Author: Birdy Hill

  • AI Agents Are Coming for Marketing — What This Means If You’re a Team of One

    AI Agents Are Coming for Marketing — What This Means If You’re a Team of One

    Every marketing publication is currently writing about AI agents like you have a technology team, a governance function, and a six-figure software budget. Most of us don’t. Here’s what actually applies if it’s just you.

    Open any marketing trade publication right now and you’ll find the same story: AI agents are the defining shift of 2026. By the end of this year, an estimated 40% of enterprise applications will include task-specific AI agents, up from under 5% just a year ago. Whole campaigns — audience discovery, execution, optimisation — are increasingly being run by coordinated networks of specialised agents rather than a marketing team working through a task list.

    It’s a genuinely significant shift. It’s also being written about almost exclusively for people who have a technology team, a martech stack with a real budget behind it, and someone whose job it is to worry about governance. If you’re running your own blog, your own small business, or working as a solo or fractional marketer, none of that coverage is really written with you in mind. So here’s the version that is.

    What an “agent” actually is, without the jargon

    A chatbot answers a question when you ask it one. An agent does something — and keeps going, across multiple steps, without you approving each individual step along the way.

    The practical difference: asking an AI tool to write a social caption is using it as a chatbot. Asking a tool to check your analytics every Monday, identify your best-performing post from the week, draft three variations of a follow-up post based on it, and queue them for your review — without you doing each of those steps by hand — is closer to what “agent” means. The system is carrying out a small workflow, not just answering a single prompt.

    That’s the entire concept. It gets dressed up in a lot of enterprise language about “multi-agent architectures” and “orchestration,” but at solo-operator scale, an agent is simply a tool doing a repeatable multi-step job for you, reliably, without you babysitting every step.

    Where this is genuinely useful if you’re working alone

    The enterprise version of this trend is about running entire campaigns autonomously. The realistic solo version is much smaller, and much more useful, because it targets the specific problem solo operators actually have: there’s only one of you, and the repetitive operational work is what eats the time you’d rather spend on the actual thinking.

    A few realistic starting points:

    Content repurposing. One long-form post becomes a LinkedIn post, three social captions, and a Pinterest pin description — as a defined, repeatable workflow rather than something you redo manually from scratch every time.

    Research and monitoring. A standing task that checks a handful of sources or a search dashboard on a schedule and flags what’s actually worth your attention, instead of you doing that scan yourself every week.

    First-pass admin. Sorting and drafting responses to routine enquiries, freeing your actual attention for the ones that need judgement rather than a template.

    None of this replaces you. All of it removes the unglamorous, repetitive layer of the work, which is precisely the layer that’s hardest to justify spending your limited time on when you’re the only person doing everything.

    The part almost nobody is saying out loud

    Here’s the uncomfortable finding buried in most of the 2026 agentic AI coverage, if you read past the enthusiasm: organisations are pouring money into content generation — something like 22% of AI budgets, with over 80% adoption — while spending roughly 3% on governance, with adoption sitting around 31%. In plain terms: everyone’s switching on the tools that produce output, and almost nobody’s building the checks that make sure the output is actually right.

    At enterprise scale, that gap creates compliance risk and brand inconsistency across huge volumes of content. At solo scale, it creates something smaller but just as real: an agent confidently publishing something in your name that’s wrong, off-brand, or simply not something you’d actually say — and nobody catching it, because there’s no second person in the loop.

    This is the one piece of the trend worth taking seriously before you adopt any of it: an agent should never be the last check before something goes out under your name. If you can’t yet look at everything an agent produces before it publishes, you’re not ready to let it publish unsupervised — you’re ready to let it draft.

    A sensible way to start

    Don’t start with a fully autonomous multi-step agent running your whole content pipeline. Start with one narrow, low-stakes, easily reviewed task, and only let it run without your review once you’ve watched it get that one task right, consistently, for a few weeks.

    The order I’d suggest: repurposing first (lowest risk, easiest to check, most obviously time-saving), then research and monitoring (medium risk, mostly about saving you a manual scan), and only later anything touching direct reader or customer communication (highest risk, needs the most oversight, least suited to being fully autonomous for a one-person operation without a safety net).

    The honest summary

    The agentic AI trend is real, and it will keep growing. But almost everything written about it assumes a team, a budget, and a governance function standing behind the technology. If you’re working alone, the version that’s actually useful to you is much smaller than the headlines suggest: a handful of narrow, well-defined, closely reviewed workflows that take the repetitive load off your one pair of hands — not a fleet of autonomous agents running your business while you sleep.

    Start small, keep yourself in the loop until you’ve earned the trust to step back, and don’t let anything publish under your name that you haven’t personally seen first.

    Your action list this week

    If you do nothing else, do these three things:

    • Pick one repetitive, low-stakes task in your workflow — content repurposing is the easiest starting point — and set up a simple, repeatable version of it.
    • Decide, in writing, what “reviewed by me before it publishes” actually looks like for your process, before you automate anything.
    • Hold off on anything that touches direct reader or customer communication until you’ve watched a lower-stakes task run reliably for a few weeks

    If you’re already experimenting with agents in your own workflow, I’d love to know what’s actually working — and what’s made you pull back.

  • Using AI to Build Resilience at Work — A Real Example

    Using AI to Build Resilience at Work — A Real Example

    This is not an article about productivity. It’s about what happens when work gets overwhelming and your inner voice starts telling you it’s your fault. And how, unexpectedly, AI helped me interrupt that pattern and find my way back to solid ground.

    There’s a particular kind of professional misery that doesn’t show up in any job description.

    It’s the misery of carrying ambiguity. Of being in a situation where the process is unclear, the expectations are unstated, the history is patchy, and yet somehow — in the quiet of your own head — you have become the person responsible for all of it.

    You run the situation over and over. You look for the moment where you went wrong. You take on responsibility for things that were never yours to begin with. And the spiral tightens.

    I know this pattern well. I’ve lived it more times than I’d like to admit. And this week, in the middle of exactly that kind of situation, I did something that helped me more than I expected.

    I talked to an AI.

    A little history — and why I thought to try this

    Years ago I came across an app called Woebot.

    It was created to support young people with their mental health — a conversational AI built on cognitive behavioural therapy principles, designed to help people recognise and interrupt unhelpful thought patterns. I tried it and I remember thinking: this is something.

    What struck me wasn’t the sophistication. It was the absence of judgment.

    When you write to a human — even a trusted friend, even a therapist — there is a layer of performance. A part of you is aware of how you’re coming across. You edit yourself. You present a version of the situation that doesn’t make you look too bad, or too weak, or too much.

    With Woebot, that layer disappeared. Because it wasn’t a person. It couldn’t think less of you. It had no memory of you from last week and no opinion of you for next week. You could write your actual thoughts — the unedited, embarrassing, circular, desperate ones — and something would respond with genuine, considered, non-judgmental reflection.

    That experience stayed with me. The idea that the absence of human judgment creates a space where you can be more honest than you would be with another person. Not because the AI is better than a person. But because the thing you need in that moment isn’t wisdom. It’s objectivity. And objectivity is very hard to find when you’re emotionally in the middle of something.

    This week

    I’ve been carrying a complex, long-running situation at work. I’m not going to go into the specifics — client confidentiality and professional discretion matter — but the shape of it will be familiar to anyone who has managed a complicated account or project over a long period of time.

    Ambiguous process. Unclear ownership. A history of decisions that weren’t fully documented. Gaps in records that made resolution harder than it should have been. And a moment this week where I needed to respond to a finance query, action something I hadn’t been clearly instructed to do, and do both in a way that didn’t make me look like I was blocking progress or failing to understand something I should have understood.

    In the past, I would have handled this by going very quiet inside my own head and very busy in an unhelpful way. Running the situation on a loop. Asking myself where I went wrong. Wondering whether I should have anticipated this. Taking on responsibility for process gaps that existed long before I was involved.

    It would spiral. And the spiral would make everything harder — the clarity I needed to respond well would be buried under the weight of self-criticism.

    This week, instead, I opened Copilot.

    What happened when I asked for help

    I didn’t ask Copilot to write the email for me. I asked it to help me think.

    I described the situation. I explained what I’d been asked to do, what information I hadn’t been given, what I was worried about, and what I needed the response to achieve. I told it I was concerned about being seen as blocking progress. I told it I wasn’t sure whether I was supposed to know something I didn’t know. I told it I felt exposed.

    And something shifted almost immediately.

    Because the AI did something I couldn’t do for myself in that moment: it separated the emotional charge from the objective reality.

    It said — in various ways, across various iterations of thinking it through with me — that what I was experiencing wasn’t a knowledge gap. It was a process gap. That the instructions I hadn’t been given weren’t things I should have known or found. That the ambiguity I was carrying wasn’t mine to carry alone. That asking for clarity wasn’t an admission of failure. It was exactly what a senior professional should do.

    It said: you’re not behind here. You’re not missing anything. You’re actually applying better control than the process itself gives you.

    I want to be honest about what that sentence did for me.

    It didn’t solve the problem. The situation was still complex. The email still needed to be written. The process gap was still real. But something loosened. The spiral slowed. I could think clearly again.

    What AI does that a colleague can’t always do

    A trusted colleague can offer something an AI can’t — the depth of human understanding, the shared history, the warmth of someone who actually knows you and your context.

    But a trusted colleague also has opinions. A relationship with you. A stake in the situation. Sometimes — often — they are the last person you can be fully honest with about how frightened or confused or overwhelmed you actually are.

    The AI has none of that. It has no stake in how you come across. It has no memory of the last time you struggled with something similar. It has no capacity to think less of you tomorrow. It responds to what you actually write rather than to a performance of what you wish you were feeling.

    That absence of judgment is not a limitation. In certain moments, it is precisely the thing you need.

    What it offers is perspective. The ability to see your situation from outside your own emotional charge. To have someone reflect back: here is what is actually happening, here is what is objectively true, here is what you do and do not have responsibility for.

    That is what therapists call cognitive reframing. And it is one of the most powerful tools in building genuine resilience — not the resilience of gritting your teeth and pushing through, but the resilience of being able to see clearly when everything feels foggy.

    The pattern I recognised in myself

    I want to name something, because I think it might resonate with people reading this.

    I have a pattern — and I suspect I’m not alone in it — of taking on responsibility for situations I didn’t create. Of looking for the moment I went wrong when the situation itself was wrong. Of believing, at some level, that the ambiguity or the difficulty or the mess must be attributable to something I failed to do.

    I’ve thought about where this comes from. Partly from being someone who cares about doing things well. Partly from years in environments where the standards are high and the margins for error feel small. Partly — I think honestly — from being a woman in a senior role in an industry where the expectation of having all the answers sits heavier on some people than on others.

    The spiral is the mind’s attempt to find control in an uncontrollable situation. If I can just identify what I did wrong, I can fix it. If I can just understand where the failure was, I can prevent it from happening again.

    The problem is that when the situation is genuinely ambiguous — when there is no clear failure point, just accumulated complexity and unclear process — the spiral finds nothing and keeps searching.

    AI interrupted that. Not by telling me everything was fine. But by helping me see, clearly and without judgment, what was actually mine and what wasn’t.

    A practical suggestion

    If you are going through something at work that feels overwhelming — a situation you keep running on a loop, a problem where you can’t find the solid ground — try this.

    Open whichever AI tool you have access to. Copilot, Claude, ChatGPT — it doesn’t matter which.

    Write out the situation as honestly as you can. Not the polished version. Not the version you’d tell a colleague. The actual version. What happened, what you’re afraid of, what you’re blaming yourself for, what you don’t know.

    Then ask it to help you think through what is genuinely your responsibility and what isn’t. What you know and what you were never told. What the process requires and whether that process was ever clearly given to you.

    You may be surprised by what becomes visible when you step outside your own emotional charge and look at it from the outside.

    You don’t have to act on everything it says. You don’t have to treat it as a substitute for a trusted colleague, a mentor, or a therapist when those things are what’s needed. But as a first step — as a way of interrupting the spiral and finding enough clarity to move forward — it is more useful than I expected.

    Why I’m sharing this

    I am not usually someone who finds it easy to admit that a situation got under my skin. Senior professional life has a particular culture around this — the expectation of apparent competence, the performance of having it handled, the reluctance to show the underneath.

    But I’ve been thinking a lot lately about what it would have meant to find something like this earlier. In the hard weeks of my career when I was running things on a loop and had no good way of stepping outside them. In the moments when I was carrying ambiguity that wasn’t mine to carry and didn’t have the language to put it down.

    I feel genuinely fortunate to have had the tool available this week. Not because it solved everything. But because it helped me be a little kinder to myself at a moment when I needed it. And it helped me show up more clearly — for the situation, for the people around me, and for myself.

    If this is useful to one person reading it, it’s worth having written.

    A note on finding the right support

    Woebot, the app I mentioned earlier, is no longer available to new users — it closed its consumer service in 2025. Worth knowing if you go looking for it. This space moves quickly, and no single app is ever the right answer for everyone.

    If you’re looking for something for a young person in your life, Collective Psychology, a UK clinical psychology practice, maintain a well-curated list of apps for children and young people, put together by working clinicians rather than a marketing team.

    But the single most useful thing I can tell you is this: if you or a young person in your life is struggling, the right first step is your GP. An app can be a genuinely useful complement, and a real first step when you’re not yet ready to say the actual thing to a real person. It’s never a substitute for what a GP can help you access — proper assessment, therapy, the right referral. No app, however well designed, can offer that.

    If something in this resonated — a pattern you recognise, a situation that felt similar, a moment when you couldn’t find solid ground — I’d love to hear about it. Leave a comment below.

    And if you know someone who might need to read this, please share it. Sometimes the most useful thing we can do for each other is pass something along at the right moment.

  • Is AI Finding You? A Practical Guide to Being Discovered in the Age of Generative Search

    Is AI Finding You? A Practical Guide to Being Discovered in the Age of Generative Search

    You’ve probably noticed that when you type a question into Google, you now sometimes get a paragraph of AI-generated answer before you see any links. That’s not a glitch. It’s the future of search — and if you run a business or a blog, it changes almost everything about how people find you.

    Something shifted in how I think about content strategy recently, thanks to a lifestyle blog I run on the side — nothing to do with my day job. I set it up originally half for a bit of pocket money, half because I wanted the lived experience of social media, blogging and the influencer world from the inside, rather than just knowing how it all worked in theory. It’s turned into a genuinely useful place to test, learn and refine my thinking.

    It had lain dormant for a while — a busy working schedule got in the way of keeping it updated — so I gave it a fresh look.

    I’d been doing all the right SEO things. Keywords in the right places. Restart of regular posting. Establishing a Pinterest strategy. Ensuring keywords are entered in the metadata. And then I started paying attention to something different.

    I started typing my own questions — the questions my readers would type — into AI tools. ChatGPT. Google’s AI Overview. Perplexity. And I started asking: when someone searches the exact questions my niche cares about — is my blog anywhere in the answer?

    Sometimes yes. Sometimes no. And the difference between yes and no had very little to do with traditional SEO.

    That’s what this post is about.

    First — what has actually changed?

    Until recently, if you wanted to be found online, the goal was simple: rank on Google’s first page. Someone types a query, Google shows a list of links, they click the most relevant one. You optimise your content for keywords, Google sends you traffic.

    That model is shifting. A rapidly growing share of consumers now start their searches with AI tools — typing their questions into ChatGPT, Perplexity, Google’s AI Overviews, or Microsoft Copilot — and getting a synthesised answer rather than a list of links.

    The AI doesn’t copy and paste from websites. It rewrites and merges information from several pages into one answer, and if your content is one of the sources it draws from, you get cited. If it isn’t, you’re invisible to that searcher, regardless of where you rank on traditional Google.

    Research from Ahrefs found that the overlap between top-10 Google rankings and the pages cited in Google’s AI Overviews fell from 76% to 38% within months — a sharp drop that reflects AI systems drawing from a much wider pool of sources than the traditional top rankings. In other words: ranking well on Google no longer guarantees you’ll appear in AI search results. They are increasingly different systems, with different preferences.

    This shift has a name: Generative Engine Optimisation — or GEO. It’s the practice of making your content visible, credible and citable in AI-generated answers. And for small business owners, bloggers and independent brand builders, understanding it now is a genuine competitive advantage, because most people haven’t caught up yet; and the big brands will look through the lens of efficiency, aim for the broadest target audience to reach and not the niche.

    Why your niche is actually an advantage

    Here is the counterintuitive thing about AI search: specificity wins.

    AI systems reward statistical predictability, structural clarity, demonstrable authority, and proprietary data. When someone asks a broad question — “what’s a good moisturiser?” — the AI draws from hundreds of sources and your site is one of thousands. When someone asks a specific question — “what moisturiser works for menopausal dry skin with Asian colouring” — the pool of genuinely relevant, authoritative sources shrinks dramatically.

    This is exactly the insight that shaped my strategy going forward for that blog.

    I write for a specific reader there, not a broad one — someone I could describe precisely, right down to what stage of life she’s in and what she’s tired of being talked down to about.

    That specificity, which can feel limiting when you’re building an audience, is a significant advantage in AI search. Because AI is looking for the most authoritative, specific, trustworthy answer to a specific question. And if you’ve built a body of content around a real niche, you are exactly what it’s looking for.

    So before we get into tactics: if you’re building a business or a blog, get specific about who you serve and what you uniquely know. The riches, as they say, are in the niches. In AI search, this is more true than ever.

    The tactics: 8 practical things to do

    These are ordered from most to least urgent — start at the top and work down.

    1. Make sure AI can actually read your site

    Before anything else: check your robots.txt file. Many sites block AI crawlers without realising it. If you use Cloudflare, your AI bot traffic may have been shut off without your knowledge.

    In WordPress, go to Settings → Reading and make sure “Discourage search engines from indexing this site” is NOT ticked. Then Google “robots.txt checker” and paste in your URL to see what your site is currently allowing.

    This is the most basic step and the most commonly overlooked.

    2. Answer the question directly — at the start

    AI engines don’t read content the way people do. They break pages into individual passages and evaluate each one for relevance, clarity, and factual density. Every section needs to stand on its own. Start each section with a clear, direct answer.

    In practice: if your post is about the best facial oil for menopausal skin, your opening paragraph should answer that question directly. Not “In this post I’m going to explore…” but “The best facial oil for menopausal skin is one that…”

    The AI extracts passages. The passage that answers the question most clearly and directly is the one it will use.

    3. Write in natural, conversational language

    Classic SEO is based on keywords. GEO is based on conversational queries.

    People don’t type “best moisturiser menopausal skin UK 2026” into ChatGPT. They type “I’m going through menopause and my skin is really dry — what should I use?” Write content that answers the question the way a real person would ask it.

    This also means thinking about the questions your readers actually ask you — in comments, in emails, in conversations. Those questions are your content brief.

    4. Demonstrate that you are a real person with real experience

    AI search has stripped away the gimmicks of the past decade. Keyword stuffing, thin content, and manipulative link-building are entirely ineffective against models trained on billions of parameters of natural language. In 2026, you earn citations by being exactly what the AI is looking for: a clear, authoritative, and brilliantly unique source of truth.

    The term used in the industry is E-E-A-T: Experience, Expertise, Authoritativeness and Trustworthiness.

    For a personal blog or small business, this means:

    • Write from genuine personal experience (“I tried this for six weeks and here’s what happened”)
    • Include your credentials naturally (“as someone who has worked in marketing for thirty years…”)
    • Use your real name and have a real About page that establishes who you are and why you know what you know

    This is why your About page matters more than you might think. It is one of the key signals AI uses to assess whether your site is a trustworthy source.

    5. Structure your content clearly with headings

    How To Schema is perfect for capturing step-by-step instructions in AI results. Article Schema defines the who, what and when of your content. In 2026, AI engines use this to verify E-E-A-T.

    In practice, for bloggers without technical knowledge:

    Use clear H2 and H3 headings that describe exactly what each section covers — not clever or cryptic, but descriptive. “The best moisturiser for dry menopausal skin” is a better heading than “My hero product.” The first one the AI can extract and use. The second one it can’t.

    Yoast SEO (the free WordPress plugin) handles the technical schema markup automatically once your headings are correctly structured.

    6. Be present beyond your own site

    AI search pulls from the entire web, not just your site. To be cited, you need to be mentioned in the places the models trust most: earned media on reputable publications, and community authority — AI models crawl Reddit, Quora, customer reviews and niche forums to understand real-world sentiment.

    For small businesses and bloggers this means:

    • Guest posts on other sites in your niche (even small ones)
    • Being mentioned by other bloggers or publications
    • Participating genuinely in relevant online communities
    • Getting product or service reviews on third-party platforms

    You don’t need to be in Forbes. You need to be referenced by multiple credible sources within your specific niche. Depth of authority in a niche matters more than breadth of fame.

    7. Keep your content fresh

    AI systems seek up-to-date information. An article written in 2022 and never updated is unlikely to appear among the results generated in 2026. Update your top content at least every twelve months, adding new information and current trends.

    Go into your five best-performing posts and add a short update section at the top or bottom: “Updated June 2026 — I’ve been using this product for another year and here’s what’s changed.” That signals to both AI and Google that the content is current.

    8. Think across all your channels — consistently

    This is the strategic point that ties everything together.

    AI doesn’t just read your blog. It reads your Pinterest descriptions, your Instagram captions, your LinkedIn posts, your newsletter. The more consistently you talk about the same topics, using the same language, across multiple platforms — the more authority signals you build around your specific niche.

    For my own blog, this means the same language appears everywhere — the blog itself, Pinterest pin descriptions, Instagram captions — describing the same reader, in the same words, consistently, wherever she might find me.

    If you’re running a styling business, a shopping guide, or a lifestyle brand — the same principle applies. Define the two or three phrases that describe exactly what you do and who you serve. Use them everywhere, consistently, in natural language. That consistency is what builds the authority signal AI is looking for.

    The honest summary

    AI search is not replacing what you already do well. It is adding a layer on top of it — one that rewards specificity, genuine expertise, clear writing and consistent presence across channels.

    If you’ve been building a real business or a real blog, serving a specific audience, writing from genuine experience — you are already doing most of what GEO requires. The adjustments are mostly structural: clearer headings, more direct answers, a check that AI can actually access your content, and a habit of updating your best work regularly.

    The businesses and bloggers who will win in AI search are not the ones with the biggest budgets or the most technical expertise. They are the ones who know their audience deeply, write for them honestly, and show up consistently across every channel where that audience might find them.

    That has always been good marketing. AI search just makes it more measurable.

    Your action list this week

    If you do nothing else, do these three things:

    ✅ Check your robots.txt isn’t blocking AI crawlers
    ✅ Update your About page so it clearly establishes who you are and what you know
    ✅ Pick your three best posts and rewrite the opening paragraph of each to directly answer the main question the post is about

    That’s your starting point. Come back next month and I’ll be writing about how to use AI tools to create content that AI search will actually cite — the practical workflow I’ve developed for my own sites.

    Have you noticed AI search changing how people find your business or blog? I’d love to hear what you’re experiencing in the comments — or email me directly at hello@ravenintegrated.com

    And if you found this useful, the best thing you can do is share it with one person who needs it. Word of mouth still beats every algorithm.

  • Use the 80/20 principle for greater productivity in marketing

    Use the 80/20 principle for greater productivity in marketing

    Use the 80/20 principle for greater productivity in your marketing effort

    If you are starting a new business, it’s important to prioritise tasks that are most important. Return on investment isn’t just a benefit of marketing investment; it’s a way to think about how much time and effort you invest.

    Richard Koch’s book “The 80/20 principle” is about concentrating attention on results. He uses Pareto’s theory that 80% of consequences come from 20% of the causes. You can use the principle to prioritise marketing activity.

    For instance, when you can identify 20% of the customers/clients that yield 80% of profits for your business, you can look for more clients that have similar characteristics.

    How about identifying 20% of the tasks that generate 80% of output? or figuring out 80% of the tasks that can be delegated to free up more time on the tasks that only you can do?

    This principle works when you consider the science. Research conducted by psychologists about our ability to multitask showed that switching between tasks is not without consequences such as reduction in memory, productivity and performance.

    Working out your priorities requires the focus of one task at a time, when combined with the 80/20 principle, you’ll find that you’ll achieve more with less.