Custom SaaS development agency: secure and ready to sell

Many SaaS products start out as internal tools, before being offered to other companies.

At Kreante, a first version and a full SaaS are both priced on a custom quote, based on a scope agreed together in writing. The budget depends mainly on five factors:

After launch come running costs: hosting, which grows with the number of customers, payment provider fees, email sending, support tools and maintenance. It is best to factor them into your subscription prices from the start.

The mockup arrives within 48 hours, during scoping. A first version, enough to welcome your pilot customers, then takes a few weeks, and a full SaaS, with advanced billing and integrations, a few months. The longest step is often finding the first customers: that is why we encourage you to show them the project from the mockup stage.

Four models come up most often:

• per-user pricing, easy to understand, which grows with your customers’ teams;
• tiered plans, from the basic offer to the most complete one, to address customers of different sizes;
• usage-based pricing, indexed on transactions, documents processed or tickets sold;
• freemium, a limited free version that makes people want to upgrade to the paid offer.

The right model is the one your customers understand immediately and that tracks the value they get from the software. It can be tested and can change: your SaaS should therefore let you change your plans without redeveloping billing.

The question has been stirring the industry since AI agents learned to produce tools on demand. Generic software, easy to reproduce, is indeed under pressure. A SaaS built on precise business expertise, on data only you control and on a relationship of trust with your customers, on the other hand, remains hard to replace. AI then becomes an extra asset: built into the product, it makes your software more useful and speeds up its development.

Yes, provided you start with an audit. A tool designed for a single company often mixes its data, rules and settings into the code. To open it up to other customers, you need to separate each customer’s data, make the settings configurable, add billing and strengthen access security. Depending on the state of the code, this transformation goes through a gradual evolution or a partial rebuild. A prototype built with an AI tool requires the same approach, with particular attention to security.

• Personal data protection: the software usually processes its customers’ data on their behalf. It then acts as a processor under the GDPR and must, in particular, govern this processing with a contract with each customer, ensure data security, keep a record of processing activities and assist its customers in the event of a data breach. The CNIL (France’s data protection authority) details these obligations in a dedicated guide^[2];
• The EU Data Act: since 12 September 2025, cloud service providers, SaaS included, must allow their customers to switch provider and recover their data. Fees charged for switching will be entirely prohibited from 12 January 2027^[3]. Adjustments for SMBs are being discussed at European level.

Planning a complete data export from the design stage meets these requirements, and reassures your customers when they sign.

Several profiles coexist: studios entirely dedicated to SaaS, product agencies that support startups, and teams that rely on AI to deliver faster.

Motives Solutions, team motivation measured in three questions. Vincent Legros and Bénédicte Levier wanted to make a scientific approach to motivation, based on Professor Michael J. Apter’s reversal theory, accessible to every team. Kreante developed a bilingual platform: each employee answers three anonymous questions, and the manager receives a report that identifies their team’s engagement levers and helps them act on disengagement or stress.

TheFlow, a ticketing platform that evolved its business model. See the case study presented above.