Workflow 27: How to Get Clients Using AI: Six Practical Ways
Part of Scale You Workflows: 31 real workflows from my business, so it runs without you. Prompts included. Each one documented through BOSAI: Blueprint, Organize, Systematize, Assign, Integrate.
Six practical ways AI helps you get clients, from validating your niche to drafting follow-ups, and the point where a person has to take over.
- Time to build
- Varies
- Difficulty
- Beginner
- Tools
- Claude, ChatGPT, Gmail, Notion, LinkedIn
Everyone says you can get clients with AI. Almost nobody says how.
Here are six ways that actually work, with the caveat that your business might need a different mix than mine.
What this replaces
Waiting to be found.
The usual version of AI lead generation is either vague ("use AI to find leads!") or grim: scraping emails and blasting strangers. Neither is what I do. What works is using AI on the parts of client-getting that are genuinely mechanical, and keeping the parts that aren't.
Blueprint
Six places AI helps.
Trigger: manual, applied to whichever part of the problem you're stuck on. Inputs: your niche, your content, your existing list, your profile, your experience. Output: better targeting, more content, consistent follow-up, a clearer profile.
- Narrowing your niche, and checking it's real
- Making content
- Following up on a list you already have
- Optimizing your profile
- Finding roles or clients worth applying to
- Researching a prospect before you speak to them
1. Narrow the niche, then validate it
Most people try to serve everyone. Serve everyone and you serve no one.
Use AI to research who you're thinking of targeting, then check whether they actually need what you're selling. That second half is the part people skip.
Part of that check is looking at competitors. You don't want to be the only person in a space. Competitors are proof that people want this, that they want it long term, and they tell you what kind of people buy it.
2. Make the content
This is how I've gotten clients for seven years. Video, articles, podcasts, all of it.
AI helps me at three points: brainstorming ideas, checking those ideas are things people actually want to click on, and managing the calendar. Mine writes straight into Notion, so I open it and see what I'm filming.
3. Follow up on the list you have
I don't scrape. If you already have a list of people you want to work with, that's different.
Both Claude and ChatGPT connect to Gmail, so AI can draft your outreach, then check daily who replied and who needs a follow-up. Mine drafts a follow-up every three days and stops after the third. I read every one before it goes out.
What changes is that you stop overthinking. A draft is already waiting, so you read it and send it.
4. Optimize the profile
I use AI on my LinkedIn constantly, mostly for an outside view.
The goal is that when someone lands on your profile, they understand what your business does and what problem you solve. If the messaging is unclear, nothing else works.
5. Find the right roles or clients
Upload your resume with the contact details stripped out, then have it search job sites and find the roles that genuinely match your experience.
You end up applying to people who are already looking for you, instead of going page by page.
6. Research before you speak
Once someone replies, research them before the call. Who they are, what they do, and whether they're actually a fit for you.
Organize
What you need before any of this helps:
- A clear sense of who you serve, or a willingness to narrow it
- Your existing list, if you have one
- Gmail connected, for the follow-up piece
- Your resume or profile, whichever applies
- Somewhere for content plans to live
Systematize
Niche validation:
I'm considering targeting [NICHE] with [YOUR SERVICE].
Research and tell me:
1. Does this group have the problem I'm solving? What's the evidence?
2. Do they have budget for it, and do they currently pay for anything similar?
3. Who else serves them? For each competitor: how long they've been around,
whether their reviews are good, whether customers return.
4. What language do they use to describe this problem themselves?
If nobody else serves this group, treat that as a warning and tell me why
it might be empty.
Be honest if the evidence is thin. I'd rather hear that now.
Follow-up management:
You manage my outreach follow-ups. Run daily.
My list: [WHERE IT LIVES]
My rules:
- Follow up every 3 days
- Stop after the third follow-up
- Never contact anyone who hasn't been added to the list by me
Each day:
1. Check which of my sent emails got a reply. Mark those done.
2. Find anyone due a follow-up under the rules above.
3. Draft the follow-up. Short. Reference the previous message. Save as a
draft, never send.
4. Tell me who's aged out after three follow-ups so I can drop them.
Never add anyone to the list yourself. Never scrape or source contacts.
Profile review:
Read my [LINKEDIN PROFILE / RESUME], pasted below.
Answer as someone landing on it for the first time:
1. What does this person do, in one sentence?
2. What problem do they solve, and for whom?
3. What's unclear or takes too long to work out?
4. What would make someone reach out?
Be blunt. Then suggest rewrites for the weakest parts, in my voice.
How to set it up
- Write down who you think you serve, then run the niche validation prompt before committing to it.
- Check the competitors it finds. If nobody serves this group, treat that as a warning.
- Pick one content channel and use AI to plan a month of posts for that niche.
- Build your outreach list by hand, from people you already know or already want to work with.
- Connect Gmail and run the follow-up prompt daily. Read every draft before it goes out.
- Run the profile review and fix the weakest parts of your profile first.
Assign
AI does the research, the drafting, the checking, and the chasing.
Every actual relationship is yours.
The line falls in a specific place, and it's the reason this list doesn't include scraping and mass sending. AI is excellent at working out who you should be talking to. It's terrible at being the person they hear from.
The follow-up workflow above only ever drafts. Someone reads each one before it goes. That's not a formality: the first message a potential client gets from you shouldn't be the first draft of anything.
The niche work has a similar catch. It'll tell you what the research says. It can't tell you whether you want to spend the next three years serving those people, which is a question only you can answer and the one that actually determines whether it works.
Rule of thumb from this build: clients come from conversations, so protect the conversation.
Worth adding: what has actually brought me clients for seven years is content. Showing up and explaining what I know, week after week. AI makes that faster. The showing up is still mine.
Integrate
These six connect to workflows elsewhere in the library:
- Niche validation uses the same competitor research as my idea filter
- Content planning and repurposing already run as their own workflows
- Prospect research is a whole agent in my stack
- Follow-ups land in the same inbox my email agent already sorts
All of it depends on one thing none of these workflows can create: someone who actually wants what you're selling. Everything above gets you in front of them faster.