NavigateMENU / 04
More
Elsewhere
Paper Crane / services / ai-consulting
We help organizations work out where AI actually fits, train the people who will use it, and build the automation underneath. Most clients come to us after a pilot or a chatbot that never changed how anyone works.
( Scroll to explore ↓ )
Note
Where AI projects usually stall
The common failure is not picking the wrong model. It is that the work never reaches operations. A chatbot goes on the website, a licence gets bought for every seat, and six months later the way people do their jobs is identical.
Meanwhile there is real usage happening that nobody has counted. Someone in finance is paying for a tool on a personal card. Someone in communications has built four things their colleagues have never seen. The gap inside most organizations is wider than the gap between them.
We work on the operations end of this. That means measuring what is already happening, deciding what is worth building, teaching the people who will use it, and building the parts that need building.
What we do
01
Strategy and readiness
A measured picture of where your organization stands, an opportunity map filtered for what your systems and data can actually support, and a roadmap in a sensible order.
02
Team enablement
Training split by fluency so nobody sits through basics they mastered a year ago, built around tools we make for the work your team already does.
03
Automation and agents
Pipelines, agents and harnesses aimed at work that is too large, too repetitive or too constant for people to keep up with.
04
Integration into what you run
Models wired into the systems you already have, grounded in your own data, rather than another tab your team has to remember to open.
05
Multi-model systems
Routing across providers where it makes sense, so you are not rebuilding everything the next time the pricing or the capability changes.
06
Governance and policy
Use policy, data handling, an approved tool list, and answers to what your board and your funders are going to ask.
01Project example
A rebuild, then the training that followed
We rebuilt ERP-One's site so a non-technical team could update it by talking to Claude. Then we ran working sessions on Skills, and their team started building research and context tools inside their own cloud environment.
FeaturedERP-One
Who we work with
Organizations with real operations
Nonprofits, associations, government and mid-sized companies where the work is done by people whose jobs are not going to be replaced by a chatbot.
Teams with nobody internal to own this
No head of AI, no data team, and a board that has started asking questions. This is the most common situation we walk into.
People who have already tried something
A pilot that stalled, a licence nobody uses, or a tool that got built and then quietly abandoned. Knowing what did not work is a useful place to start.
Common questions
Do we need to be on a particular AI platform?
No. Our strategy work assesses whatever you are running and recommends honestly, including telling you to stay where you are. Our deepest hands-on work is in the Claude ecosystem, and we will say so when it is relevant.
Is this training, or is it building?
Both, in different engagements. Training without tools does not stick, and tools without training do not get used. Which one you need first depends on what the assessment finds.
Will this replace jobs?
It is not what we are hired for and it is not work we take. Most of our clients are trying to do more with the team they have, not fewer people with the same output.
What happens to our data?
That is part of the governance work: what leaves your organization, where it lands, what the vendor retains, and what PIPEDA or FOIP mean for it. For public sector and health clients this is usually the first question, and it should be.
Do we own what you build?
Yes, outright. Code and content are yours, on infrastructure you control. That is true of everything we build, not just the AI work.
How long does an engagement take?
A strategy engagement is three to four weeks. Enablement runs four to six. Automation work depends entirely on what is being automated, and we will size it before you commit.
Next step
Start with a conversation
Tell us what is happening in your organization now, including the parts that are not working. We will tell you honestly where we would start, and whether it needs us at all.