Your Biggest AI Automation Questions Answered (From 1000+ Comments)
Nick Saraev · 31 min of video, folded into 8 min of reading · about 22 minutes of real substance
Verdict
Watch from 17:47 if you sell automation services, since the intro offer to recurring service math is worth hearing him work through out loud. The other eight answers are short enough that this digest carries them.
The point
Nick Saraev works through questions from nine commenters about starting and scaling an AI automation business. His answers keep returning to one test: how much leverage or revenue does this move add, then what does it cost when it breaks. Beginners get pushed away from deep coding plus from automating anything that moves money, toward templated offers that lead into recurring contracts.
10 takeaways · 4 actions · 5 recall questions
Who's talking
Nick Saraev builds no-code automation systems for clients with tools like Make.com while answering viewer questions daily on a second YouTube channel. He sells Maker School, a 90-day accountability program he says has over 2,500 members, which is also why he holds back part of the sales-call answer.
Read with these in mind
- How much coding does someone need before they can sell automation services?
- Where does the money in an automation business come from once the first project ships?
- Which processes pay you back when you automate them, which ones quietly cost more than they save?
Do this
Takeaways
Template every client build. A custom $1,000 project means doing 100% of the work. The same $1,000 against a saved template means roughly 20% of the work, so a proposal system that took about 5 hours from scratch becomes 45 minutes of connections plus a 15 minute walkthrough video. 18:53
→ The leverage shows up on the second sale of the same system, so pick systems you can sell repeatedly.
Intro offer feeds recurring. The product line he recommends starts with a fast, low internal cost intro offer around $1,500, a proposal system for example, then converts those buyers into a recurring service. At $4,000 a month against a four month average churn, one recurring client is worth about 16K, close to ten times the intro offer. Between 10 and 30% of intro buyers convert. 21:04
→ Sell the cheap thing to earn the right to pitch the expensive one.
Finance automation risk math. Automating payroll for 10 staff twice a month saves roughly 40 minutes, about $15 of labour at $20 to $30 an hour. One misfire that adds a zero turns $1,000 into $10,000, so he puts the risk against the reward at 600x. 9:11
→ Weigh the downside of a broken run before you count the minutes saved.
Revenue beats cost savings. Lifting the margin of a $10,000 a month business by 5% puts 500 bucks a month in the owner's pocket. Lifting their topline by 20% puts $2,000 there, which is why he steers automation work toward revenue systems. 12:14
→ Pitch systems by the revenue they add rather than the hours they save.
Cold email compounds weekly. A 1.2% reply rate on 1.8K sent emails is a starting position he calls not terrible. Improving the campaign 10% a week for 12 weeks lands it near 3.8%. He says a single change can move a campaign 200%. The weekly loop costs 20 to 30 minutes. 13:08
→ Judge a campaign by its improvement rate across weeks.
Skip Python, learn JSON. His coding answer is 3 or 4 hours of a Code Academy JavaScript tutorial, then stopping. No-code builders work better with JavaScript or Node than with Python. The real bar is knowing what JSON is plus where the curly brace goes. 1:42
→ Spend the saved weeks getting the first few customers with Make.com instead.
SaaS after proven demand. He puts productizing an automation into SaaS after the service already makes $5,000 to $15,000 a month. Agency revenue climbs quickly early. SaaS takes longer to get going with steadier MRR-driven growth that eventually passes the agency, the path Instantly took from a managed B2B email service. 6:30
→ Use agency cash to fund the switch, expecting a revenue dip through it.
Two-step no-show sequence. Within 5 minutes of a booking he sends a message that reads as hastily typed, built in Make by watching the booking event and feeding the booking page answers to AI for a paraphrase. Two hours before the call he sends a second note saying he will be about 30 seconds late. 29:23
→ A small admission of fault makes a templated message read as personal.
Ideas from adjacent niches. For creator clients he names two systems beyond repurposing: a daily idea feed that scrapes YouTube videos or Instagram reels in a niche, summarizes them into a Google sheet, plus title generation from high performing templates. His own version scrapes reels from video editors or personal trainers who use the term AI, so he sees which tools land with domain experts outside his niche. 15:36
→ Look one niche over for content that will travel back to yours.
English before automation skills. For a beginner in India choosing between automation skills and communication, he points at his 110 step roadmap for the tooling, then tells him to block the first hour of every day for English training. 27:53
→ Fix the skill that gates every client conversation first.
How the video runs
- Does a beginner need to code?
Shruy has no coding background but wants to sell Make services and leans on templates for the code parts. Nick answers with the JavaScript position above, then pushes him back toward no-code work plus first customers.
- Sales calls, then two books
He gives the spine of the sales script he used at LeftClick: build rapport, set the schedule of the call, cover why me why now why this, cover a solution, send a proposal. He argues against hard closes for technical services, then recommends The Goal for business thinking plus Never Split the Difference for consultative questioning.
- Turning automations into SaaS
Phoenix asks whether productizing makes sense. Nick draws the two revenue curves on screen, then sets the revenue bar described above.
- Why he avoids finance automation
A Calgary viewer with a finance background asks what breaks when you automate accounting or operations. Nick works through the payroll numbers above, then lists the safer uses: bookkeeping, statements uploaded to Google Drive, pages parsed with pdf.co, then a ChatGPT analysis run against a template.
- 1.8K emails, 1.2% replies
Anasi got his first clients on Upwork but is stuck on cold email. Nick tells him to stay with the channel, points at the previous day's video for the writing process, then reframes the numbers as a starting position that compounds. He adds that monthly maintenance on a campaign is small, so running a second channel alongside it is possible.
- Copywriting or automation
Yousef is stuck choosing between the two. Nick picks automation on longevity grounds, giving human copywriters maybe a couple more years, per the quote above.
- Services for content creators
Sync system already sells long to short repurposing plus scheduling. Nick adds the idea feed and adjacent-niche scraping covered above, mentions using deep research for outlines, then flags titles and thumbnails as the highest leverage pieces where a human touch still matters.
- Packaging templates into an offer
The longest answer, close to ten minutes. Bite Bliss asks how to structure and price a templated service. Nick draws the leverage math, the intro offer to recurring funnel, then his own recurring package: a roadmap, weekly strategy calls, unlimited maintenance, a defined two hour Slack window, unlimited systems at a three to five business day turnaround that runs to about eight in practice. He closes with the alternative of building the whole recurring service around one offer, for example a lead gen service at 10,000 leads a month with weekly split tests plus reporting.
- English or automation first
Sanka from India knows JavaScript but rates his spoken English as just okay. Nick handles the tooling side quickly, then makes the communication case covered above.
- Cutting no-shows
Tommy already runs SMS and email reminders plus a 4 minute case study video before calls. Nick's addition is the timing and tone of the two messages described above, helped by booking calls within a week rather than far out.
Worth quoting
“Ask them enough questions, have them talk about their pain points enough, and they will sell themselves on your service.”
“you can only ever save 100% of your income, but you can make a quadrillion times your income.”
“this is the worst your campaign will ever perform. You are only going to get better from here on out.”
“We're getting to the point where pretty soon models are going to be able to copyright better than us”
“I don't think I've ever gotten a higher ROI from anything than improving my communication skills.”
Test yourself
Try to answer from memory before revealing. That's what makes it stick.
What weekly improvement rate does Nick apply to a 1.2% reply rate, over how many weeks?
10% a week compounded over the next 12 weeks, which he says brings the campaign to about 3.8%.
In the payroll example, what does the automation save per month against what a single misfire costs?
About 40 minutes, roughly $15 of labour, against a hypothetical $9,000 error from one extra zero. He calls that a 600x risk against reward.
How much of the work does a templated project take compared with building from scratch?
About 20% of the work for the same $1,000 fee, so roughly 5x leverage. His proposal system drops from about 5 hours to about 1 hour.
What are the two messages in Nick's no-show sequence, when does each go out?
A short note within 5 minutes of the booking that paraphrases what the prospect wants, written to look hasty. Then a message two hours before the call saying he will be about 30 seconds late.
What revenue does Nick want to see before an automation service becomes a SaaS?
$5,000 to $15,000 a month from selling and implementing the service, so demand is verified and there is some financial stability before the build.