This guide is for buyers about to hire a consulting partner. It belongs to our series on choosing a delivery partner, which starts with large or boutique vendors, and shows how pre-code work feeds custom software development.
In this guide
- What does a software development consulting company actually do before development?
- Why does the work before code decide the price and the risk?
- Five pre-code steps and what each must hand over
- How should the pre-code work run, step by step?
- When do you need paid discovery, and when not?
- What does a fixed price buy, and what does it move into change control?
- Where AI fits
- What delivery record exists, and what does not
- Alternatives: other ways to start a software project
- Limits of this guide
- Frequently asked questions
- Next step
What does a software development consulting company actually do before development?
A consulting company's first job is to decide what should be built, and whether anything should be built at all, so that the expensive part of a project starts from agreed problems, decisions and tests rather than a feature wish list.
That work has five parts, from problem framing to a milestone proposal, and each ends in a document. With them, a buyer can compare partners and, if needed, take the work elsewhere.
Netbase treats this as the first step of its six-step delivery lifecycle, which runs from discovery and strategic alignment through team assembly and architecture planning to agile execution with outcome-based milestones, rollout and ongoing support. Consulting, solutions, custom development and AI-enabled digital transformation are Netbase's lead offers, which is why the lifecycle starts with consulting rather than with staffing.
Why does the work before code decide the price and the risk?
A fixed price is only as reliable as the scope underneath it. When discovery is skipped, the proposal prices a guess, and the gap between the guess and the real need comes back later as change requests, disputes or an unfinished system.
The GOV.UK Service Manual applies the same logic to public services: discovery is for understanding users, constraints and the problem, not for building, and stopping after discovery is not a failure if the research says so. A partner that never suggests stopping is selling development, not consulting.
The pre-code documents also become the contract. Our guide to scope-creep risk in a proposal shows the clause wording to check once the proposal exists.
Five pre-code steps and what each must hand over
If a step's document is missing, ask the question in the last column before you sign.
| Step | What the consulting company does | Document you should receive | Ask if it is missing |
|---|---|---|---|
| 1. Problem framing | Turns the request into a business problem, an owner and a measure of success | One-page problem statement with the decision it supports | Which business result is this project for, and who owns it? |
| 2. Discovery | Interviews users, maps the current process, data and systems, lists constraints | Discovery findings: users, process map, integration list, risks, open questions | What did you learn that changed the original request? |
| 3. Architecture decisions | Chooses the system shape, integrations, hosting and data model, with the reasons | Context diagram plus architecture decision records with trade-offs | Which options did you reject, and why? |
| 4. Scope and acceptance criteria | Splits the work into features and writes how each will be accepted | Scope document with acceptance criteria and an explicit out-of-scope list | How will we both know this feature is done? |
| 5. Milestone proposal | Groups the scope into milestones, each with an outcome, acceptance criteria and a payment point | Proposal with milestones, assumptions, change-control rules and IP terms | What happens to the price if an assumption turns out wrong? |
An architecture decision record captures one decision and its rationale, so a later team sees the trade-offs instead of guessing. The C4 model draws the system at context, container, component and code level, which keeps architecture readable for business owners.
How should the pre-code work run, step by step?
-
Agree the problem and the decision
Write the business problem, the owner and what success changes, on one page.
-
Run discovery with real users
Talk to the people who do the work today, trace the process and list every system the software must read or write. Record open questions rather than hiding them in assumptions.
-
Decide the architecture and write the reasons down
Draw the system context, choose the integration and data approach, and keep one decision record per significant choice, including the options rejected.
-
Write scope with acceptance criteria
Turn each feature into a testable statement and add an out-of-scope list; the Scrum Guide's Definition of Done is a useful model.
-
Price the scope as milestones
Each milestone gets a written outcome and acceptance criteria agreed before it starts, a payment point on acceptance and the assumptions behind it. Unknowns that cannot be priced yet become a small paid spike, not a padded estimate. For a product that will grow, plan milestones per lifecycle stage.
-
Review the package together before signing
Anything either side reads differently is fixed in writing now, when it costs least.
When do you need paid discovery, and when not?
Not every project needs the full sequence; the depth depends on how much is already known.
| Situation | Recommended pre-code work | Why |
|---|---|---|
| New product or new business process, many open questions | Paid discovery first, then a fixed-price proposal for the build | The scope does not exist yet; a proposal without discovery prices a guess |
| Replacing or integrating existing systems | Discovery focused on data, integrations and migration, then architecture decisions | Most risk sits in the systems you already run |
| Clear, well-specified change to a live system | Short scoping call and a direct fixed-price proposal | The specification already does discovery's job |
| Throwaway prototype to test an idea | Skip formal architecture; agree a time-boxed prototype | Decision records are wasted on code you will discard |
| AI feature whose quality is uncertain | Discovery plus a paid spike that tests the model on real data | Acceptance has to be measured before it can be priced |
For AI features, see fixed-price AI projects. A product discovery workshop usually starts a new product; solution architecture consulting covers step 3 alone.
What does a fixed price buy, and what does it move into change control?
A fixed price buys certainty for the scope and acceptance criteria written down, not unlimited change. Anything outside the scope document moves into change control, where it is estimated and agreed separately.
That is fair only when the scope is good. An estimate is a sales number, not consulting output, when it has a single total with no milestones, no out-of-scope list, no assumptions, or acceptance described as "client satisfaction". Each hands the risk back to the buyer.
When the roadmap is still moving, another model may fit; our comparison of dedicated teams and fixed price and the engagement models page describe when each fits.
Where AI fits
AI speeds up the documentation-heavy parts of pre-code work: it summarises interview notes, drafts process maps, proposes acceptance criteria and checks a scope document for gaps. Consultants and engineers still decide what the findings mean, choose the architecture and confirm that every acceptance criterion is testable and agreed. Netbase works with the major commercial and open-source AI tools and models, chosen per project, and integrates AI into every service, including consulting.
When a proposal mentions AI, use our AI-augmented vs AI-native checklist; AI-native software development covers contract terms.
- In delivered work
RAG knowledge assistant
A retrieval-augmented knowledge assistant delivered for a client that is not named.
- Available capability
AI-assisted discovery and scoping
AI drafts of findings and acceptance criteria held to consultant review; not yet tied to a published case with figures.
What delivery record exists, and what does not
Netbase's lead offers are consulting, solutions, custom development and AI-enabled digital transformation; dedicated teams and outsourcing are secondary delivery options. The delivery lifecycle begins with discovery and strategic alignment, followed by team assembly and architecture planning, agile execution with outcome-based milestones, modular components, rollout and ongoing support.
Most Netbase projects are delivered on fixed-price contracts; milestone-based, monthly team retainer and KPI-linked models are also offered. In Netbase's outcome-based milestone model, each milestone has a written outcome and acceptance criteria agreed before it starts. For custom development, the client owns the IP created for it; Netbase's productized modules are licensed, not transferred.
The delivered AI feature named with a client is 4over4's recommendation engine, which suggests products from browsing history and previous purchases. Beyond it, Netbase has delivered AI projects for clients that are not named, including retrieval-augmented knowledge assistants, document AI and MLOps pipelines.
What does not exist: no published figure for how often discovery changed a Netbase project's scope, no published comparison of fixed-price outcomes with and without discovery, and no named client for the anonymised AI projects. This page describes a method, not results.
Alternatives: other ways to start a software project
| Alternative | Strength | Risk | Choose it when |
|---|---|---|---|
| Write the specification in-house, then request proposals | You keep full control of scope | Vendors price your gaps differently, so proposals are hard to compare | You have a business analyst and an architect on staff |
| Hire a time-and-material team and discover while building | Starts fast, adapts easily | Total cost is open; acceptance is harder to define | The product direction will change every few weeks |
| Pay one firm for discovery, tender the build | Independent scope you can take to several vendors | Handover loses context; two contracts to manage | Procurement rules require competition on the build |
| One consulting partner from discovery to delivery | Context carries through; one owner for scope and result | You depend on that partner's discovery quality | You want a fixed price backed by the team that wrote the scope |
Whichever route you choose, insist on the five documents; they make the work portable. Digital transformation consulting describes Netbase's consulting-first route.
Limits of this guide
- It describes good practice and Netbase's registered approach; it names no companies and gives no prices or durations.
- Legal terms such as IP transfer and liability need review by your own counsel.
Plan the next step with a Netbase consultant
Frequently asked questions
You should receive a problem statement, discovery findings, an architecture overview with decision records, a scope document with acceptance criteria and an out-of-scope list, and a proposal that splits the work into milestones with assumptions and change-control rules. Together these let you compare partners and check every claim before you sign.
It depends on how well the request is specified. A clear change to a live system can go straight to a fixed-price proposal, while a new product or a replacement of existing systems usually needs discovery, because the risk sits in unknown users, data and integrations rather than in the features you already listed.
It can, but the price then covers a guess, and the gap usually returns as change requests. A fixed price is reliable when scope, assumptions and acceptance criteria are written down first.
That depends on the contract, so check the IP clause before signing. At Netbase, the client owns the IP created for its custom development, while Netbase's productized modules are licensed rather than transferred.
Next step
Bring the request or the proposals you are comparing, and Netbase will go through the five pre-code documents with you: book a solution review.
Related services and solutions
AI-enabled custom software development built around your workflow Netbase provides custom software development for companies whose workflow no off-the-shelf product fits, from discovery through architecture, build, testing and handover, with AI where it removes manual steps and the IP left with you. Since 2012 Netbase has worked with 500+ businesses across 80 countries; according to the published USticker case study, design-tool users converted 40% more often.
Learn More
AI-enabled digital transformation consulting that ends in a delivery plan Netbase provides digital transformation consulting for mid-market companies that need a prioritized roadmap with a delivery path, not a deck. We rank every change by value and risk, including where AI removes manual work, and the same team builds it. According to CEO Huy Nguyen (David), Netbase has served 500+ businesses across 80 countries since 2012.
Learn More
Product discovery that validates AI features and ends in a release roadmap Netbase provides product discovery workshops for founders and product leaders to turn a product idea into validated scope, architecture options and a release roadmap. We test the problem with real users, prototype the riskiest flows, check whether planned AI features have the data and accuracy they need, and agree what the first release must prove before committing a build budget.
Learn More
Solution architecture for AI-ready systems that settles trade-offs before build Netbase provides solution architecture consulting for CTOs and product owners to define how systems, data, integrations and AI services fit together before build, with every trade-off documented. You receive a target architecture, decision records and an integration map that any competent team can build from, ready for AWS, Google Cloud, DigitalOcean or Cloudflare.
Learn More
Work with Netbase remotely
Every Netbase engagement is delivered remotely, so clients in any country and any industry can work with us; Netbase has worked with clients in 80 countries. Contracts and invoices are issued by Netbase JSC, and payment is made by bank transfer to Netbase JSC's company account or PayPal Business. Talk to the CEO, Huy Nguyen (David), on WhatsApp or by email.
Discuss a project
Netbase JSC helps organizations design, build, modernize, and operate digital products and AI-enabled business systems.project@netbasejsc.com
+84 937 869 689
91 Nguyen Chi Thanh, Dong Da, Hanoi, Vietnam
Get in touch
Tell us what you want to build, modernize, or operate.