"Zazu Runs My Calendar, Tinkerbell Runs My Website"

Sixteen years in B2B, six of them working with AI, Louise Read has built herself fifteen AI agents to do her bidding, all named after Disney characters, and still isn't sure what to tell the next generation. She tells Rich Fitzmaurice why.

Louise Read built her AI agents on maternity leave. Nursery days are Tuesday, Wednesday, Thursday, which means the rest of the week is jammed, and this particular week she was turning around a LinkedIn campaign with deliverables landing faster than she could quite keep up with. “I just feel like my head’s all over the place,” she said, by way of hello.

•   •   •

Rich: What's your general sentiment towards marketing and AI?

Louise: That we're not ready. But I feel very positively about it. I think I'm lucky in the sense that I'm much further down in my career. I've worked in B2B for sixteen years now, and with AI for six. So, I've had the opportunity to work with incredible brands that have used AI to fight counter-terrorism, working with the White House and the US Army, and with an insurtech company that helped pay out people faster in the LA fires so they had homes and places to stay. I've worked with health tech solving blood testing problems. AI has an incredible way of solving very complex issues, things that have been problems for years. So, I have a positive attitude towards it, and I've seen the benefit from it in my own career.

However, I do worry about the younger marketers, because I'm able to be that orchestrator who's had the opportunity to work through a huge range of different marketing specialisms. I have a good view of what good looks like, and I'm able to orchestrate AI to make it work in my favour. I don't know that necessarily the younger generation are able to do that.

There's a huge sway of marketers who ask me how they get a job, how they get into the industry. And I find that really challenging to have an answer to, because I'm not sure what the answer is right now, as we're automating that initial layer.

•   •   •

Rich: Things are very cyclical, and part of me quite enjoys the fact that you always hear people on stage saying tactic X is dead. SEO is dead, PR is dead, events are dead. And actually, they're all coming back. Content writing and PR was supposedly dead, and right now it's more important than ever, ironically, because of AI search.

Louise: I hate that, the “SEO is dead” thing. Back at the beginning of my career, at Experian, we were selling data management software. When people are making a huge purchase like that, spending half a million pounds, and they're a big bank, they are looking at the Gartners and the Forresters of the world for their magic quadrant, to understand what technology in the category they should be investing in. Because they can't make a mistake or a poor decision at that level. So, I kind of disagree when everyone says the funnel was never linear, because in some cases in B2B I do think it actually was. But now we have this proliferation of channels, and there's an expectation that as a brand you need to be in so many different places. So of course, the funnel can never be linear now.

I originally thought, eight months ago, that video was a really good place to invest in, from a LinkedIn point of view, to show that you're the human. But now with AI avatars, I'm not sure video is that channel anymore. Podcasts, you can manipulate someone's voice. So the moat isn't content anymore, content is too easy to fake. The moat has got to be in person, community driven, and possibly experimental activations that are outside the norm. Shifting B2B into a B2C model, which honestly, we've needed to do for years.

Rich: We've always looked on with envious eyes at B2C.

Louise: We have. And I think we've relied on old ways for too long.

•   •   •

Rich: Are there any particular job titles you feel are at risk more than others?

Louise: Yes, unfortunately. Someone was asking me at Pilates about a promotion into paid media, advertising, Google, LinkedIn, Meta, X. It's very analytical, and it's algorithmic. You're looking at a data point and then changing an ad off the back of that.

At scale, AI is probably a lot better at that than humans. I can say that because I was a performance marketer, I owned that channel, and I've automated that entire flow. I used to spend my entire day pulling data from across different platforms, aggregating it into a spreadsheet, and making decisions off the back of it. Now I've automated it. There are tools you can just purchase that do it off the shelf.

Rich: You may not eliminate it entirely, there's always going to be average adverts and good ones, so there's always an opportunity for creativity. But in terms of the volume of jobs in that particular skill set, logic dictates it's going to be a lot lower.

Louise: It's going to be a lot more automated. But you're right that there'll probably still need to be someone orchestrating it. And there'll be other channels that come out, AEO for one. We've just heard that ChatGPT is allowing advertising. So it's only a matter of time before the other LLMs have a paid media option too. You'll still want someone able to think in that capacity and orchestrate it. But if I think really long term, as in a decade, do I really think humans will still be doing paid advertising like that? Honestly, I'm not sure. There's a need for it now, but I don't know that there necessarily will be as the models get better.

•   •   •

Louise: There's a stat, about 95% of existing marketing teams haven't yet seen ROI on AI, because a lot of the use cases we're building towards are time saved, as opposed to commercial objectives, a revenue number or a KPI you'd tie a human team member to. Can you imagine going to the board and saying, we're going to hire someone to save time?

As marketing functions we need to come back to what we're good at, tying the programmes we're running back to commercial objectives. In the same way you'd build out your marketing team, you need to build out your agentic team and think about the KPIs you're going to measure them against.

Rich: Attribution is something people claim to have. I've been a CMO in very large organisations, and honestly, as long as the pipeline's there and the sales number is great, I'll take the last point of identification and have faith the campaigns did the job.

Louise: It's hard, but businesses need some key metrics they can track, revenue, net new logos, influenced pipeline, and correlate that to AI. They need to know what's working against those metrics before even thinking about automating it. What I see happen a lot is businesses hiring junior marketers because it's cheaper. Nothing against junior marketers, but they don't have the experience to build out a strategy. So they throw more and more at the wall and hope something sticks. Six months down the line the CEO or the CRO comes back and says, I'm fed up, marketing isn't generating against the commercial goals. It's unfair on the junior marketers, they don't have the know-how. And it's unfair that we've decided, because the CEO and the sales team think everyone knows marketing, that anyone can do it.

Rich: What would your advice be to a young marketer right now, straight out of university, trying to get that first full-time role?

Louise: God, honestly Rich, it really scares me. I've actually been asked this question, and I'm not sure I would encourage people to go into marketing. I know I need to be optimistic, but I don't know that necessarily I would.

Models will become increasingly more sophisticated, and 85% of roles that will exist by 2030 don't currently exist today. If a marketing team is made up of five to fifteen people now, within the next decade we'll probably see this marketing engineer role, a master orchestrator pulling together different agents doing different things. Realistically, do we think marketing teams will be fifteen people in ten years? We can't, because we know we're going to have the orchestrator, we know there's going to be more automation. So what areas will still be prevalent? The person who understands it all and can manage the agents, that's someone who's been in a career a long time, who's gone deep across those areas. It's not necessarily open to someone entry level.

If you're already in a role, my advice would be to hone in on AEO, get as much experience and own as many projects as you possibly can. Get involved with event and community projects, building activations, panels, webinars, doesn't matter, just events and community, and put your name to building a community. And brand, if your company has the money to spend on it, get involved in those activation projects.

•   •   •

Rich: You mentioned, very in passing, fifteen AI agents. What on earth are they doing?

Louise: They're all Disney characters, because I'm a mum. Let me tell you about my agents. I have some business ones and some personal life ones.

Zazu is my chief of staff, he manages all my agents, and he's the only one that really talks to me and tells me what the others are doing. He sends me a daily summary of what's happened and what actions I need to take.

Belle helps me with my newsletter, because Belle likes reading. She doesn't write it, she understands my ICP, B2B marketers and sales leaders, and looks at different topics that are front of mind for that audience, surfacing articles, podcasts, active discussions and debates from the last week, and summarises the top three or four that are currently being talked about. She also surfaces language patterns, dos and don'ts, don't use the word AI slop, don't use MQL, everyone hates that word. She helps me pick a topic, and I send her a voice note with the themes I want to talk about, and she finds data points to add.

Lumiere is my client manager, she prepares me for all my meetings, takes my Granola notes from the last one, and makes sure we're staying on track against the project plans. Al does all my reporting, to clients and for myself. Jeannie is my social assistant, she takes my tone of voice and condenses my voice notes into social posts that I then edit. Mary Poppins manages my personal finance, because I spend too much money on coffee.

And then there's Tinkerbell, who does my website, and my fairy godmother. Zazu is my chief of staff, but my fairy godmother is the only agent that can override him. She asks me every morning how I'm feeling, whether I slept okay, how much I've got left in the tank. If I tell her I'm feeling really low, that I can't do any work today, it changes the output Zazu has planned for me. It'll move my meetings around and change my to-do list for that day.

Rich: Crazy, isn't it. And this is all Claude agents?

Louise: All Claude, all self-built. Just actions, things I needed solving. I've only built all these since April, and I'm entirely self-taught. I'd describe myself as an expert power user, but you can teach yourself this very quickly and go very deep if you have the time. I had the time because I was on maternity leave.

•   •   •

Rich: What have you seen from marketers using AI that's actually impressed you?

Louise: I worked with a marketing manager, a really strong content writer, who'd come up as a content writer and been promoted into the role. She wouldn't necessarily say she knew the fundamentals of SEO or GEO, and there are a lot of content marketers out there who haven't yet learned those areas. She felt comfortable enough to be vulnerable and say, I know I'm a good content writer, but I don't yet know how the mechanics of SEO or GEO work. I worked with her to build a skill that took my experience of how I'd think about SEO or GEO and used it to create her content briefs, so she could lean into her own skill set, writing, without needing to know what keyword to pick or where to focus her time that week. That's what's building her content calendar now.

I'd say that's one of the most impressive things I've seen, partly from a human empathy angle, because it's not common to see people admit that.

Rich: What frustrates you? What do you not like seeing?

Louise: One example, and this comes up a lot with the series A to B SaaS companies I work with, is a junior marketer under pressure from senior leadership to produce the strategy. They lean on AI to build it up, and AI presents this wonderful document that sounds really great, but it's rubbish. It means nothing. It's not correlated to the data. It doesn't understand the ICP, the customers, the pain points, the competitive landscape. That's a complete waste of time, because you're just setting yourself up for failure. Everyone in the organisation knows you've used AI to do it, and you can't even execute against it because it doesn't make sense. So definitely not for marketing strategy.

I'd argue the same about founder-led personal brand work. A lot of marketers are suddenly being tasked with writing the LinkedIn of their CEO, and I think that's wrong. The CEO needs to take ownership of that themselves, it will never be authentic in their voice otherwise. I don't mind if the CEO does a brain dump voice note and the marketing exec uses AI to help structure that into a post. But personal branding should never be written directly, going into Claude and saying, write me five posts for LinkedIn.

Rich: I reckon that's pretty prevalent though. The antithesis of that is the COO of Blackstone doing his viral videos running round Central Park. Enormous following, unscripted, not sanitised, whatever's on his mind while he's out for a run. That goes back to the fundamentals of marketing, you're not going to stand out by doing average. And AI, at the moment, is a very different definition of average, because it can't think for itself, it's an amalgamation of everything that's already been created. There's no creativity, no new ideas, no strategy to it. People are far too quick to delegate their thinking to a machine. It's quite lazy.

Louise: I like that, far too quick to delegate their thinking to a machine. That's actually spot on. I've seen people writing about the fact that the last thing we should automate is our own thinking. We need to continue to train our own thinking, our own decision making, our own thought process. As there are more things we can automate, that's the one thing that will always remain true. Make my thinking clearer, help me put some structure to my ideas, help me visualise them, fine. But the issue comes right at the beginning if you go down the route of, come up with an idea for me.

The AI is also, and this is the thing I catch myself fighting with constantly, too accommodating. It'll always give you the answer you want to hear. And I do think we have to be careful about not leaning on it too much, because it will never call you out and say, actually, you're wrong.

•   •   •

Louise Read is a fractional CMO and AI marketing consultant via her company Codi Marketing, ex-Klaviyo marketing leadership, working with CEOs, CROs, founders and B2B marketers to tie AI to commercial outcomes rather than time saved.