela (40:57)
outop o utopian societies. I mean utopian

Tihon (41:00)
Yeah.

Danijela (41:01)
they sold us utopian idea but I don't think it will be Utopian.

Tihon (41:05)
Absolutely, absolutely. Daniella, and if we can in just like three, four minutes have

Tihon (41:12)
A short snippet for the listeners, because the the first case you've mentioned, and thanks you a lot for being willing to share this workflow. How exactly you solve content problems with AI. I saw this document, this playbook that you've shared on LinkedIn. It's super useful. Great. We will put it also as a link for the podcast. For the second use case, how you've built these six AI agents. Maybe you have a short

Tihon (41:40)
How to to share, like how it works.

Danijela (41:43)
Definitely,

Danijela (41:43)
I will definitely share it with you and your audience. So be I because I understand when you explain it,

Tihon (41:48)
Mm-hmm.

Danijela (41:49)
especially in this quite short conversation, it's difficult to understand on all levels or layers.

Tihon (41:54)
Exactly. Especially how to

Danijela (41:57)
Exactly, exactly. So I will share a document where it's explained step by step. So

Tihon (42:03)
Uh-huh.

Danijela (42:04)
don't get confused with my

Tihon (42:06)
Perfect.

Danijela (42:07)
jumping

Danijela (42:07)
from one to another how to workflow.

Tihon (42:10)
Perfect, perfect. Thanks a

Tihon (42:11)
lot. And I think like people who are really serious about implementing something, they will find out how to. This is never the question. It is always, you know, like the last step. When when you are sold on something and want to try, like, okay, you always can ask AI how to build something and it will just tell, right?

Danijela (42:30)
Exactly. And that's exactly

Danijela (42:32)
what I do. I mean, for every obstacle, for every step where I don't understand what to do next, you

Tihon (42:40)
Yeah, yeah.

Danijela (42:41)
you take a screenshot, you literally ask AI, okay, I

Tihon (42:45)
Yeah.

Danijela (42:45)
encountered this problem, what to do next? And it helps you. This is great. It's

Tihon (42:50)
Mm-hmm.

Danijela (42:51)
like having best coach by your side all the time, twenty twenty four seven, yes.

Tihon (42:54)
Yes, yes, yes. Personal tutor. Yeah,

Tihon (42:59)
another another probably one of the best use cases of AI. And I see we have just a couple last minutes as a last question

Danijela (43:08)
Mm-hmm.

Tihon (43:08)
and a bonus to our listeners. What are

Tihon (43:12)
the sources of information like your favorite you can share, maybe LinkedIn names, YouTube channels, books, blogs, whatever.

Danijela (43:22)
Now we talked a lot about AI, so I will share one name and

Tihon (43:28)
Uh-huh.

Danijela (43:29)
profile. We can also list it below this episode. It's

Tihon (43:32)
Uh-huh.

Danijela (43:33)
it's a man who writes a lot about AI and I personally learn a lot from him. I his name is Ruben Hasid. He's

Tihon (43:38)
Mm-hmm. Mm-hmm.

Danijela (43:41)
very famous in the AI space. He has his

Tihon (43:44)
Yeah.

Danijela (43:44)
blog on Substack and

Tihon (43:46)
Yeah.

Danijela (43:46)
sends

Danijela (43:47)
I think weekly newsletters if you subscribe.

Tihon (43:49)
Yeah.

Danijela (43:50)
and he's very honest, very methodological and pedagogical,

Tihon (43:56)
Uh-huh.

Danijela (43:58)
so explains really for ev for people who don't have any clue about AI or or don't have previous knowledge, they can follow his tips because he really does it step by step and even reflects on his previous posts.

Danijela (44:14)
Sometimes and say, I was talking about this way of working for a month ago, but now I realize that there is a better way. So I like that with him. He he's not stating things like there

Tihon (44:23)
Yeah. Amazing. Amazing. Yeah. Yeah. I know. Mm-hmm.

Danijela (44:28)
are absolute truth truth. So he's very self reflecting as well. So

Tihon (44:32)
Yeah. Yeah.

Danijela (44:34)
if you want to learn about AI and especially Claude, I would recommend

Tihon (44:37)
Mm-hmm.

Danijela (44:38)
to follow him. But to be

Tihon (44:39)
Mm-hmm.

Danijela (44:40)
honest with you, you learn the most from AI.

Danijela (44:43)
Actually, because you can ask any question to any LLM models and get pretty good answers. So I guess most of your listeners, including myself, we stopped using Google or searching for documents elsewhere.

Tihon (44:58)
Mm-hmm.

Danijela (44:59)
AI is today the best source for learning. And YouTube, of course. I hope people will watch this episode

Tihon (45:06)
Right.

Danijela (45:07)
on YouTube, and I'm personally

Danijela (45:09)
spending a lot of time on YouTube listening to long

Tihon (45:12)
Absolutely likewise.

Danijela (45:13)
long form formats, long form materials I mean or videos.

Tihon (45:18)
Yeah, yeah, a hundred percent. YouTube is is a treasure for information. Yeah, all right. Thanks a lot, Daniela. I see somebody is probably

Danijela (45:24)
So no no it it's fine.

Danijela (45:29)
I mean let's just wrap if you want this like a wrapping up part. Yeah, my colleague is here

Tihon (45:32)
Yeah. Yeah.

Danijela (45:35)
but no. Okay, thank you so much Tychon. It was a pleasure to be in your episode and

Tihon (45:39)
Thank you so much, Daniela.

Danijela (45:42)
really looking forward to continue our conversation even offline.

Tihon (45:46)
Absolutely like West Daniela, it was a pleasure. Thank you so much for sharing your thoughts. And I'm a hundred percent sure our listeners will find a lot of valuable scenes here. Thank you so much.

Danijela (45:58)
I really hope so. Thank you.