Clipto raised $15 million at a $250 million valuation after building AI search for files people already have. The product points to a practical priority for companies planning their AI work: make existing material usable.
Start with material people already need: meeting recordings, project folders, customer notes, and training documents.
Check retrieval before adding more output
Before adding generation to a workflow, check whether employees and approved AI systems can locate the source files and earlier decisions they need.
The signal inside the Clipto raise
TechCrunch reported on August 31 that Clipto has been used by more than 30 million people, has hundreds of thousands of paying subscribers, and reached $15 million in annual recurring revenue at the start of 2026. The product indexes videos, audio, images, meetings, and other local files so users can search by describing what they need.
A team may have the answer scattered across call recordings, proposals, transcripts, and versioned files with names like final-final-v7. Generating another document adds little if people cannot locate the useful work already there.
Clipto’s founder told TechCrunch that content is already abundant. Much of it sits unused.
When planning a drafting or summarization project, check whether the team can retrieve its source material. A faster draft still needs the relevant decisions and examples.
Search is becoming part of the work itself
Deloitte’s 2026 State of AI in the Enterprise report says worker access to AI rose by 50% in 2025. It also says only 34% of surveyed organizations are using AI to deeply remake products, services, or core processes, while 37% are using it at a more surface level with little or no change to existing processes.
Consider a hypothetical client handoff. Relevant context sits in a call recording, a support note, and a proposal from nine months ago. If the person preparing the handoff cannot find them, the assistant may draft an answer that misses an earlier commitment or an approved exception. Retrieval belongs in the handoff design.
Test whether access to approved source material improves the handoff draft. Keep the source references so a reviewer can check the answer.
Start with a workflow and its source material
Pick one workflow where people regularly recreate work because the old context is buried. Sales handoffs, onboarding, and legal review are possible starting points. Choose based on the recurring retrieval problem: people know the answer probably exists somewhere, but finding it takes long enough that they stop trying.
Then inventory the source material around that workflow:
- Which files contain the real context?
- Who is allowed to access them?
- What should never be exposed to an AI system?
- What questions does the team ask repeatedly?
- What would a good retrieved answer include?
Use the inventory to test retrieval against questions the team actually needs answered.
Clipto is interesting because it keeps the user’s indexed material local and requires an active request and authorization before an AI application accesses it, according to TechCrunch. Check how those authorization controls fit the chosen workflow. Not every assistant should see every file just because the company bought an AI platform.
Test retrieval with the people doing the work
Ask the team to retrieve a known example, such as a client document from last quarter. Check whether the system returns the correct file or clip, whether the answer points back to it, and whether the requester is allowed to see it.
Assign an owner for source quality and permissions. That person needs a way to remove outdated material, correct a misleading result, and check whether sensitive files remain restricted.
Before expanding the trial, record which recurring questions retrieval answered correctly and where people still had to search manually. Use that record to decide what to index next.