Three workflows, built and handed over working

The Build Series

Three workflows your team does by hand today, built properly and handed over running. Each one starts with a written definition of done, gets built against your real tools and data under NDA, and is tested by your own people against that definition before anyone calls it finished. Then we train the team on it, because a build nobody knows how to use is shelfware.

Starting at $25,500

3 builds · roughly 90 days · scoped on the call

Some work is too big to build in a room, and everybody knows which work it is.

Your team can learn Claude and build their own Projects and Skills, and most teams should. That is what the Custom Program is for. Then there is the other list: the workflow that touches four systems, the one that has to match a house standard exactly, the one where being wrong costs a client relationship.

Those need a brief, a written standard for what finished means, access to the real data, and somebody testing it properly before anyone trusts it. That is a build, and ninety minutes in a room with the team does not get you one. Treating it that way is how teams end up with something that demos well and gets quietly abandoned.

This engagement does three of them, to the same sequence each time, over roughly 90 days. Then it trains your team on each one, which is the step everybody skips. If you already know which three, that is the starting price. If you do not, we open with intakes and a short written build plan naming which workflows to build and in what order, and that costs more.

The sequence is The Build Method: Understand and Decide once at the start, then Prototype and Validate for each of the three.

What’s included

A written brief per build, not one for the engagement

Each of the three starts with its own conversation: what the workstream is, what is in scope, what is out, and the list of what we need from you before anything starts. A workstream whose brief cannot be written down is not ready, and we say so instead of starting it. Teams who do not yet know which three they want start a step earlier, with intakes and a short build plan. That is priced on top.

A written definition of done, agreed before we build

What it takes in, what it emits, which surfaces it has to work on, and what it must never do. Anything your own standard has not settled is written down as an open decision with an owner and a date. You sign this off before a line of it exists, so acceptance is measured against a document and not against a conversation on delivery day.

One named owner on your side, per build

One person who agrees the definition of done and signs the acceptance. Usually the person who will own the workflow afterward. Builds without a named owner drift, and the engagement is structured so they cannot.

NDA, then real access to your tools, data and systems

A signed NDA comes before any of your data reaches us, then we work against the real thing. This is the step that decides whether a build is production or a demo, and it is where build engagements most often stall, so we start it on day one instead of discovering it in week three.

The build itself, in your environment

A Skill, a plugin, a scheduled task, or a reporting workflow, built by us and delivered into your own account. Where several pieces belong together, they ship as a plugin so you can enable the whole thing for the firm in one move instead of asking every person to switch it on.

A test pack your own people run

One row per case, written from the definition of done, with the must-pass cases marked and each surface tested separately. Your team runs it and logs what they find. We triage inside two business days, fix, and reissue with a new build version so a retest proves the fix instead of overwriting the evidence for it.

Acceptance in writing, against the bar you set

Every must-pass case passing on every surface in scope, and no open finding that produces a wrong result. Both halves are counted on the sheet, not asserted. If you want a different bar, it goes on the spec before the build starts, never negotiated at delivery.

Training the team on each build once it is accepted

A session per build, for the people who will use it. This exists because of what happens without it: on our largest engagement ten things got built and only four were switched on for everyone, and the gap was never the quality of the build. Presence in an admin console is not adoption.

Recordings, reference guides and a private microsite

Every training session recorded, a written guide per build, and one page holding all of it with the specs and the test packs. When the person who owned a build leaves, the next one inherits a document instead of a rumor.

Starting at $25,500

How it works

  1. 1

    Fit call

    Thirty minutes to find out whether you have three workstreams worth building, whether you already know which three, and whether the access to build them is gettable. That conversation is what sets the quote.

  2. 2

    Sign, and the NDA goes with it

    SOW and NDA together, so access work starts immediately instead of waiting on legal while the clock runs.

  3. 3

    Build one: intake, spec, build, test, accept, train

    The full sequence on the first workstream, roughly a month. The first one also calibrates the other two, because what your team flags in testing changes what we scope next.

  4. 4

    Builds two and three, same sequence

    Each with its own intake, its own definition of done and its own named owner. Roughly 90 days end to end for all three, and we tell you early if one is running long instead of at the end.

Starting at $25,500

Why the price is a starting point

Three straightforward builds for a team that already knows what it wants and has its data in order is a different piece of work from three complicated ones where the standard lives in somebody’s head. Pricing it as one number would mean overcharging the first team to cover the second.

So here is what we are weighing when we quote yours. All of it comes out of the fit call and goes in writing before anyone signs.

Whether you already know what you want built

Arriving with the list is the floor. Most teams do not, and then the engagement opens with intakes across the groups that would use it, and a short written build plan naming which workflows to build and in what order. That costs more and it is usually worth it, because the most expensive build is the one nobody needed.

How complicated the builds are

A Skill that reformats a document against a house template is not the same work as one that reads four systems, applies judgment and has to be right every time. We price per workstream, and we tell you which of yours is the hard one.

Whether your data is clean and reachable

A build is only as good as what it reads. Where the source is inconsistent, spread across systems or needs work before anything can use it, that work is real and it goes in the scope instead of being discovered in week three.

Whether your systems, tools and workflows are written down

If the standard a person applies by hand today exists as a document, we build against it. If it lives in one person's head, we are deriving it from examples and checking it back with them, which takes longer and carries more risk of building the wrong thing well.

How many builds

Three is the floor. Teams with a longer list run more, and the per-build price comes down as the count goes up because the access and setup work is done once.

Whether we can get the access we need

NDA, a provisioned seat, and permissions on whatever the workflow touches. Where that is straightforward it costs nothing. Where it needs security review, a procurement process or a workaround, it is work and we scope it as work.

Before anything gets built

What we need from you

This list is on the page, not buried in a kickoff deck, because it is where build engagements go wrong. On our largest one, access and NDA constrained the entire build phase, and the client sponsor ended up proposing contractor status for us to solve it. Worth knowing in week zero.

A signed NDA, and then real access

To the tools, data and systems the workflow touches. Usually a provisioned seat and whatever permissions it needs. This is the step that decides whether a build is production or a demo, and it is the one that most often stalls an engagement, so it starts on day one.

A named owner for each build

One person who agrees the definition of done and signs the acceptance. Usually whoever will own the workflow afterward, not whoever booked the call.

Three or four real examples

Of the input, and of what a good output looks like. This changes how close the first version lands more than anything else on this list.

The standard it has to match

The template, the house format, the rules a person applies by hand today. Where no standard exists yet, that becomes an open decision with an owner and a date, never a guess we make for you.

Material your team can test on

Synthetic, or work that is closed and announced. Testing never runs on live client matters.

Which of these you want depends on how many, and on who learns

Three engagements on this site end with Claude doing work it was not doing before. They differ on who does the building and how much of it there is.

This page · from $25,500

The Build Series

Three workstreams, built by us, over roughly 90 days, each one trained into the team after it is accepted. Where you do not yet know which three, it opens with intakes and a build plan.

Quoted per workstream

The Build Sprint

One workstream, same sequence, delivered and handed over. No training attached. The right size when there is a single thing worth building and you want to see how we work before committing to more.

From $9,500

The Custom Program

Your team builds, with us in the room, and keeps the capability. Cheaper, slower to a finished artifact, and the only one of the three that leaves people able to build the next thing without us.

Nicole Patten

Ex-Google Senior UX Engineer. 7 years building products used by millions. Now she builds custom Skills, orchestration systems, and automations that turn Claude into your team’s operating system - not just a chatbot. Nicole’s neuroscience background shapes how she builds adoption, not just technology.

Frequently asked questions

Who does the building?

We do. This is the offer where the work gets done for you and handed over running, which is the opposite of <a href='/custom-program'>the Custom Program</a>, where your team builds with us in the room and the point is that they learn to do it. Teams who want both usually run a program first and this after.

What if we do not know which three we want?

Most teams do not, and that is a normal place to start. The engagement opens with intakes across the groups who would use this and a short written build plan naming which workflows to build and in what order. It costs more than arriving with the list and it is usually worth it, because the most expensive build is the one nobody needed. If what you want is the firm-wide version of that, ranked across every department with ROI framing, that is the <a href='/rollout'>90-Day Rollout</a>.

Why is the price a starting point?

Because three straightforward builds for a team that already knows what it wants and has its data in order is a different piece of work from three complicated ones where the standard lives in somebody’s head. The things that move it are listed on this page. We quote the whole scope on the call, in writing, before anyone signs anything.

How is this different from a Build Sprint?

Count, commitment and what happens at the end. A <a href='/build-sprint'>Build Sprint</a> is one workstream, quoted per workstream, delivered and handed over. This is three on a fixed price over roughly a quarter, and each one is trained into the team after it is accepted. If you only have one thing worth building, buy the Sprint.

What do you need from us?

A named owner per build who agrees the spec and signs acceptance. A signed NDA. Access to the tools, data and systems the workflow touches, which usually means a provisioned seat and whatever permissions it needs. Three or four real examples of the input and what a good output looks like. And the standard it has to match: the template, the house format, the rules a person applies by hand today. The last one changes how close the first build lands more than anything else on this list.

What if we cannot give you access to the real systems?

Then say so on the fit call, because it is the single most common thing that stalls a build engagement and it is better found in week zero than week three. Some of it has workarounds, synthetic material or closed and announced work for testing. Some of it does not, and where it does not we would rather scope around it or tell you this is the wrong engagement.

What if a build takes longer than expected?

You hear about it when we know, not at the end. The number of fix-and-retest rounds included is named in the scope before we start, so the thing that usually makes a build overrun is bounded in writing. Where a workstream turns out to be bigger than the engagement, we scope the rest separately instead of quietly delivering less of it.

What if something does not ship?

The definition of done is agreed before building, so there is a written standard to measure against instead of an argument. Where a build cannot meet its bar for a reason outside the work, usually access or a platform limit, we say so, document it, and swap the workstream with your agreement.

Why train the team at the end? We bought builds.

Because the alternative is measurable and we have measured it. On our largest engagement ten things were built and four were enabled for everyone, and the gap had nothing to do with build quality. A workflow that only its owner knows how to run is a single point of failure you paid for.

How is this different from the Rollout?

This builds three things for one team. The <a href='/rollout'>90-Day Rollout</a> trains every department, builds across them, and adds governance, a written roadmap of what to build next, and measured before-and-after adoption numbers for leadership. Anyone who is going to be asked to prove adoption with a number should read the Rollout page first.

What happens if Claude is down on the day of a session?

Claude occasionally experiences outages or degraded performance. If Claude is unavailable or significantly impaired on a scheduled session date, we’ll reschedule that session to the next mutually available time at no additional cost - and it doesn’t count against any reschedule limit.

What if our team’s Anthropic accounts get suspended?

Anthropic occasionally suspends or restricts accounts for reasons that aren’t always clear or contestable. Elevate Online doesn’t control Anthropic’s policies and isn’t responsible for account actions. If account issues prevent a session from running, we’ll work with you to reschedule or adjust scope in good faith.

Name the three workflows.

Thirty minutes to find out whether you have three worth building, whether you know which three, and whether the access is gettable. Starting at $25,500.