Skip to content
Menu

What we do

Four services. One disciplined way of making AI actually work.

Every service below solves a specific problem. Either your AI investment is not producing returns, or you have not invested yet and do not know where to start.

The four services

Engagement path AI Stack Audit Implementation Custom AI Agents Ongoing Optimization Enter at any point

These are not standalone packages. Most engagements start with an audit, move into implementation, and continue with ongoing optimisation. Some start somewhere else entirely, with a paid media account nobody can read, and widen from there. You can engage us at any entry point.

01

AI stack audit

2 to 3 weeks Most common entry point

Before we touch anything, we inventory everything.

Most businesses we meet are paying for more AI tools than they can name. A ChatGPT subscription. A chatbot vendor. Three automation platforms. A voice agent nobody configured. The audit is how we find out what is actually earning its place.

  1. 01 Full inventory. Every AI tool, subscription and workflow. What you pay, what it produces.
  2. 02 Team interviews. We talk to the people doing the work. That is where the manual tasks AI should be handling show up.
  3. 03 Cost attribution. Every subscription matched to a real outcome. Anything not producing gets flagged.
  4. 04 ROI shortlist. The three highest-return implementations we would prioritise next, with the arithmetic behind each one.
  5. 05 Written recommendations. Tools to keep, cut and replace. You walk away with a document whether you continue with us or not.

Why this matters. Most AI vendors show up selling. We show up asking questions. The audit frequently finds enough stack sprawl to pay for itself before we implement anything new.

02

Implementation

2 to 4 weeks per system Follows the audit

We deploy the systems. We measure from day one.

Implementation is where most AI projects die. Not because the tools do not work. Because nobody set up the measurement layer, nobody connected the systems to the existing stack, and nobody stayed around to fix what broke in the first week.

  1. 01 Connect systems. New AI tools wired into your CRM, calendar, email and existing workflow.
  2. 02 Define every KPI. Every system has a measurable success metric configured before it goes live.
  3. 03 Build measurement alongside. Reporting infrastructure deploys with the system, not after it.
  4. 04 Train your team. Exactly what the system does. Exactly what it does not.
  5. 05 Tune the first two weeks. Real usage always differs from plan. We monitor and adjust closely.

Why this matters. The last mile of implementation is where projects fail. That is the mile we specialise in. By the time we deploy anything we have mapped what could break, and we stay close for the first two weeks to catch what we missed.

03

Custom AI agents

3 to 6 weeks, scope dependent Standalone or bundled

When the right tool does not exist yet, we build it.

Sometimes the right tool does not exist. A solo operator needs a digital chief of staff. A service business needs an agent that handles leads while the owner is on a job site. When the market has not built it, we do, on open-source frameworks you own.

  1. 01 Scope against outcomes. We define the agent by specific tasks and results, not vague capabilities.
  2. 02 Choose the right framework. Hermes or OpenClaw. We pick on your use case, not our preference.
  3. 03 Train on your context. The agent learns your business, your tone, your workflows.
  4. 04 Test across real scenarios. Before deployment we run the agent through the edge cases that actually happen.
  5. 05 Deploy where you already are. Your email, your Slack, your tools. No new interfaces to learn.

Why this matters. We build on open source, and you own the agent outright: code, data, configuration. If we disappeared tomorrow it keeps running, and any competent engineer could maintain it. Most custom-build shops lock you into their platform. We do not.

04

Ongoing optimisation

Monthly, ongoing Month to month

We do not hand over a dashboard and leave.

The fastest way to waste an AI investment is to deploy it and walk away. Systems break. Contexts shift. The thing that worked in month one is quietly underperforming by month four. Optimisation means we stay, review every month, adjust what is not working, and document the returns in numbers.

  1. 01 Monthly 30-minute optimisation call. Review performance against KPIs. Decide what to adjust.
  2. 02 Shared communication channel. Slack or email, your call. Reachable during business hours.
  3. 03 Quarterly deep dives. Every three months we surface new automation opportunities we see forming.
  4. 04 Continuous tuning. Agents get sharper. Automations get tighter. Integrations get retested.
  5. 05 Documented change log. You always know what was adjusted and why.

Why this matters. Most AI agencies ship and disappear. We stay because it is how the systems produce returns. You can end the engagement at any time. We built the model so that ending it does not end the systems.

In practice

What it looks like when the four run together.

ENG

One engagement, widening

Anonymised

One engagement began as paid media management for a specialty retailer that sells through a website and through offline channels. It did not stay there.

Running the ad account led to building the measurement layer around it, because the account could not be judged without one. That led to AI systems, then to custom agent work, then to software. Not because an expansion was pitched. Because each piece of work made the next one obviously worth doing, and because we were still there to do it.

That sequence is the entire argument for co-managed operations. An agency that ships and leaves never gets to the second step.

Why there is no revenue figure here. Analytics can only see part of how this business sells. We could publish the number it does see. It would be a real number, and it would be measuring the slice the tracking reaches while calling it the business. The same discipline is why the rest of this site carries the sources for every figure on it.

Where to start

You do not need to know exactly what you need when you book the first call.

Most clients arrive in one of three situations, and all three lead to the same conversation.

You are already paying for AI and are not sure it is working.

Start with the audit. We map what you have, find the stack sprawl, and give you a written shortlist. No commitment to continue past the audit.

Go to the audit

You know what you want to implement but not how.

Start with a call. We scope the implementation, estimate the timeline and give you a written proposal.

Book a call

You are tired of tools that do not talk to each other.

Start with a conversation. We will tell you honestly whether we can help.

Book a call

You are spending on ads and cannot tell what they return.

Start with the account. We read what is actually happening in it, separate the spend that produces from the spend that does not, and show you the arithmetic. Engagements often begin here.

Book a call

Questions

Questions

How is this different from hiring an AI consultant?

Consultants sell strategy. We sell implementation. A consultant hands you a roadmap and leaves. We build the systems, measure the returns, and stay through the optimisation phase. You can engage us for a one-time audit if you want strategy only. Most of our work is in the doing.

How is this different from just using ChatGPT?

ChatGPT is a good tool and many of our clients already use it. We are tool-agnostic. If it is the right fit we will help you get more out of it: better prompts, connected workflows, team-wide playbooks, and measurement so you can see what it produces. If you need something it cannot do, such as remembering your business across sessions, connecting to your CRM, or taking action on your behalf, we will bring or build the right tool.

Do I need to already have AI tools to work with you?

No. Roughly half our clients arrive with tools they are not sure are working. The other half start fresh. If you have nothing yet we skip the inventory and go straight to identifying your highest-return first implementation.

What size business is this for?

Our focus is service businesses and e-commerce brands in the $500K to $5M revenue range, plus solo operators and consultants who want custom agents. We have worked with larger businesses. The engagement model is built for operators who want hands-on support without enterprise pricing.

What does this cost?

Pricing depends on scope, and we scope it on the first call. Audits are a fixed fee, implementations are priced per project, and ongoing optimisation is monthly. We do not publish a rate card, because a published number that does not survive contact with your situation is worse than no number.

Can I end the engagement if it is not working?

Yes. No long-term contracts. Engagements are month to month. We built the model this way because we want you to stay because it is producing returns, not because you are locked in. Everything we build is yours. If you leave, the systems keep running.

What if Breaking Trail goes away?

Everything runs on tools and frameworks you own. Open-source agent frameworks. Standard automation platforms. Your own CRM and email systems. Any competent technical person could maintain what we built. We are not selling a proprietary platform. We are selling the work.

Start the conversation

The first call is a conversation, not a pitch.

Sixty minutes. You tell us what you have and what you are trying to fix. We tell you whether we can help. If we cannot, we will say so.

Book a call Free AI ROI Scorecard