What to automate
Repeated tasks, re-entry work, alerts and handoffs that already drain the team.
AI consulting for SMB leaders
GPT-5.5, Claude Fable 5, Claude Opus 4.8, agents, open source, automation, copilots. The point is not to chase every announcement. The point is to know what changes for your company, your margins and your priorities.
A clear roadmap, not a list of tools to try.
Business decisions: what to change, keep, build or avoid.
Leadership support to turn ideas into tradeoffs.
GPT-5.5
Models are becoming more useful for reasoning, coding, searching and taking action.
Claude Fable 5
Long-horizon agents make workflows credible that used to feel too ambitious.
Claude Opus 4.8
Choosing the right model, cost and level of control is now a strategic decision.

Who helps you decide
I am not here to sell a magic tool. I help you sort the real levers, the bad bets and the projects that can become concrete inside an SMB.
Market noise
Your team may already use ChatGPT. A sales rep pitches a new tool. A competitor talks about AI agents. A vendor promises to automate everything. Meanwhile, your real question remains: what deserves your time, budget and attention as a leader?
A good AI strategy for an SMB is not a grand abstract plan. It is a series of sober decisions: where AI can change your offer, margin, responsiveness or organization, and where it will mostly add disorder.
Decisions to make
A leader does not need to know every AI architecture. They need to know which initiative to launch, in which order, and with which level of risk.
Repeated tasks, re-entry work, alerts and handoffs that already drain the team.
The people already using AI quietly, and those who will decide, review or transmit work.
Market tools that are good enough, without building an unnecessary custom system.
Agents and workflows that deserve custom work because your trade or data makes the difference.
Projects that sound attractive but are too vague, risky or expensive for your current stage.
SMB use cases
We start with your company: margin, offer, sales, quotes, customer relationship, admin, reporting. Only then do we choose the right AI layer.
Model
01If customers can get a first answer, draft or analysis from a public assistant, your perceived value can shift.
We separate what becomes a commodity, what should be reinforced, and what could become a new offer or premium service.
You see how to adapt your positioning before the market forces you to.
See my approachSales
02Files, emails, CRM entries and follow-ups exist, but no one has time to turn them into clean actions.
We decide whether the right topic is a sales agent, better CRM hygiene, training or a preparation workflow.
Sales becomes a prioritized initiative, not a stack of tools opened in parallel.
Discover HermesQuotes
03Field photos, voice notes, customer constraints and history are scattered. Quotes arrive late or depend too much on one person.
We check whether AI should structure inputs, draft a quote, detect gaps or automate the CRM handoff.
The quote becomes a sales lever, not just an admin document to produce.
Read the field quote caseSupport
04The same questions come back, but some requests are sensitive and should not be automated blindly.
We separate simple replies, requests prepared for a human, and topics that require approval.
You gain responsiveness without handing the customer relationship to a robot.
See AI agentsAdmin
05Invoices, follow-ups, incoming documents, supporting files and checks take mental space every week.
We identify what belongs in a reliable workflow, an AI agent or a simple business rule that needs clarifying.
You know where to save time without creating a fragile system.
Read the 2026 invoicing guideLeadership
06Revenue is in Pennylane, opportunities are in the CRM, urgencies are in email and priorities live in your head.
We scope a simple leadership report: sources, frequency, alerts, summaries and confidence limits.
You start the week with decisions to make, not ten tabs to reopen.
Discover OpenClawRight entry point
Consulting is not here to replace the other offers. It is useful when you first need to sort, prioritize and choose the right path.
When your team needs to use ChatGPT, Claude, Copilot or Notion AI better on real business cases.
See trainingWhen the process is already clear and CRM, invoicing, email, files or alerts need to be connected.
See automationWhen a task needs to read, interpret, prepare an action and alert with a human in the loop.
See AI agentsWhen you feel AI will change your market, but you do not yet know where to place your bets.
Stay on consultingMethod
We clarify your activity, margins, sales stakes, tools, urgencies and concerns.
We identify flows, repeated tasks, friction points, sensitive decisions and available data.
We classify initiatives: training, automation, AI agent, existing tool, topic to avoid or topic to postpone.
You leave with a readable roadmap: order, objective, dependencies, risks and next step.
We keep a steering session if you want to secure choices, review quotes or move forward initiative by initiative.
Formats
The right format depends on your maturity, team size and the level of transformation you are considering.
Short session
To sort an idea, challenge a decision, review a proposal or know whether an initiative deserves to start.
Complete diagnosis
Interviews, mapping, use cases, tradeoffs, risks and prioritized roadmap to decide without fog.
Monthly or punctual
To keep direction: tool choices, quote review, prioritization, team adoption and first project steering.
No public pricing. Scope depends on diagnostic depth, number of stakeholders and expected follow-up level.
Scoping
In 20 minutes, we look at your context, urgency level and the decision you need to make. If the right answer is training, an agent, a workflow or nothing for now, I will tell you.
After the call, you receive a proposal with scope, deliverables, timeline and support level. You keep control of the next step.
Commitments
I do not sell licenses and I do not earn tool commissions. The choice must serve your company.
Recommendations talk about margin, workload, risk, customer relationship, team and priorities, not useless jargon.
Ideas stay connected to what can actually be trained, automated or built.
If an AI initiative is not relevant now, it should leave the roadmap.
FAQ
Describe your situation in a few lines, no commitment.