R&D operations

The AI opportunity assessment: from ambition to R&D project in 4 weeks

A week-by-week methodology for turning a fuzzy AI ambition into a scoped, fundable R&D project, including what gets done each week, what the outputs are, and how it feeds into a Business Finland application.

Most companies arrive at an AI R&D funding conversation with ambition but without a project. They know they want to do something with AI, they have a commercial goal in mind, and they have a rough sense of the technology. What they do not have is a defined R&D question, a project structure, or a budget with any bottom-up logic behind it.

The AI Opportunity Sprint is a four-week structured process that turns that ambition into a scoped, fundable R&D project. This is the methodology, made public.

What the sprint is for

The sprint answers three questions before writing starts:

Is there genuine R&D uncertainty here, or is this implementation?
If there is real R&D, what is the specific question the project needs to answer?
What does the project structure, cost, and timeline look like, built from the bottom up?

Getting these right before writing is what separates applications that get funded from applications that don’t. Most rejections can be traced back to writing starting too early, before the project was actually shaped.

Week 1: technical assessment

Input: whatever the company has, an idea, a product problem, existing vendor conversations, prior internal experiments.

Activities: a structured session with the technical lead and, ideally, one external specialist familiar with the relevant AI domain. The session maps the actual state of knowledge: what the company wants the system to do, what existing approaches have already tried, what performance is needed for the outcome to be commercially useful, and where the gap is.

The goal is to locate the uncertainty honestly. Many companies discover during week one that what felt uncertain was actually implementation: the approach was known, the outcome was predictable, the work was just large. Others find the reverse: a real open question buried under optimistic language. The session is designed to force that distinction.

Output: a written technical assessment. What the system needs to do, what the baseline performance of existing approaches is, what the specific gap is, and why it matters commercially. If the gap is real, the assessment becomes the foundation of the R&D uncertainty section. If there is no gap, the sprint stops here and the company is told not to apply.

Week 2: scoping and work package design

Input: the technical assessment from week one.

Activities: the project is structured into work packages, coherent chunks of R&D each with an objective, a set of activities, and a milestone or deliverable that represents a genuine learning event. For a typical AI R&D project, packages often follow the arc of the uncertainty: baseline evaluation, method development, evaluation and hardening, integration and piloting.

Each package is scoped to be independently auditable: what was done, what was learned, and what it cost. This is important for both the application and for post-funding reporting.

Output: a work package structure with objectives, activities, and milestones for each phase. A first-pass timeline.

Week 3: budget and eligibility

Input: the work package structure.

Activities: the budget is built bottom-up from the work packages. Personnel is costed from realistic hours against each package, using actual loaded salaries. Subcontracting is identified and scoped. The indirect cost allowance is applied correctly. Eligible and ineligible costs are separated.

Common errors found and fixed during this week: personnel time that does not match the work package activities, subcontracting proportions that exceed the eligibility threshold, costs that overlap with the overhead allowance, and a total that bears no relation to the scope.

Output: a bottom-up budget with eligible cost totals, funding rate applied, and an estimated funding amount. A one-page cost narrative explaining the logic.

Week 4: application readiness

Input: the technical assessment, work package structure, and budget.

Activities: a review of the company’s eligibility (company size, financials, ownership structure), a check that the commercial narrative is connected to the R&D question rather than running parallel to it, and a review of any prior Business Finland relationship. If an early dialogue with Business Finland has not happened, this week includes preparing for it.

Output: an application-ready project brief: R&D question, work packages, budget, company eligibility summary, and a draft of the technical-commercial logic the application will need to make. This brief is either taken directly into writing by the company, or handed to a writer or consultant to develop.

From sprint to application

A company that finishes the sprint has everything a strong application needs, except the prose. The shaping work is done. The uncertainty is defined. The budget is grounded. The commercial link is clear.

What typically takes four to six weeks to write now takes two, because there are no structural decisions left to make during the writing process. That is where the sprint’s value shows: not in producing documents, but in making everything that follows faster and less likely to fail.

Companies that go to writing without doing this work first tend to discover mid-application that the project is not actually fundable, or to submit an application that is rejected because the evaluator cannot find the R&D question in it. The sprint exists to find that out before the writing starts.