Partner Program · for ERP firms, MSPs and agencies

Your clients are asking about AI.
You don't have to build a team to answer.

You already have the relationships, the domain knowledge and the trust. What you may not have is an AI product to sell or an engineer who can deliver one without putting your reputation at risk. We are the delivery arm you can put behind those conversations — with shipped, purchasable products, a demoable ERP assistant, read-only architecture and an on-premise option your more conservative clients will ask for first.

Request the Partner Pack See the products you'd be selling

This is a new programme. We are onboarding a small number of partners deliberately, so that each one gets real engineering attention rather than a portal login.

Who we partner with

We are deliberately narrow here too. Partnership works when you own a client relationship we could not reach on our own, and we supply capability you would otherwise have to hire.

The firm

  • Dolibarr implementation companies and ERP consultancies
  • Managed service providers who already run their clients' systems
  • Fractional CFO firms, inventory consultants and accounting-technology practices
  • Digital agencies with real clients and no internal AI team

The situation

  • Clients are asking what you can do with AI, and the answer today is "not much"
  • Saying "we don't do that" quietly hands the relationship to somebody else
  • Subcontracting to a generalist shop once already went badly, or you can see how it would
  • You want to add a line of revenue without hiring specialists you cannot keep busy

Where this doesn't fit

If you do not yet have a client base — a partner programme is not a substitute for one, and we cannot supply demand. If you want a reseller margin with no involvement in discovery, scoping or the client relationship, the economics do not work for either of us and delivery suffers. If you expect us to fund your sales cycle with unpaid pilots, custom demos on speculation or free proof-of-concept work to win business you have not yet qualified, we will decline. We would rather say this on the first call than three months in.

The four models

Pick the one that matches how much of the client relationship and delivery you want to own. Partners often start with one and move to another as trust builds.

Referral

The lightest option. You make the introduction and step back.

  • Who fronts the client: we do, under the SiliconBlaze name
  • Who delivers: we do, end to end
  • How you earn: commission on what the engagement bills

Reseller

You sell our Dolibarr modules and deployments into your own book as part of your offer.

  • Who fronts the client: you do, as their supplier
  • Who delivers: we build and support; you handle installation and first-line contact
  • How you earn: your margin on licence and deployment, plus your own implementation fees

White-label

We are invisible. The client sees your brand, your process and your team.

  • Who fronts the client: you do, entirely
  • Who delivers: we do, under your brand and to your schedule
  • How you earn: you price the client; we invoice you at partner rates

Joint delivery

The model that usually produces the best work, because both sides are in the room.

  • Who fronts the client: you do — the relationship stays yours
  • Who delivers: shared discovery and demonstration; we take the AI engineering, you keep the ERP and process work
  • How you earn: you bill your own scope directly and take agreed terms on ours

Whichever model you choose, it is written down before we speak to a client — including who owns the account, who is on which call, and whose name is on the deliverable.

What you can put in front of a client tomorrow

Nothing on this list is a mock-up or a roadmap item. These exist, they run, and you can show them on a call this week.

Stock Intelligence (v1.5)

A shipped, purchasable Dolibarr module on the official DoliStore. Demand forecasting, reorder planning and stock health scoring inside the ERP your client already runs.

See Stock Intelligence →

Financial Intelligence

Our second shipped DoliStore module. Financial reporting and analysis for finance teams that have outgrown the standard reports but do not want another system.

See Financial Intelligence →

Doli Brain — the live ERP assistant

A working assistant that answers plain-English questions against live ERP records and shows the underlying figures. This is the demo that changes the conversation, because the client can ask it their own question.

See the live demo →

GroundTruth — document intelligence with citations

For clients whose problem is contracts, policies and reports rather than stock. Answers are grounded in their own documents, with the source shown. Useful when the objection is "how do I know it didn't make that up".

Open GroundTruth →

Why this matters to you specifically

When you introduce an AI partner, your client's first unspoken question is whether you have brought them a real engineering team or a reseller with slides. Shipped products on a public marketplace, a demo they can drive themselves, and a read-only architecture you can explain in one sentence answer that question before it is asked. Accuracy is measured against a 480-case internal test suite run on a real database, and there are 15+ years of production engineering behind delivery.

How an engagement runs

From your side of the table. Every stage has a defined end, so you are never telling a client that a project is "ongoing" with no date attached to it.

Stage Duration What ends it
Introduction One call You brief us on the client and the opportunity. We tell you honestly whether it is a fit, and which of the four models suits it. Terms are confirmed in writing before anyone contacts the client.
Joint discovery 1–2 calls We meet the client together, under whichever brand the model calls for. You lead on business context; we lead on feasibility and the demo. Ends with an agreed problem statement.
Assessment 1 week A written roadmap and a fixed-price proposal the client keeps whether or not they continue. This is the low-risk first sale, and the easiest thing for you to get signed.
Proof-of-Value Sprint 2–4 weeks One workflow running against the client's real data, with accuracy measured. If it does not work, you find out in weeks rather than after you have staked the relationship on it.
Production deployment 6–12 weeks Live system, role-based access, logging, training, documentation and handover. Where you hold a retainer or support contract, handover is written for your team to operate.

Ranges reflect scope, not uncertainty — the assessment fixes which end of the range applies before the client commits to anything, which is exactly what makes it safe for you to recommend.

The security model

This is what your client's IT lead will ask you the moment you mention AI — usually in the first ten minutes, and usually with some scepticism. Here is what you can tell them, and it is all defensible.

"Can it change anything in our ERP?"

No. Our ERP assistants connect with read-only credentials. They can query and report, but cannot modify records. At the database level, writing is not available to them.

"How much of our data can it see?"

Only what a given feature needs. Access is scoped, queries can be inspected, and data can be allow-listed. The client decides what is visible, not us and not you.

"Does it have to go to the cloud?"

No. The client chooses where it runs. On-premise deployments can use locally hosted open models, so their data never leaves their environment. For regulated or conservative clients, this is usually the answer that unblocks the deal.

"Will our data be used to train AI?"

No. Where a product uses a hosted model, we use established providers via their APIs — or a local model where preferred. We do not use client business data to train AI models.

"What happens when it doesn't know?"

It is designed to say so rather than guess. Accuracy is measured against a 480-case internal test suite run on a real database — a number you can quote, because it is a real test suite rather than a marketing figure.

"Can we see what it was asked?"

Yes. Personal data can be masked and result sets limited, and systems can be configured with audit logging of queries and actions, so the client always knows what was asked and what it saw.

An honest note on certifications

SiliconBlaze does not currently hold formal certifications such as SOC 2. We state that plainly rather than implying otherwise — and we would rather you heard it from us than discovered it in a client's vendor questionnaire after you had vouched for us. What we offer instead is transparent, technically specific documentation of how our systems handle data, and the option to run everything inside the client's own environment. Full detail is on our security page, which you are welcome to forward directly.

What we ask of you

A partnership is an obligation in both directions. These four are the ones that decide whether an engagement goes well, so we would rather be explicit about them up front.

A real introduction, not a list

One warm introduction to a client with an actual problem is worth more than a mailing list, a webinar audience or a spreadsheet of contacts. We do not run cold outreach against partner databases, and a shared list is not a referral.

Access to the decision-maker

At some point we need the person who can approve budget and grant data access in the room. Engagements scoped entirely through an intermediary tend to be re-scoped later, at somebody's expense — usually yours, because it is your relationship.

Honesty about the data

If the client's ERP was migrated last quarter, if stock records are known to be unreliable, or if half the history lives in spreadsheets — tell us before the assessment, not during it. We can usually work with poor data. We cannot work with a surprise.

No overselling on our behalf

Please do not promise capability we have not agreed to, timelines we have not quoted, or accuracy we have not measured. If a client asks something you are unsure about, the right answer is "let me check with the engineering team" — and we will answer the same day.

What we commit to

The same four obligations, pointed the other way. You are lending us your reputation, and these are the terms on which we accept it.

Named engineering accountability

You get a direct line to the people building the system, not a ticket queue and not an account manager relaying messages. When something breaks or a client asks a hard technical question, you are talking to whoever wrote the code.

We do not go around you

Clients you introduce are yours. We will not approach them separately, quote them directly, or use the engagement as a route into the rest of their business. If they contact us independently, we tell you.

Honest scoping, including "no"

If a deal is a bad fit — thin data, an unrealistic timeline, a problem AI is the wrong tool for — we will tell you before you have committed to it in front of your client. A declined engagement costs us one project. A failed one costs you a relationship.

Commercial terms in writing first

Commission, partner rates, scope boundaries, support responsibilities and account ownership are agreed and documented before any client conversation happens. Nothing about your economics is decided after the deal lands.

What it costs — and what you earn

There is no fee to join, and no volume commitment. Commission rates and reseller terms are agreed per partner — they depend on which model you choose, how much of discovery and delivery you take on, and whether the client is yours or ours to support afterwards. We put them in writing before your first client conversation rather than publishing a single rate that would be wrong for most partners.

What we can show you is the client-facing ladder, so you can see the size of the deals you would be working with:

Easiest first sale

Assessment

$390

Fixed price, one week. Data and process review, three prioritised use cases, feasibility and security notes, a roadmap and a fixed-price proposal the client keeps either way.

Proof-of-Value Sprint

from $2,400

2–4 weeks. One data source connected, one workflow built and running against real data, accuracy and security evaluated, production plan written.

Production Deployment

from $9,000

6–12 weeks. Data integration, role-based access, the assistant and intelligence modules, logging and evaluation, training, documentation and handover.

AI Operations Retainer

from $600/mo

Optional and cancellable. Monitoring, evaluation, pipeline maintenance, new reports and workflows, monthly review. Often the piece a partner keeps for themselves.

Prices in USD. If you or your client bill in EUR, tell us on the call and we will quote in EUR at the equivalent rate.

Start with the Partner Pack

The four models side by side, the commercial terms we work from, the security documentation you can forward to a client's IT lead, and what a first joint engagement actually looks like. Then one call to work out whether there is a fit. This is a new programme and we are onboarding deliberately, so tell us about the clients you have in mind.

Request the Partner Pack

Or see the products you'd be selling — no form required.