Back after a pause. Sanity check on manual offer before building

[Shougan]

Context​

It’s been a while since I last posted here. Earlier this year I shared weekly updates around an AI concept for manufacturing, but I went quiet for about 2–3 months.

That pause wasn’t inactivity. I’ve been focused on execution and decided to postpone posting updates on the forum while things evolved.

Why the direction changed​

The work I did around AI for manufacturing is what brought me in contact with two developers. One thing led to another, and we decided to work together.

As we explored ideas together, it became clear that the path we were collectively most motivated to pursue was slightly different from the original manufacturing angle. Rather than force alignment on a direction that wasn’t there, we chose to collaborate on a new path while keeping the same core approach: talk to customers first, validate manually, then build.

What we are actually exploring​

At a high level, we’re working on an AI-driven solution aimed at reducing manual, repetitive work for real estate agents in a service-heavy workflow. The focus is not “AI for AI’s sake”, but helping agents spend less time producing materials and more time on high-value activities like client interaction and closing deals.

The current hypothesis is centered around speed and consistency of output rather than deep automation from day one. In other words: help agents get from raw input to usable output faster, even if parts of the process are manual initially.

Key learning so far​

Over the last month I’ve had roughly 10 direct conversations (mostly phone calls) with potential customers. The explicit goal was to understand their workflows end‑to‑end, including where time, energy, and attention are currently being wasted.

All of them were positive toward the overall idea and agreed to see a demo once we have something concrete.

Signals from the market​

During these conversations, one recurring signal stood out:

• 6+ people explicitly called out content creation for social media as a real pain
• They described it as time‑consuming, cognitively heavy, and increasingly important
• One person literally said: “I would pay for this.”

This is not the core focus right now, but rather a feature or capability we’re considering further down the roadmap. That said, the signal is strong enough that it’s worth keeping on the radar as we shape the initial offer.

Where I’m at right now​

At the same time:

• I’m working with two developers who are building in parallel
• I personally see strong short‑term potential in manually delivering a very narrow version of the solution to 3–5 of the people I’ve already spoken to

Proposed next step​

My current idea is to:

• Manually sell and deliver a simple pilot to 3–5 customers
• No full product, no heavy automation
• A clear promise, a clear output, and me doing part of the work manually

The intent is to:

• Validate willingness to pay
• Understand what actually creates value
• Reduce the risk of building the wrong thing

The question​

From a Fastlane perspective:

Is this the right move, or am I at risk of falling into a consulting trap that I should avoid at this stage?

If you were in my position, what would you optimize for right now?

Any feedback would be greatly appreciated.

 

Dislike ads? Become a Fastlane member: Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)

[Jaywalker]

Uwis vji metv nupvj O’wi jef suahjmz 10 fosidv dupwistevoupt (nutvmz qjupi demmt) xovj quvipvoem datvunist.

Aq vjot pancis vu toyvz dupwistevoupt op e fez.

Et e sigisipdi quopv, nz uggodi timmt qjupi qmept gus EV epf V op vji goimf. Vji csepf jet ciip esuapf topdi 1867. Velit vjev enuapv ug uavsiedj vu hiv ipuahj dupwistoupt vu qez vji commt iwip xovj vji dusqusevi neslivoph epf ep optepi uggis.

 

[Jobless]

Ximduni cedl,

Ov't hsiev og tuniupi xepvt vu qez, cav qsigisecmz ov tjuamf ci tactvepvoem. Iyenqmi, optvief ug timmoph vu e siem itvevi ehipv, zua timm vu vji siem itvevi ehipdz us e siem itvevi nesliv qmedi.

Zit zua piif vu veml vu nusi vjep 10 qiuqmi. Neli e finu epf tjux ov vu vjin. Ov fuitp'v jewi vu xusl. Ov dep ci e geli tiswodi gus fotqmez qasqutit upmz. Etl gus nusi nupiz vjep zua vjopl ot sietupecmi. Og vjiz caz, zua dep caomf ov raodlmz.

Emtu djidl xovj DIPVT nufim. Xjev esi vji IPVSZ epf DUPVSUM etqidvt op zuas gewus?