Consultancy is the older half of Droid Minds and still the half that pays for the lab. Engagements are scoped in weeks, start with the part of your operation that hurts most, and end with something your team can run without us.
Assistants that live where your customers and staff already are — WhatsApp, web chat, your internal tools — trained on your documents, prices and policies. They qualify enquiries, answer the repeat questions, and hand the awkward ones to a human with context attached.
3–6 weeksFixed scopeHuman handoff built in
You end up with
—
A working agent in one channel, live with real customers
—
A knowledge source you can edit without calling us
—
Escalation rules and a transcript trail for every conversation
AI automation consulting
A week inside the work as it actually happens, then a short list: what to automate now, what to leave alone, and what it is worth. We build the first one so you can judge the rest on evidence.
Two-week assessment
Data & integration work
Most AI projects stall on plumbing. We clean and connect what you have — spreadsheets, accounting, bookings, a CRM someone set up in 2019 — so one system stops being why the others cannot be trusted.
Scoped per system
Web & app development
Internal tools, client portals and customer-facing products, built deliberately small so a team of two can keep running them after we leave — the same discipline we hold on our own products.
Project or retainer
Staff AI training
Half-day, hands-on sessions using your own documents and real cases — writing, summarising, quoting, research — plus a plain policy on what should never go into a chatbot. People leave with three habits, not a deck.
Up to 15 people
Every engagement ends the same way: you can run it without us.
01
A call, free
Forty minutes on what is slow and what it costs. If AI is the wrong answer, we say so and point you somewhere better.
02
A written scope
One page: the outcome, the fixed price, the date, and exactly what we need from your side.
03
Build in the open
Weekly demos while we build, so scope changes happen in week two instead of at the deadline.
04
Handover
Training for the people who will use it, documentation for whoever maintains it, and a month of support.
Engagement models
Three ways to buy the work
Every model is quoted after the free call — we would rather price the actual problem than publish a number that fits nobody.
Assessment
Two weeks, one answer
We shadow the work, quantify the cost, and hand you a prioritised plan with prices. Credited against the build if you continue.
We have done useful work for two-person firms. If the problem is worth a few thousand dollars of your time a year, there is usually something worth automating. If there is not, the free call ends with us saying that.
Do we need clean data first?+
No. Nobody has clean data. Cleaning and connecting what exists is usually the first third of the project, and we price it as part of the work rather than as a prerequisite you have to solve alone.
Who owns what you build?+
You do — code, prompts, configuration and documentation, in your accounts wherever possible. We run on Canadian hosting by default and your data is never used to train shared models.
What if AI is the wrong tool?+
Then we say so. A good share of our work is ordinary integration and process fixing that happens to be far cheaper and more reliable than a model. We would rather be recommended than booked.
Tell us what is slow
One email describing the messy part of your week is enough to start. We answer with questions, not a proposal template.