AI build plan for founders and small business leaders

Feel behind on AI? You just can't see where it fits.

The pressure to "do something about AI" got you a stack of subscriptions and a team that's still stuck on ChatGPT. What you didn't get is a clear direction that makes you feel like you know how to actually implement AI. We'll show you exactly where AI fits in your business today and how far it can realistically take each function, with the costs on the table and an honest line on what should stay human.

Start free assessment → No card required · 10 minutes · Free sample report
Take your AI assessment
Which best describes your business?
Free · no cc requiredNext →
// the cost of having no plan

Most AI projects don't fail on the technology. They fail because tools get installed with no groundwork.

Generic AI advice everywhere. A plan built for you? Nowhere.

ERR_01
Generic playbooks
Advice written for any company, which means it fits none. Your stack, your team, your starting point never enter the picture.
ERR_02
Custom plans cost a fortune
Consultancies charge tens of thousands, then upsell the follow-ups. Ours is EUR 450.
ERR_03
Vendors sell tools, not outcomes
Everyone wants you to install their thing. Whether it fits your company is somebody else's problem.

You can't plan a route without knowing where you're standing. The assessment gives you that starting point in 10 minutes.

Read: 8 mistakes to avoid before buying your first AI tool →
Start free assessment →
// the honest starting point

A full assessment built around your company. Not just another whitepaper.

We've watched good companies buy nine tools, wire none of them together, and quietly stall at 'everyone uses AI now' without a single function actually running better. That's not a discipline problem. It's a missing map.

Treethree builds that map. We don't sell you a tool, and we're honest about what shouldn't be automated at all. The assessment is run by AI we trained on real implementation paths, and every recommendation is costed, sequenced, and gated, so you can see exactly what each step buys before you commit to it.

// the four AI modes

From a few AI tools to a team that runs on agents.

Mode 1most start here
Incremental
A few AI tools, used by hand. Nothing connects.
ChatGPT, Copilot
→ You're spending, but you can't point to what changed.
Mode 2
Assisted
AI drafts, a person finishes. Output gets repeatable.
Claude Code, Codex
→ The same work, in a fraction of the time.
Mode 3reachable
Augmented
Workflows run end to end; a person reviews.
n8n, evals + observability
→ Your team stops doing the work and starts checking it.
Mode 4
Autonomous
Agents run routine work and escalate the edge cases.
Custom agents, LLM swaps
→ The routine runs itself. People work on what only people can.
Read: the four AI modes in detail →
// the framework

Your implementation plan, built for your company, and only yours.

Five steps that calibrate on top of each other into an execution plan.

01
Landscape of Choices
The full map of what running on AI could look like for you.
Online Assessment - Free Sample
Start free assessment →
02
Direction Setting
A working session on your functions, team, and tools. We set the direction together. Not a form.
in-person
03
Path Deep Dives
Function by function, in payoff order.
in-person
04
Tailored Implementation Plan
Phased, costed, every step gated on the last.
in-person
05
Execution Cascade
The live build: an iterative plan that keeps evolving and adapts as your team builds and executes. Hand-holding only where you want it.
in-person
Read: the 5-step framework in detail →
// case study

A D2C brand went from ChatGPT to AI-first in five months. No new hires.

“We climbed from only chatbots to fully automated customer service, financial reporting, order management and a full content pipeline in five months. 15+ agents deployed, some already fully autonomous. We did it with the team we already had.”

0
new headcount
5 mo
M1 to M3
15+
agents live
// two ways this year goes

Same year. Two very different endings.

× without_a_plan
·Nine tools bought, a graveyard of half-used subscriptions.
·Everyone 'uses AI.' Nobody can say what changed.
·Spend climbs, the team is no faster, momentum stalls.
·Still stuck at M1, a year later.
✓ with the treethree framework
·One sequenced plan: what to automate, in what order.
·First function live in three months, the team supervising.
·Spend tied to gates, every step earns the next.
·M1 to M3 in five months, agents compounding.

The difference is the groundwork. Almost every plan skips it, so agents run on bad data, and that's why they fail. Our first recommendation is always the same: get your data clean and structured first (Month 1). From there, the first agent stands up with your team supervising (Month 3), and by Month 5 it runs with no hand on it.

See your own version, 10 min →
Read: why data is the groundwork →
// inside the report

We dissect your business, analyse each function, and show you the path.

A few are previewed free; the full report unlocks the rest for the price of one tool you'd have stopped using in three months anyway. A consultancy charges tens of thousands for less.

Executive summaryFree
The whole picture in one page: where the four modes could take you, the honest cost of each, and the first moves worth making.
Your AI vision: the four modesFree
A day in your company at each mode, where you stand today, and where you could realistically go from here.
Function by functionFree sample
Every function scored and mapped to a mode, with the order that pays off first. One shown in full as a sample.
The tool, model, and cost catalogue
The specific tools and models per function, with real monthly run and one-off setup costs for your size.
Knowledge and data architecture
What your agents need to read, where it lives today, and the groundwork to get your data ready first.
Non-negotiables and what stays human
The lines you don't cross, the calls that stay with people, and where a human always signs off.
Team and org readiness
Who runs this, what changes in roles, and the skills to build before you scale agents.
Forks in the road
The decisions that branch your path, with the trade-off behind each so you choose with eyes open.
What is still open
The questions a form can't answer, surfaced honestly, and how the calibration closes them.
Sequencing your path
The phased, gated build order: what to automate first, second, third, and what each phase unlocks.
The risk register
Where this can go wrong, ranked, with the mitigation and the early warning sign for each.
What to do now
Your first 90 days, step by step. Groundwork you can start Monday, then the exact next moves.
Methodology and resources
How the report was built, the benchmarks behind it, and the sources to go deeper.

See what's possible in 10 minutes.

Online assessment, evaluated by our trained AI. A real sample report, not a teaser. No sales call required to get it. No card required to start. Full report EUR 450.

Start free assessment →
// before you start

Questions, answered straight.

No. You answer the questions online, our trained AI evaluates them, and you get a real sample report back: executive summary, your four-mode position, what should stay human. No call required to receive it.
The assessment and your report are written by the flagship models from Anthropic, OpenAI, and Mistral (the same class of model behind Claude, ChatGPT, and Le Chat), steered by prompts we built on real implementation paths. The model does the reasoning and writing; our framework decides what to ask, how to score it, and what to recommend. You get current-generation AI quality, not a static template.
Yes. The whole point is to meet you where you are. The assessment locates your actual starting point and sequences from there, whether you're nine tools deep or haven't started.
The free sample includes your executive summary, your AI vision across the four modes, and a function-by-function sample. The full report adds the tool, model, and cost catalogue, the data architecture, what stays human, the risk register, sequencing, and your specific 'what to do now' plan.
The free report is genuinely free, and so is the calibration step that follows it. You only pay at the point we generate the full document, because that is where the heavy reasoning runs: the costed tool and model catalogue, the data architecture, the risk register, and your sequenced plan. We never charge for something already produced.
About 10 minutes for the assessment itself. You don't need to prepare anything: answer in plain language about how your team works today, and the questions branch based on what you tell us. The optional calibration afterward goes deeper only if you choose to continue.
Your answers are used to generate your report and nothing else. They are not sold, and they are not used to train the underlying models. Free-text answers are sanitised before they reach the model, and the paid report is generated server-side after payment confirms.
Surprisingly accurate. We use a carefully trained AI with access to industry and company benchmarks. A form still can't know your real margins, exact tool costs, volumes, or how brittle your integrations are, so the free report flags those gaps openly rather than guessing, and the calibration step is what closes them for the paid report.
No. The entire flow, from assessment to free report to full report, runs without a call. If you want a human afterward, you can ask, but nothing about getting your report depends on it.
Three stages. The opening questions take about 10 minutes. The calibration that follows goes deeper, roughly an hour, and it stays free. Only then do you pay, and the full report generates in around 10 to 15 minutes. It lands in your account, ready to read on screen or download whenever you want.
Yes and no. The report is your starting point, step 1 of 5, and you can act on it as it stands. Tuning it further is bespoke calibration, the kind that can't be done through an online assessment. If you want to go that far, contact us and we'll take it from there.
Five stages that build on each other into a plan you can start Monday. 1. Landscape of Choices: the full map of what running on AI could look like for you, and this free report is a sample of it. 2. Direction Setting: a working session on your functions, team, and tools, where we set the direction together. 3. Path Deep Dives: function by function, in payoff order. 4. Tailored Implementation Plan: phased, costed, every step gated on the one before. 5. Execution Cascade: the live build, an iterative plan that keeps evolving as your team builds and executes, with hand-holding only where you want it.
Not for now. We can guide you all the way to a concrete playbook and implementation plan, but who builds it, and how, is your call. That independence is the point: we have no stake in which tools you pick, so the assessment stays honest and unbiased. The choice is fully yours.