How to retire a SaaS tool without losing data or work

Contents

1. Why retirement goes wrong
2. What the law now says about leaving
3. The retirement sequence
4. A checklist for the final month
5. Questions buyers ask
6. Sources

Deciding that AI or an owned system can replace a SaaS tool is the easy part. Retiring it cleanly, with the data exported, the work moved and the contract ended at the right date, is where most of the effort and most of the risk lie.

Why retirement goes wrong

Companies rarely retire software deliberately. Zylo reports that organisations continue to uncover unused seats, and that the average application count barely moved in a year even as spending rose⁠12026 SaaS Management IndexZylo. Tools linger because nobody owns their exit, data is scattered across them, and renewals arrive before anyone has checked what would break.

Klarna's experience is the large-scale version. Its chief executive said the company had shut down Salesforce and around 1,200 other SaaS platforms after finding its data fragmented across them, and that the work started with consolidating that data, not with switching the tools off⁠2“No, I don’t think it is the end of Salesforce”: Klarna CEO clarifies why it stopped using SalesforceITPro · 24 Mar 2025.

What the law now says about leaving

In the European Union, the Data Act has applied since 12 September 2025 and sets a framework for customers to switch between providers of data-processing services⁠3Data ActEuropean Commission. The Commission's explainer states that Software-as-a-Service providers must export data in a commonly used, machine-readable format, and that switching charges, including data-egress charges, are removed entirely from 12 January 2027; until then they were allowed only as a transitional measure⁠4Data Act explainedEuropean Commission. Outside the EU, the contract governs, which makes the export clauses worth reading before each renewal.

The retirement sequence

Seven steps, in order

• Name an owner for the retirement, with authority to set dates.
• Map the tool's work and data: who uses it, for what, and what it holds, including attachments, history and custom fields.
• Map its connections: which other tools read from or write to it.
• Stand up the replacement for the same work, with the data it needs.
• Run both in parallel for at least one full business cycle.
• Export everything and verify it, then archive the export where the business can find it.
• Cancel at the renewal date, remove the seats and close the integrations.

Step by step

Map the work, not only the data. A support tool holds tickets, but it also holds routing rules, canned replies and reporting definitions. A CRM holds contacts, but also pipeline stages and automation. Write these down; they are the parts most often lost.

Choose the replacement by what it does with the work. If an AI agent is taking over, check how it is priced and what it covers. Intercom's Fin, for example, can sit on the existing helpdesk and is billed per outcome⁠5Fin pricingIntercom, which means the helpdesk itself is not retired. A consolidated system goes further: OpenLazarus's builder shows which tools a design replaces and what plan it needs⁠6Build your OSOpenLazarus, and its FAQ states that the code and data in the new system belong to the customer⁠7FAQOpenLazarus.

Run in parallel with a measure. Pick one number the old tool's work produced (tickets resolved, invoices sent, orders processed) and compare it across the two for the parallel period.

Export twice: once at the start of the parallel run to test the migration, and once on the final day to capture everything since.

Cancel deliberately. Cancel at the renewal date rather than mid-term, confirm in writing that data will be deleted after export, and remove single sign-on access and integrations on the same day.

A checklist for the final month

The end state is a smaller, better-documented software estate: the work runs in the replacement, the old data sits in a known archive, the seats are gone from the bill, and the next tool on the list can be retired the same way. Our guide to which categories AI replaces first helps choose that next tool.

Questions buyers ask

Sources

1. 2026 SaaS Management Index. Zylo. Accessed 8 Nov 2026.
2. “No, I don’t think it is the end of Salesforce”: Klarna CEO clarifies why it stopped using Salesforce. ITPro, 24 Mar 2025. Accessed 8 Nov 2026.
3. Data Act. European Commission. Accessed 8 Nov 2026.
4. Data Act explained. European Commission. Accessed 8 Nov 2026.
5. Fin pricing. Intercom. Accessed 8 Nov 2026.
6. Build your OS. OpenLazarus. Accessed 8 Nov 2026.
7. FAQ. OpenLazarus. Accessed 8 Nov 2026.

Updated 8 Nov 2026: First published.

Further reading

• Analysis · Updated quarterlyWill AI kill SaaS? What the evidence shows in 2026In February 2026 investors wiped about a trillion dollars off software companies in a fortnight. Companies, meanwhile, kept spending more on SaaS. Both facts are true, and together they show what AI is actually doing to rented software.
• Guide · CategoriesWhich SaaS categories AI replaces firstExposure depends less on the vendor than on the work the software supports. Routine, observable, rules-based work goes first; regulated, knowledge-heavy work last.
• Guide · CostsSaaS seat costs vs owning your softwarePer-seat pricing made software easy to buy and hard to shrink. AI has added a fourth way to pay. Here is how the four compare on the numbers vendors publish.