Software your customers can figure out on their own.

SaaS, or software as a service, is software your customers sign into and pay for by subscription. We help founders and small software companies decide what to build, design it around how customers actually work, and ship it. That could be a first version, a redesign of a product that has grown hard to use, or AI features that do a real job.

Service
SaaS products
Timeline
Longer than a website, planned in phases with dates up front
Pricing
Fixed quote after a free audit
Built with
Figma, React and Next.js, TypeScript
  • Product strategy
  • MVP scoping
  • Customer research
  • Dashboard design
  • Design systems
  • Full-stack builds
  • AI features

Sounds familiar?

Signs it's time.

What a first version should and shouldn’t do

Most software businesses don’t stall because the code was bad. They stall because they spent their budget building features nobody asked for, before finding out whether anyone would pay for the core one. For a small team, the first version is the most expensive guess you’ll make.

An MVP, or minimum viable product, is the smallest version a real customer will pay for or commit to. It isn’t a cheaper copy of the full product. It’s a complete answer to one narrow problem. We start by naming the first customer, the job they need done and what they do today instead, which is often a spreadsheet and a lot of email.

Then we cut everything that doesn’t serve that job and write down why, so you can see each trade-off and make the final call. Where it helps, we put a clickable prototype in front of those first customers before writing the code for the real product, so the scope is tested by the people who would pay for it rather than argued over in meetings.

Designing screens people use all day

Business software has a different job from a consumer app. Your users open it every morning, work through long lists of records and need to trust the numbers. A dashboard that looks impressive in a sales demo can be slow and confusing by the second month.

For Deep6 AI, a platform for recruiting patients into clinical trials, we ran 18 research sessions and nine design iterations on its table and map views. That’s what this kind of product needs: watching people do their actual work, then testing changes with them before anything gets built.

What we run into most often in existing products:

  • Tables with no saved filters, bulk actions or sensible defaults
  • Charts that show data without answering a question
  • Empty screens on day one, with no hint of what to do next
  • Settings and permissions added wherever there was room
  • Five slightly different versions of the same button

A design system is what keeps a redesign from drifting back. It’s a shared library of components, colors, type and rules, kept in Figma and matched in code. Our design-systems work at bp came with more than 100 resource pages and over 25 workshops, because a system only holds when the people building with it understand it. Yours will be far smaller, with plain-language notes your developers can follow without a designer in the room.

Adding AI where it does a real job

Plenty of products now have a chat box in the corner that nobody uses. AI earns its place when it removes a step your customers already take: summarizing a long record, drafting a first reply, sorting incoming requests, or answering a question about their own data in plain language.

We start with that step, then design for the moments when the model is wrong. That means showing where an answer came from, letting people edit before anything is sent, keeping a person in charge of decisions that matter, and tracking what each request costs so your margins hold as usage grows. We build on major model providers through their APIs, the connections that let your product send requests to their models, and use settings that keep your customers’ data out of model training.

We test each AI feature on realistic examples before it reaches customers, and if a simple rule or a clearer screen would do the job better, we’ll say so. We use AI in our own work the same way: to speed up research, drafting and prototyping, with people reviewing everything that ships.

From marketing site to first session

A SaaS product is more than the app. Customers find you through a marketing site, make up their minds on a pricing page, sign up, and either reach a useful result in their first session or leave. We look at that whole path in the audit, because a strong product behind a confusing website still loses trials.

We can design your marketing site alongside the product so the two look and sound like one company, and plan launch campaigns when you’re ready for more traffic. If your customers also need a mobile app, it can share the same design system.

When the work is done, you own the code, the design files and every account. You can take it in-house, hire your own developers, or keep working with us in fixed-price phases. If you’re not sure where to start, a free audit gets you an honest read on your product or plan, and a concept to react to, before you commit any budget.

Who it's for

Founders scoping a first version

You have a problem worth solving, maybe a few pilot customers, and a long list of ideas. You need to decide what the first release does, what it leaves out, and how to get it in front of paying users.

Products that have grown messy

Years of features added one request at a time have left screens that don't match and customers who need a call to get started. You want a redesign that fixes the structure, not just the colors.

Business tools built around data

Your customers manage hundreds or thousands of records in tables, reports, maps and admin panels. They use your product all day, so small slowdowns cost them real time.

Teams adding AI to a product

You want AI to remove a step your customers already take, like summarizing, drafting or searching, rather than adding a chat box nobody opens.

What's included

Everything below is part of the work — not an upsell waiting at the end.

Product strategy and MVP scope

For a first version, or MVP (minimum viable product), we name the first paying customer, the one job they need done and what they use today instead. Everything that doesn't serve that job goes on a later list with a note on why, so the cuts are decisions rather than accidents.

Research with real customers

Interviews and usability sessions, where we watch people try the product, with the people who use or would buy it. We recruit them from your customer list or pilot users, and what we learn shapes the flows, the wording and the order we build things in.

Flows and clickable prototypes

We map the core flows, such as sign-up, setup and the main task, then build a clickable prototype that customers and investors can try. AI prototyping tools let us test ideas quickly, before anyone writes the code for the real product.

Dashboards and data-heavy screens

Tables with saved filters and bulk actions, charts that answer a specific question, map views, and detail pages that keep context. Each screen also gets its empty, loading and error versions, so a brand-new account or a slow report never shows a blank page.

A design system for your team

A shared library of components, colors, type and spacing rules, kept in Figma, where the screens are designed, and matched in code, with plain-language notes on when to use what. New features then look consistent without a designer on every ticket.

Full-stack build

Both halves of the software: the screens people see (the front end) and the server and database behind them (the back end), plus sign-in, user roles and permissions, subscription billing through Stripe, system emails and admin tools. Everything runs on hosting accounts in your name.

AI features with guardrails

Summaries, drafting, sorting and plain-language search, built on AI models from major providers. We design for the moments the model is wrong: answers show their sources, people can edit before anything is sent, and we track what each request costs.

Onboarding and first-run experience

The first session often decides whether a trial becomes a customer. We design sign-up, sample data, setup checklists and in-app hints so new users reach a useful result without booking a call.

How it works

Longer than a website, planned in phases with dates up front. You'll always know what's happening this week and what's next.

  1. 01Days 1-3

    Free audit

    We review your product, or your notes, deck and prototype if you're pre-launch, along with your marketing site, sign-up and onboarding. You get a clear read on what's working, what's getting in customers' way and where to focus first.

  2. 02Usually within a week

    Preview and quote

    We turn one screen into a designed concept: your marketing homepage, or a key product screen such as the main dashboard or the first screen after sign-up. Then we walk you through it in a short video. If you like the direction, you get a fixed-price quote and a dated plan, with no obligation.

  3. 03Phase 1

    Research and design

    We talk to your customers or the people you plan to sell to, map the core flows and design the product screen by screen. We test a clickable prototype with them before writing the code for the real product, and the design system takes shape here too.

  4. 04Phase 2

    Build and test

    We build feature by feature, so you have a working version to click through as we go. Sign-in, billing, permissions and any AI features are tested with realistic data before launch.

  5. 05After launch

    Launch and learn

    We launch to your first customers, watch how they use the product and agree with you what to change first. The next round of features is scoped and quoted the same way, based on what customers actually do.

How pricing works

Every SaaS project gets a fixed price, quoted after the free audit and preview once we both understand the scope. The main things that move the price are how many types of user and core tasks the product handles, connections to other systems, how much customer research it needs, and whether we're designing only or also building. Larger products are split into phases, such as research and prototype first, then the build, each with its own fixed quote, so you can change course between them. Third-party costs like hosting, AI usage and payment fees are billed to your own accounts.

Get a fixed quote →

What you have at the end

  • Source code in a repository you own
  • Hosting, database, payment and AI accounts in your name
  • Figma files for every screen and flow
  • A documented design system and component library
  • Technical notes on running, updating and handing off the product
  • Research findings and a prioritized roadmap

Tools & platforms

  • Figma
  • React and Next.js
  • TypeScript
  • Node.js
  • PostgreSQL and Supabase
  • Stripe Billing
  • Vercel
  • Anthropic and OpenAI APIs
  • Replit and Bolt.new

Questions about saas products

Something else? Ask us directly.

I only have an idea and some notes. Is that enough to start?

Yes. For an idea-stage product, the free audit becomes a scoping review: we read your notes, deck or rough prototype, look at what your future customers use today instead, and come back with what a first version should include. The preview then gives you a concept to react to before you commit to a build.

Can you work with our existing code and developers?

Yes. During the audit we look at your current product and, if you share it, the codebase. We can design and hand off to your team with a component library they can build from, or build alongside them in the same repository.

How long does a SaaS project take?

Longer than a website, and it depends on scope. After the free audit you get a dated plan that shows when you'll see a clickable prototype, when you'll have a working version to test and when you launch. The number of user types, core tasks and connections to other systems, and how much research the product needs, are what stretch or shorten it.

What kind of AI features make sense for my product?

Ones that remove a step your customers already take, like summarizing long records, drafting replies, sorting incoming requests or searching your data in plain language. We prototype the feature first, test it on realistic examples and design what happens when the AI gets it wrong. If AI doesn't fit your product yet, we'll tell you.

Who owns the code and the accounts?

You do. The code lives in a repository, the online home for your code, that you own, and hosting, database, payment and AI accounts are set up in your name from the start. If we part ways, there's nothing to transfer back; we hand over the technical notes your next team needs.

Can you build products that handle sensitive data, like health records?

Our Deep6 AI work was design for a healthcare platform used to recruit patients into clinical trials. When we build, people only see the data their role needs, and we choose hosting and AI providers that fit your requirements. We're not a compliance auditor, so if you need HIPAA (US health privacy rules) or SOC 2 (a security audit standard), we work alongside your compliance advisor and build to their checklist.

Pairs well with