Learn · Articles
AI in Practice #0 — How AI Actually Helps Run a BSF Case in an Afternoon Instead of a Month
On Monday I said most BSF cases are killed by 5 non-biology surprises.
00 · Overview
How AI helps in a BSF case
Today: how AI actually helps me move from "we have biowaste" to a yes-or-no answer in an afternoon instead of a month.
Not a pitch. Not "AI will transform BSF". The boring, useful things that save hours every week in a small BSF company.
01 · Price picture
1. A country-specific feed and protein price picture in 10 minutes.
Every conversation starts with "what do local soy meal, fishmeal, and corn cost this quarter?" That used to be half a day across FAO tables, USDA reports, local industry posts. Now it's 10 minutes with sources cited. The AI isn't being smart — it's just fast at pulling together data I already trust how to read.
02 · CAPEX check
2. Sanity-checking a customer's CAPEX against reality.
A prospect sends a 5 t/day budget. Instead of arguing from memory, I compare it line by line in minutes against what similar facilities have actually cost. Usually: "your number is 30–50% off in this line, here's why." Not AI being smart — 8 years of our own expertise made queryable.
03 · Feedstock data
3. Turning messy waste data into a clean feedstock profile.
Customers send waste data in every shape — spreadsheets, photos, three emails, a PDF. Turning it into a clean stream profile (volume, moisture, seasonality, contamination) used to be the worst part of my week. Now it's the first 15 minutes.
04 · Sourcing
4. A sourcing shortlist for a new country in an afternoon.
When a customer wants a facility in a country we haven't built in yet, the question is "what equipment is available locally, at what lead time?" Researching suppliers, pulling quotes, cross-checking against our BOM, flagging imports — used to take weeks. Now it's an afternoon.
05 · Business case
5. A first-pass business case the customer can react to.
The worst way to start a customer relationship is with a blank page. Now I produce a first-pass case — feedstock, revenue, CAPEX, OPEX, break-even at realistic local values — inside an hour. The customer doesn't get a perfect answer. They get something concrete to push back on. That moves the conversation forward.
06 · Industry watch
6. Keeping up with the industry, weekly.
BSF is a small industry with loud headlines. A structured weekly review keeps me current in about an hour instead of ten.
07 · One to four
7. Turning one insight into four formats.
A LinkedIn post, a blog post, a sales-deck slide, an investor update — same numbers, different audiences. A small team can only do this with AI helping.
08 · The real edge
Where AI stops and knowledge starts
None of this replaces field experience or judgement calls. It replaces the slow, repetitive knowledge work that used to take 80% of a small-company day.
The real edge isn't AI. It's knowledge — 8 years of BSF cases across 50+ customers on 6 continents — made queryable and actionable. Anyone can buy an LLM subscription. Almost nobody has the knowledge base to point one at.
Running a small company in an expertise-heavy field and trying to figure out what "AI in daily work" looks like? DMs open — happy to compare notes.
Keep exploring
Related topics and pages
- BSF business course — where a customer's numbers and markets shape the business case
- BSF start-up course — the validation steps behind a first-pass case
- Articles library — the written notes this post joins
Read next
- AI-Enhanced Workflow at Manna — the practical workflow behind the AI help described here
- Your AI Agent Has Amnesia. Here's How We Fixed It at Manna. — the shared memory that keeps a fresh agent consistent
- What shows up first in a real biowaste case (it's almost never the biology) — the five non-biology surprises that decide a BSF case first
- BSF Bible teaser: the transparency gap — why honest numbers, not vendor pitches, are what a BSF case still lacks
Next step
Bring the case to Manna
Manna can help organise the material, prices and records behind a BSF case so the next answer rests on written, shared context.