In this blog post I'm going to walk you through how I, a non-technical fifty-something marketing veteran with the coding skills of a confident teenager, actually use AI in my business every single day. Specifically the bits that work, the bits that don't, the bits the bro-marketer crowd is lying about, and the seven-day starter programme to get you off zero and into using this stuff for real outcomes.
Most of what you'll read about AI for non-technical business owners this year was written by someone with a developer in the next room. They've got a Notion template, three tools they're affiliated with, and a hot take about how easy this all is. The hot take is wrong on both ends. It's easier than they're making it sound for the things you actually need, and it's harder than they're making it sound for the things they're pretending you can do in a weekend.
I'm not a developer. I'm not a prompt engineer. I'm not someone who learned this stuff at a bootcamp. I've been a marketing consultant for twenty-one years. My clients have included IBM, Twitter, Dropbox, monday.com and Greenpeace. I run a newsletter that sits at 15,000 subscribers with a 70 per cent open rate. I went all in on AI in 2024 because I had to. The reason I had to is not because of FOMO. The reason I had to is that the cost of producing marketing work without AI overtook the revenue I could earn from it, and AI was the only thing in the market that could close the gap.
That's the honest origin story for the non-technical, mid-career business owner. It's not about innovation. It's about not getting eaten by the cost curve.
By the end of this blog you'll know exactly which AI tools I use, what I use each for, the ten specific workflows that have changed my business in eighteen months, the things AI is genuinely terrible at, the seven-day starter programme if you're at zero, and when to hire help versus when to do it yourself.
Last reviewed: May 2026 · Lilach Bullock
TL;DR
You do not need to be technical to use AI in your business in 2026. You do need to be willing to try things, fail repeatedly, and accept that the first month will feel slow before it feels fast.
The mistake non-technical founders make is buying twelve tools, learning none of them, and concluding AI doesn't work. The fix is to pick one tool, run one workflow on it every working day for two weeks, then add the next workflow. Compound learning beats every shortcut.
The myth keeping you stuck
The most damaging myth in AI for non-technical business owners in 2026 is "I need to learn to code or I need to hire an engineer."
You don't, on either count. The current generation of AI tools, integration platforms, and no-code workflow builders means a non-technical founder can ship genuinely useful AI workflows in their own business without writing a line of code or hiring an engineer.
I'm proof. I built the custom WordPress publishing pipeline, the SEO recovery tooling, the sales call prep automation, and the inbox triage system on this site — all on weekends, all in plain English, all without an engineer. The tooling has matured to the point that "non-technical" is no longer a useful gate.
What it does require: willingness to learn the basic patterns (what an integration is, what a webhook is, how AI fits into a workflow), patience for the first few iterations to work properly, and a clear view of what you're trying to achieve.
My actual AI stack in 2026
I'll tell you what I actually use, and what I use each one for. Not a list of the twenty most popular tools. The ones in my actual rotation, this week, this month.
HubSpot (already in my stack). Most of the AI features I rely on are inside the CRM I've been using for years. Draft email replies. Auto-summarise contact activity. Auto-suggest follow-up timing. None of this requires me to learn a new tool. I just turned on the AI features that were already in the product.
Kit (already in my stack). My newsletter platform. AI subject line suggestions. AI segment naming. AI drafting of welcome sequences. Same logic as HubSpot: it was already in the stack, I just turned on the AI features.
Notion (already in my stack). Notion AI for internal notes, summaries, the meeting transcripts I never want to read but need to scan. Mid-tier useful. Wouldn't pay for it standalone if I weren't already using Notion for other things.
You do not need more than this. You really do not. Every non-technical founder I've watched fail at AI implementation has too many tools, not too few.
The ten things AI is actually doing for me in 2026
These are real. They are the daily workflows. Specific, with timing.
One. Newsletter drafting. I write a 1,200-word newsletter every Sunday. The first draft is done by Claude in about ten minutes, based on a brief I write in about five. I then spend ninety minutes editing for voice, accuracy, and the specific bits that need to be exactly right. Total: ~2 hours. Pre-AI: 4 to 5 hours.
Two. Blog refresh. I have roughly two thousand blog posts on my site, written over twenty-one years. Most are out of date. I refresh roughly four posts a week using Claude to draft updates based on my brief, then editing for accuracy and voice. Pre-AI this was unrealistic at the volume I needed. With AI, it's part of my weekly rhythm.
Three. Client research before discovery calls. Perplexity, fifteen minutes per call. I get a current view of the prospective client's company, recent news, leadership, and likely problems before I sit down with them. Pre-AI this was forty-five minutes of LinkedIn scrolling per call. Calls I have per week: roughly six. Time saved per week: three hours.
Four. Sales email drafting. Either ChatGPT or Claude for the first draft, then voice-edited by me. The trick here is to give it the previous five emails in the thread so it understands context. Time per email: five minutes. Pre-AI: fifteen minutes.
Five. Podcast and stage prep. I speak at events and on podcasts roughly twice a month. Granola transcribes the prep call with the producer. ChatGPT summarises into the topics I need to cover. Notion holds the speaker prep doc. Total prep time per appearance: thirty minutes. Pre-AI: two hours.
Six. Inbox triage. Gmail's built-in AI suggestions are good enough now that I can clear an inbox of fifty emails in twenty minutes. Pre-AI: an hour.
Seven. Calendar and meeting summaries. Granola records, transcribes, and summarises every meeting I take. I can find a quote, a decision, a next step from any meeting in the last six months in under thirty seconds. Pre-AI this was lost information.
Eight. Customer service drafting. When a client emails with a complex question, I draft the response with Claude, edit for voice, and send. Time: ten minutes. Pre-AI for a complex client email: forty minutes of drafting and second-guessing.
Nine. Pricing and scenario modelling. I'm not a finance person. I use Claude to walk through pricing scenarios when I'm thinking about a new offer. Not as a calculator. As a thinking partner. This is the one most non-technical founders miss because it doesn't look like AI work. It is.
Ten. Internal documentation. Anything I do twice, I document. AI does the first draft of the documentation based on my workflow. I edit. Documents I'd have never written exist now.
The throughline across all ten: AI does the boring first draft. I do the judgement, the editing, and the bits that need to sound like me. The combination is faster than me alone, and the output is higher quality than the AI alone. Neither of those statements would be true if I'd tried to use AI without twenty-one years of marketing context.
What AI is genuinely terrible at right now
This list matters because the bro-marketer crowd will tell you AI is good at everything. It isn't.
One. True strategic judgement. Deciding what the business should do next year. Deciding whether to launch the new product. Deciding which client to fire. AI can structure your thinking, but it cannot make the call. The call is yours.
Two. Voice mimicry beyond about three hundred words. AI can sound like you for a paragraph. It cannot sound like you for an entire essay. The longer the writing, the more the seams show.
Three. Original facts about your specific situation. AI will invent numbers about your business if you let it. It will hallucinate citations. It will give you a confidence level that doesn't match its accuracy. You have to verify any fact that matters.
Four. Anything emotional in client relationships. Apologies that need to land. Difficult conversations about money. Conversations with clients who are upset. Write these yourself.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
Five. Designing the workflow itself. AI cannot tell you which problem in your business to solve next. It cannot prioritise across competing claims on your time. It cannot see your business well enough to design its own integration. You have to know what you want.
Six. The first hour of any project where you're inventing the brief. AI gets faster and faster as the brief gets clearer. Until the brief is clear, AI is no faster than you working alone. The trick is to spend more time on the brief, not on the output.
Seven. Anything that needs to be exactly true and is checkable. Specific dates. Specific names. Specific numbers. Citations. Check everything.
The seven-day starter programme if you're at zero
This is the version I'd give my own mother if she were starting from a standing start tomorrow.
Day one. Pick one tool. Claude or ChatGPT. Sign up for the paid plan. Yes, the paid plan. The free tier is a loss leader designed to make you upgrade. Just start with the paid plan and skip the wall.
Day two. Pick one workflow. The first one should be small. A sales email reply. A blog headline. A first draft of a social media post. Use the tool to do it. Read what came back. Edit it until it sounds like you. Send.
Day three. Same workflow. Same tool. Different real example from your real work. Read. Edit. Send. Notice how fast or slow it felt this time compared to yesterday.
Day four. Same workflow. Same tool. Different example. By day four you will start noticing patterns in what the tool gets right and what it gets wrong. Write down two things you've noticed.
Day five. Same workflow. Same tool. Different example. The output should be starting to feel useful. If it isn't, the issue is the brief, not the tool. Spend five more minutes on your brief next time.
Day six. Same workflow. Same tool. Different example. By now the workflow should feel faster than doing it without AI.
Day seven. Step back. Track how many minutes it took you for this workflow today versus the equivalent two weeks ago. Even if you didn't measure exactly, you'll know.
Week two. Add a second workflow. Do not add a second tool. Just add a second job-type for the same tool.
Month two. Add a third workflow. Maybe try a second tool if you've outgrown the first for one specific use case. Maybe not.
Month three. You will have three workflows running and you will be saving roughly two to three hours per workday. From here you compound.
The thing every non-technical founder gets wrong is they try to do all of this in week one. They sign up for five tools, watch ten videos, take three courses, build nothing, and conclude AI is not for them. The path that works is slower at the start and dramatically faster at the end.
When to hire help versus when to DIY
I'll be straight with you because I have a vested interest in this question and I want to be honest about it. I am an AI implementation consultant. I make money when small business owners hire someone like me. I'm going to tell you when not to.
Do it yourself if: You have five to ten hours a week to learn and experiment over the next three months. You are happy reading documentation. You have someone on your team who is technical enough to handle integrations if a workflow needs to talk to your CRM or email tool. Your business is straightforward enough that a few standard workflows would cover most of your time savings.
Hire help if: You don't have ten hours a week to spare. You need this implemented in eight to twelve weeks. You want someone else to figure out which workflows are worth building first. You want to skip the part where you waste three months learning the wrong tools. You have a team that needs to be trained, not just yourself.
The thing I will not tell you to do is hire the cheapest consultant you can find. The cost of getting this wrong is much higher than the cost of paying a senior consultant once. The cost of doing it slowly yourself is much lower than the cost of paying a junior consultant.
FAQ
What's the difference between ChatGPT and Claude in practice? ChatGPT is faster, more eager, slightly less accurate. Claude is slower, more careful, slightly better at long writing. For most non-technical founders, ChatGPT is the right first tool. Claude is the right second tool, six months in, when you start caring about voice.
How long until I see ROI? Eight to twelve weeks for measurable time savings on one workflow. Six months for ROI across multiple workflows. Twelve months for the AI-assisted version of your business to feel normal.
Can AI replace my marketing person? No. It will let one marketing person do the work of two. It will let you keep your existing marketing person at the same headcount while doing twice the volume. Those are very different things from replacement.
What's the biggest mistake non-technical founders make with AI? Trying to do too many tools at once. Pick one. Use it daily for two months. Add the next only when you've outgrown the first for a specific use case.
Is the AI bubble going to burst? The bubble in AI tooling investment will probably correct. The underlying utility of AI in marketing and business operations is real and is not going anywhere. Build your workflows on the assumption that the tools you're using today might cost twice as much in two years, but the workflow is still worth it.
Will my customers know I used AI? Possibly. If the work sounds like the AI default voice, yes. If it sounds like you, no. The edit step is what makes the difference. Spend the time on the edit.
The thing to take away
You do not need to be technical to use AI in your business. You need to be willing to start small, do one workflow daily for two weeks, then add the next. You need to be willing to pay for the paid tier of one tool instead of using six free tools badly. You need to be willing to spend the editing time the AI doesn't do.
If you can do all of that, you will be using AI better than ninety per cent of "AI consultants" inside six months.
If you can't, hire a senior AI implementation consultant who has shipped this in their own business first.
If you want to talk about which path is right for your business, the discovery call is twenty minutes and free.
Either way you'll be further along than the version of you reading this who hasn't started yet.
About Lilach Bullock
I'm Lilach Bullock, an AI implementation consultant and fractional CMO based in the United Kingdom. I've been a marketing consultant for twenty-one years. In 2024 I went all in on AI and rebuilt my consultancy around it. I now help founders and marketing leaders implement AI workflows that move business metrics, not just tool stacks.
Recognition includes: Forbes Top 20 Social Media Power Influencer (twice listed), Oracle Social Influencer of Europe, Number One Digital Marketing Influencer in the UK (Career Experts), Best Mumpreneur of the Year (Downing Street recognition), Global Women Champions Award. I've spoken at over 100 events worldwide and run a weekly newsletter with 15,000+ subscribers.
Connect: LinkedIn · Newsletter · Get in touch · Wikidata
Related reading
- What is AI implementation?
- How to evaluate an AI consultant
- AI marketing consultant: the 2026 hiring guide
- AI for service businesses
Related: deeper reading
More on AI implementation and consulting
Related reading on AI consulting
- /ai-consultants-uk-small-business/ — AI consultants for UK small businesses — UK-specific framing
- /fractional-cmo-ai/ — Fractional CMO + AI ops — the embedded senior role
- /ai-marketing-consultant-usa/ — AI marketing consultant — USA-focused page
- /ai-marketing-consultant-israel/ — AI marketing consultant — Israel-focused page
- /ai-marketing-consultant-canada/ — AI marketing consultant — Canada-focused page
- /ai-marketing-consultant-australia/ — AI marketing consultant — Australia-focused page
- /how-much-does-an-ai-consultant-cost-2026/ — How much an AI consultant costs in 2026 — engagement scopes explained
