AI transformation as a sequence you can staff

Transformation is the longest Nidrosoft service. It includes the other doors on purpose: audit and strategy to rank work, product or AI systems to ship, workshops so more people than the first operator can run and extend the system. The method stays Embed, Diagnose, Scope, Build, Own. The difference is that Own is expected to survive a second and third team, and that we plan the handoff before the first build starts.

Companies ask for transformation when a single automation feels too small and a strategy pack feels too empty. That instinct can be right. It can also be a board word looking for a budget. We will not accept the word unless there is a sponsor, a first department, and a measure that already exists. Past 250 employees, that constraint is how we stay out of the McKinsey and BCG style lane. Those firms win when the artifact is a company-wide narrative. We win when the artifact is a series of live processes.

Cyriac Zeh founded the studio after Anthropic, Microsoft on Bing and Edge, Intuit, and Gap Inc. One hundred twenty-five products shipped. Sixteen public. That history is a reason to trust sequencing. It is not a reason to hire us as a hundred-person rollout army. If you need that army, hire a firm that staffs one. We will still take department one if the sponsor wants a studio on the first system.

Phase one is the audit. Two weeks in the work if access is ready. Ranked processes, systems, data, risks. The packet names team one. If the packet cannot name team one, stop. Buy seats or a leadership sprint instead.

Phase two is the first build. A written outcome. Human approvals on money, contracts, and customers. Code and prompts in your accounts. An operator who can run Tuesday. How-we-work often places implementation after the plan, in a multi-week window. Transformation does not magically shorten that window. It only funds what comes after.

Phase three is training beyond the operator. Leadership learns to reject bad add-ons. Adjacent roles practice on the new path. The playbook from the training service becomes the internal course. If you skip this phase, team two will treat team one’s system as folklore.

Phase four is the next team. Same method. New map slice. New scope. The people from team one teach. We stay as needed. We step back where Own is already true. Improve-and-support is available. It is not an invented forever fee on this page.

Between phases we write a go or no-go. The sponsor can stop after team one and still have a complete purchase. That stop is not a failed transformation. It is a company that got one owned process and declined to fund the next. We would rather see that honesty than a zombie program that keeps a steering slot on the calendar with nothing new in production.

Copying team one’s prompts into team two without a new Diagnose is a common failure. The nouns change. The approval path changes. The unofficial spreadsheet changes. We will charge for a thin remap rather than a full audit when the systems are shared, and we will still sit with the new operators. Embed does not expire because the banner is already printed.

Communications inside the company stay factual. We will help the sponsor describe what is live and what is not. We will not write an internal campaign that implies every desk is automated. People who still do the unofficial path deserve a sentence that names their work. That sentence is how they later join the map instead of fighting it.

• Team one named before any program banner
• One measure per build that the business already believes
• Training files in your drive
• A written rule for what team two is allowed to copy
• A written rule for when we leave a workstream

Under 30 people, transformation is usually the wrong word. You need one build that pays for itself and a founder who will live in the approval step. We will sell you an audit and a build. We will resist a twelve-month program banner that exists to impress an investor.

Between 30 and 250, transformation can mean two or three departments over a longer calendar. Ops, then finance. Clinic front desk, then billing. Dealer desk, then title. The sponsor still has to protect the first outcome from a wish list.

Past 250, start in one department even if the CEO wants a town hall. Town halls are training, and even then they work better after a live path exists. A strategy firm can run the town hall and the governance pack. We can run department one. Write the boundary so both invoices make sense.

Remote programs are normal. Cyriac is in San Diego. Teams elsewhere already work this way. We will not invent offices. If a works council or a regulator requires on-site days, say so in the audit. We can plan travel. We cannot pretend to be local staff in twelve countries.

Each build still leaves code, prompts, eval notes, runbooks, and accounts you control. The program adds a shared index: which process is live, who operates it, which measure moved, which prompts are canonical. Think of Nerlude’s instinct (one place for the messy infrastructure) applied to AI operations. You do not have to buy Nerlude. You do have to have a place.

TokenTra-shaped visibility should appear once more than one team spends on models. Teams have cut 10 to 30 percent of wasted spend when they can see tokens by feature. A transformation that adds agents without adding a bill view will hide waste in the program’s success story.

Protectron-shaped trails belong if you operate under the EU AI Act or you sell into teams that do. The program can fund the paperwork habit. It cannot replace counsel. Write that sentence into the charter.

Playbooks from training stay in your drive. Strategy packs from another firm stay theirs to update. Do not mash the files. Assistants and new hires will quote the wrong one.

We will not staff a transformation that is only communications. We will not staff one whose first deliverable is a homepage bot. We will not staff one that forbids a human on customer send. We will not staff one that requires a fifty-person bench in month one. We will not staff one that treats Nidrosoft as a logo next to a strategy firm while we are asked to write the pack.

We will not staff one that ignores the unofficial stack. If inventory still lives on Marketplace, the program must look at Marketplace, the way Nexuvo does, or it will automate the wrong desk. If briefs still take three hours, look at the brief, the way NanoBrief does. If audio is the product, look at Clear Audio AI and MixGuidera as proof that media is in scope, then still rank it against the inbox.

We will not staff one that cannot survive Cyriac being in one workstream at a time. Studio capacity is a constraint. Write it down. If you need parallel factories, hire them and keep us on the first stubborn process.

If quarter one must produce a board pack and nothing in production, hire the strategy firm. If quarter one must produce Copilot usage, buy seats. If quarter one must produce a module on a single platform, hire that platform’s SI. If quarter one must produce a live process in tools you already run, stay here.

A multi-vertical instinct already exists in Omnira, with Omnira Dental live and other verticals in build. That is a product fact, not a claim that your company will receive Omnira. It shows that a studio can carry more than one operations surface without changing the method.

Across the public archive the same pattern repeats. Eviction Wizard: one dashboard, a legal path, a human on the writ. Nexuvo: overnight work, a 3x inventory result, a person on the purchase. TokenTra: visibility, then a 10 to 30 percent cut in wasted spend. NanoBrief: a document job collapsed from three-plus hours to under two minutes. Octana, Guidera, Protectron, Nerlude, Clear Audio AI, MixGuidera: edges that still have a Tuesday user.

A transformation program inside your company should be able to point at similarly concrete leftovers. If the only leftover after six months is a steering deck, you bought the wrong row in the table.

Every build in the program keeps a human on consequential actions. That rule does not relax because the banner says transformation. Clinics, landlords, dealers, and funds already know why. We will write the signer’s name into the scope. If the signer is a committee that meets monthly, the process cannot be overnight. Change the process or change the committee.

Security and legal join at the audit and stay on a cadence. They do not replace operators. EU hosting and residency appear when the work needs them. Audit trails are part of how we work, not an upsell surprise.

If another firm owns enterprise risk language, let them. We will implement inside that language where it is specific. We will not rewrite their pack for sport. We will flag where the pack assumes a write path that does not exist.

Cyriac Zeh can hold a limited number of serious workstreams. Transformation planning must respect that. Team two may wait. Waiting with a clear map is better than three half-built agents. If waiting is unacceptable, hire a shop for the specified work and keep the studio on the unspecified process.

Rent-an-AI-engineer can extend capacity after the first outcome, still as a scoped seat, still without a public monthly number on this page. Book the audit. Decide whether you need a program, a single build, or a monthly embed.

Southern California companies can mix on-site days with remote weeks. Everyone else should expect remote embed plus planned travel. We will not staff a fake local office to win an RFP in London or Dubai. The regions pages say the same thing.

Use measures the department already argues about. Time to complete a file. Inventory acquired. Tokens per feature. Hours per brief. No-show rate. Days outstanding. We will add AI-specific notes (eval fail rate, approval lag) as supporting lines, not as replacements.

Public numbers we can cite stay limited. Nexuvo’s 3x inventory and 80 percent less manual outreach. TokenTra’s 10 to 30 percent wasted spend. NanoBrief’s under two minutes versus three-plus hours. We will not invent a program-level ROI for a slide. Your finance team can compute one after team one has a month of real use.

If a vendor promises a multiple before seeing your stack, treat it as marketing. Our audit exists to prevent that conversation from becoming a contract.

Publish the measures where the department already looks, not on a separate innovation dashboard that only the program team opens. If the dealer already lives in a board for cars acquired, put the new number there. If finance already lives in a token invoice, put the waste cut there. Separate dashboards die when the program sponsor changes jobs.

Qualitative notes still belong in the index. Operators will tell you when a path is technically live and socially unused. That sentence is a measure. We will not bury it under a green status. A transformation that ignores unused systems will repeat the same build for the next vendor.

Write cyriac@nidrosoft.com or book the public calendar. Name the sponsor, the first department, the system of record, and the measure you already trust. Attach a strategy pack if you have one. We will say whether a Nidrosoft program, a single build, training only, or another firm is the honest next step.

Scope sets the commercial terms. No euro figure and no retainer appears here because a two-department software company and a multi-site operator are not the same program.

Done for a transformation is not a party. Done is team two running a process with files they can edit, while team one still runs theirs, and Nidrosoft is optional on Tuesday. If you want a party, throw one after that is true.

Bring the last three vendor decks you were shown. We will mark which leftovers those decks actually sell. A platform SI, a seat bundle, and a strategy pack often arrive in the same month as a Nidrosoft conversation. Sorting them in the first audit saves a year of overlapping retainers.

If works councils, security boards, or a parent company must approve the first write, name the date. A program that ignores that date will look slow and then look reckless. We would rather wait in Diagnose than ship a shadow path. Enterprise buyers already know this. Mid-size buyers sometimes learn it in week six. Learn it in the email instead.

Stopping after team one is allowed in the confirmation email. Print that sentence. A program that cannot stop will invent work for team two even when team one is unused. Unused systems are how the next vendor gets hired to rebuild the same desk.