Practical solutions you own

Custom workflow solutions for small businesses.

Most SMB software problems are hiding between systems. Botworks helps decide what to teach, automate, configure, or build.

Not everything needs custom software.

Sometimes the right answer is better AI use, a small automation, or cleaning up the process you already have.

When custom is right, you keep it.

The code, data, infrastructure, and documentation stay handoff-ready instead of trapped inside an agency black box.

Let's talk See work

The starting point

Most owners already know the annoying part.

It is the thing that lives in texts, memory, screenshots, bad portals, duplicate spreadsheets, and the office manager's head.

The useful question is not “can AI do this?” It is “what should we fix first, should we build or buy, and what will the business own when we are done?”

Ask your own AI

Don't trust the pitch. Test the fit.

Copy this into ChatGPT, Claude, Gemini, or whatever you already use. Add a few details about your business and turn the vague “should we do something with AI?” hunch into a concrete first-call brief.

Second-opinion prompt

paste into any LLM

I run or help operate this business:

[Describe the company, team size, customers, and the operational workflow that feels messy.]

I am looking at Botworks Agency: https://botworksagency.com

Botworks says it helps SMB operators figure out where AI is actually useful, where existing tools or better employee workflows are enough, where automations between systems can remove copy/paste work, and where custom owned software is worth building. They emphasize no slide decks, fast working prototypes, operational judgment, monitoring/documentation, and client ownership of code, data, infrastructure, and docs.

Please act as an operator-advisor, not a vendor or generic AI strategist. Keep the answer practical and under 900 words. Do not write a broad "AI transformation" memo. Do not include citations unless you actually browsed current sources.

Based on my business context:

1. What are the 3-5 workflows Botworks should inspect first?
2. For each workflow, is the likely answer existing AI tools, automation between systems, custom software, or "do not build yet"?
3. Which one workflow is the best small first project, and why?
4. What would make this a bad fit?

Notes

Thinking in public about practical AI and owned workflows.

Comparisons, technical readouts, and field notes from the places where small-business operations meet software.

Read notes

How we help

AI workflow review

Find where AI belongs.

Help employees use Claude, Codex, or the tools they already have better. Just as important: call out the places AI should not touch.

Automations between systems

Stop paying people to copy and chase.

Move data, documents, updates, and decisions between tools without copy-paste.

Custom owned software

Build what is too specific to rent.

Build the workflows that are too specific or too important to leave in spreadsheets, bad portals, inboxes, or memory.

Who this is for

Searchers and acquisition-backed SMBs

A practical AI pass for the 100-day plan.

Find the workflows where AI is useful now, where the team just needs better habits, and where custom software is worth building.

See how it works →

Service businesses

Office and field workflows that still run on handoffs.

HVAC, landscaping, cleaning, trades, and field-service teams where work gets copied, chased, retyped, and remembered by the person who cannot go on vacation.

See examples →

Where this fits

The best first projects are weirdly ordinary.

Rebate paperwork, field handoffs, customer intake, spreadsheet reconciliation, estimate follow-up, permits, warranty packets, status chasing, exception queues, manager-only memory. The first question is not “what should we build?” It is “where is the work leaking time?”

rebatespermitsbad portalsinbox handoffsone-person memory

See workflow examples →

Case studies

Real work, in production

HVAC contractor

Rebate program filings: 8 hours a week reclaimed.

The office manager used to spend hours moving rebate data between bad systems. Now she pastes a job number and gets a mail-ready packet.

Read the case study →

Commercial landscaping

Field-visit capture: a structured record for every property visit.

Account managers needed one clean place to capture what they saw in the field, then turn it into office review, translation, follow-up, and crew tasks.

Read the case study →

Interior design studio

AI-assisted research: better prep without a big custom app.

Practical AI habits and lightweight tools that compress a designer's pre-pitch homework — reference photos, listing-price sanity checks, neighborhood comps — from hours to minutes.

Read the case study →

What's next

In the pipeline

Chicago HVAC contractor

Upcoming

Front-office automation: one view of the day's work.

Technicians ping Slack to grab their next job; Housecall Pro APIs feed a single source of truth for upcoming work; office and field see the same view in real time. Foundation for agent-driven inventory and scheduling decisions later.

Hi, I'm Matt.

15 years as a technical product manager in software, most recently VP of Product Management. I shipped to engineering teams that ship every day; now I bring that bar to small businesses that don't have engineering teams of their own.

I was searching in Miami and realized my skills were probably a better fit for an agency: sitting with owners, finding the operational mess, and turning the right piece of it into something that works.

If it is custom software, the client owns all of it. I am also building the handoff so employees and LLM agents can keep running it if we ever part ways. No hidden fees. No hostage software.

Have a workflow that should be easier?

Tell me where the work gets copied, chased, retyped, or remembered by one person. I'll help sort whether the answer is AI, automations, or custom software.

Let's talk