Working with Prolego, engagement practices and responsibilities

Prolego is a boutique specialist that has worked exclusively in AI since 2017. Our small team of Forward Deployed Engineers (FDEs), supported by engagement and engineering leadership, works directly with your experts to build and operate reliable AI inside your environment. Our experience explains who we are and where our point of view comes from.

Start with shared principles and ambition

The first consideration is whether you share our principles, especially our belief that AI will continue improving at an exponentially fast rate. We view investment choices through that lens: what should we build now, how should we build it and what should we avoid as the technology advances?

The second is what you want to accomplish. We work with leaders who want to re-engineer the economics of their business with AI. We help identify the larger opportunity, choose where to invest, deliver useful systems and develop the organization's ability to scale and adopt them. If you simply want someone to carry out an engineering assignment, we're the wrong company.

You should gain a working system, a methodology, an AI-native architecture and experience working as an AI-native product team. The aim is a business that can keep expanding what it does with AI.

Keep your options open

You can keep Prolego operating and extending the AI or take it over with your own team. You do not have to make that decision before we start. We build in your environment and preserve your ownership of the work so both paths remain open as the capability and your team develop.

Getting started: find the business opportunity

Look at where your business makes money, where it spends money and its major workflows. Send your company's name and a sentence or two about what you're considering, by email or through our contact form. We use AI tools and our experience to research the business, explore the larger opportunity, estimate potential ROI and identify likely headwinds before the first call.

The scoping approach explains how we develop that opportunity with you. The fit guide explains when Prolego is the right partner.

Planning delivery: structure the project for success

Once we choose a starting opportunity together, we plan how to deliver it. We work with you on the business outcome, scope, first 90 days, staffing, expert participation, access and responsibilities. The proposal records the agreed approach and commercial terms.

This planning turns an opportunity into a project designed to succeed. Prolego leads the process with your team; you do not need to work out the delivery plan before contacting us. PDD explains how we organize the engineering around reliability from the start.

A small team with senior responsibility

The team covers three responsibilities:

Engagement leadership: business priorities, product direction, strategy and operating decisions.
Engineering leadership: architecture, methodology, quality and the production approach.
Forward-deployed engineering: our AI systems engineers work as Forward Deployed Engineers (FDEs), directly with the client's experts and technical teams. They help shape what to build and develop and operate the workflow across sources, prompts, tools, evaluation and interfaces.

Staffing and allocation follow the work and are specified in the proposal. These are engagement responsibilities; Prolego's historical company size of five to 15 people is a different measure. Judge us by the value delivered for the money spent, rather than the number of engineers or hourly rates.

Client participation during delivery

As we structure the project together, we agree how your team will contribute and who will cover these responsibilities.

Client responsibility	What we need
Executive sponsor	Owns the desired business outcome, KPI and investment decisions.
Business owner	Defines priorities, provides the baseline and settles operating disagreements.
Dedicated subject-matter expert	Supplies real examples, defines correct outputs and gives prompt feedback from the start.
Technical operations counterpart	Coordinates access, approved tools, security, infrastructure and release requirements.
Production owner	Owns continuing deployment, support and incident responsibilities alongside the AI team.
Representative work	Inputs, sources and completed outputs that reflect ordinary work and consequential exceptions.

Approved Workbench and coding-tool access must be available for the agreed work. The client supplies frontier-model and coding-tool access and pays model usage and infrastructure costs. Production access is needed before deployment; an approved development source set can support the beginning.

Prolego prepares concrete outputs and focused questions in formats experts already use, including Word, Excel and Teams. We minimize review effort while preserving the client's judgment. What sets the pace of an AI project explains why expert feedback and organizational agreement usually determine deployment speed.

What happens at 30, 60 and 90 days

The proposal sets the scope and prerequisites. Our planning sequence is:

Checkpoint	What the client should see
First-month goal	A performance report, working workflow and evaluation framework that let major stakeholders inspect expected behavior, potential impact and remaining risks before deciding where to invest next.
Around day 60	Refined workflow and controlled release to an initial group as access, performance and approvals permit.
Around day 90	An operable production system for the defined task, followed by tuning and expansion from real use.
Continuing work	Production review, source and guidance upkeep, evaluated improvements, model comparisons and the next useful applications.

For a six-month engagement, the work can run in two phases with a program review at 90 days. Milestones describe the intended delivery path; the proposal identifies the client's access, expert and release dependencies.

Pull risk forward so you can choose the next investment

The first-month performance report helps major stakeholders decide whether to continue, adjust the approach or redirect Prolego to another opportunity. Pulling AI risk forward explains how we make expected behavior, potential impact and remaining risks visible early.

Production begins with a defined scope

We launch useful work with human review of critical decisions, then improve it in days and weeks from real feedback. Launch and improve explains the production-learning approach; PDD describes the engineering loop.

Keep the business running during the transition

The current process keeps running while teams move to the new workflow. We preserve the operating continuity the business needs and retire the previous process once the new one is ready.

Fixed team, fixed monthly fee

We generally work on a retainer for a fixed team and monthly fee. Priorities can change without a new change order for every request. The team keeps finding ways to use AI and the shared foundation to deliver more, while operating applications already in use.

We agree priorities together and make capacity and dependencies visible. A new request may move another item in the sequence or require a specialist outside the team. The arrangement supports adaptation throughout the engagement.

For budgeting purposes, $250,000 to $500,000 is a good starting point for an engagement. Specific pricing and terms are discussed privately. The proposal names the starting result, team, sequence, client contribution and commercial terms. We do not promise predictable ROI; pulling risk forward explains how we help clients make informed investment decisions early.

Ownership and exit

Our work lives in the client's environment from day one: code, repositories, prompts, workflows, guidance, evaluations and documentation. The client owns what we create, apart from open-source libraries. Third-party products and model services retain their own licenses.

Cancellation terms are negotiated with the price and commitment. A cancellation period is often 30 days; another arrangement may use 60 days or a different agreed period. The contract sets the terms. The client retains completed work, and ownership and documented interfaces support changing providers or platforms. The first-month performance review serves a separate purpose: understanding risk and choosing where to invest next.

After launch

Keep Prolego operating and extending the systems or take them over with an internal team. Our case collection includes both paths: a regulatory-services client took over after three months; a large bank took over specialist components before production, as planned.

Model and supporting infrastructure costs are usually modest for the document, policy, record and question workflows we build. High-throughput workloads need their own estimate. The proposal should show the expected usage and operating costs.

Before a handover, identify who maintains sources and guidance, reviews exceptions, approves releases and responds to incidents. Repositories, evaluations, operating context and documentation stay with the client.

Plan for the pace of organizational change

What sets the pace of an AI project explains the four conditions behind the schedule, the lessons from delivery and how we address likely delays. Investing for compounding value explains how the same team extends the foundation as the organization gains experience.

Contact Prolego with the business change you want to make and the work that currently constrains it.