Skip to content
Delsivo
Who it's for

Run every application window on your own portal

Open each application window on your own portal, give reviewers scoped access, and interview every shortlisted applicant on the same questions.

See white-label portals
Sample portal for a made-up programme, showing an open application window and its links
Spring cohort applicationsOpen until 31 March

Apply with a PDF deck. Save and come back any time before the window closes.

  • Public linkOn your website and social postsPilot
  • Partner referral linkFor investors and corporate partnersPilot
  • Alumni referral linkFor founders from past cohortsPilot
  • Scout linksOne per scout, with attributionLater

Sample programme. The name and address are made up.

Why accelerators

A whole cohort's applications in a few weeks

Accelerators get their applications in bursts, review them with a team that changes every cohort, and have to explain afterwards why some teams got in. That is a different job from a fund's steady inbound.

  • One window, one branded portal

    Publish the portal when applications open and unpublish it when they close. An unpublished portal can't receive applications, so late submissions don't slip in by email.

  • Many reviewers, the same criteria

    Your selection criteria are published as a rubric, so every reviewer sees each application checked against the same rules, with sources and reasons.

  • The same interview for every shortlisted team

    Shortlisted applicants pitch and answer the same questions in a live AI interview or a text interview, so selection rests on comparable answers instead of on who got a call.

The workflow

An application window, end to end

Each step with the release it arrives in. Your programme team makes every selection decision.

  1. Pilot

    Open the window on your portal

    Your logo, colours and copy on a subdomain such as yourprogramme.delsivo.app, with the closing date in your own words on the landing page.

  2. Pilot

    Hand out referral links

    Private referral links for partners, alumni and investors, so you know where each application came from. Dedicated scout links with their own tracking are planned for later.

  3. Pilot

    Screen against programme criteria

    Eligibility as hard rules, what you look for as weighted preferences, and Unknown values turned into questions for applicants.

  4. Pilot

    Split the work across your selection team

    Analysts work the queue on the deals they are given. Every override and decision carries a reason.

  5. First release

    Bring in mentors as guest reviewers

    Named guests see only the applications and files you share, until their access expires.

  6. First release

    Interview every shortlisted team the same way

    One question set, live or in text, with timestamped transcripts your reviewers can cite.

Delsivo is at the design partner stage. Pilot features come first, first release features follow, and later features are on the roadmap. AI suggests and cites its sources; your team makes every decision.

Many reviewers

Who reviews, and what each person sees

Selection involves your programme team, partners and often outside mentors. Each gets a role, and every page, search result, AI answer and export goes through the same permission check.

Reviewer roles for an accelerator
PersonRole in DelsivoWhat they seeRelease
Programme managerAdminSettings, branding, the rubric and every applicationPilot
Selection teamAnalystApplications they are given or permitted to seePilot
Managing directorPartnerEvery application, and approves decisions and messages to foundersPilot
Mentor or industry reviewerGuest expertOnly the applications and files shared with them, until access expiresFirst release
ApplicantFounderTheir own application, interview and requests, never ratings or notesPilot
Consistent interviews

Every shortlisted team answers the same questions

When some teams get a call with a partner and others get a junior reviewer, selection measures luck as much as the company. One live AI interview for the whole shortlist, taken whenever each team chooses, gives every team the same chance.

First release
  • One question set for the cohort, plus follow-ups drawn from each team's deck
  • A screen-share pitch, then the AI asks your questions out loud, within the cap you set. A text interview under the same rubric is always available
  • The AI assesses only the transcript, on substance and evidence; accent, fluency, polish, appearance and the video are excluded
  • Timestamped transcripts, so reviewers cite the moment they rely on
  • Your team decides who advances and can override any assessment with a reason
After selection

What you can explain after the cohort starts

Boards, funders and rejected applicants ask why. Everything needed to answer is on the record, written as the work happened rather than reconstructed later.

Why each team was shortlisted
Criteria met, mismatched and unknown, each with the slide or answer it came from.
Who decided, and when
Every stage change, pass and override is in the deal's audit history with the person and the reason.
What changed between rounds
Rubric versions are kept. Re-running a screen creates a new run and keeps the old one for comparison.
What applicants were asked
Every request and reminder sent through the portal, and what came back.
In practice

An eight-week application window

Sample timeline. Guest reviewers and interviews are first release features.

  1. Week 1

    Publish the portal and send referral links to partners and alumni. Each application is read and screened as it arrives.

  2. Week 3

    The queue shows each application against your eligibility rules. Teams outside your geography are marked Mismatch for a person to check, not rejected.

  3. Week 5

    Applications close. You unpublish the portal, and the selection team splits the queue by sector.

  4. Week 6

    Three mentors get guest access to the fintech shortlist, expiring at the end of week 7.

  5. Week 7

    Every shortlisted team takes the same interview. Reviewers cite timestamps when they recommend a team.

  6. Week 8

    Your programme team makes the selection. Each decision and its reason is on the record for the post-cohort review.

Fair questions

What programme teams ask us

That works until reviewers disagree and nobody can say why a team was cut. Delsivo keeps the deck, the criteria, the sources and each decision's reason together on one record, so selection can be explained after the cohort starts.

Mentors can join as guest experts who see only the applications you share, until access expires, in the first release. How guest reviewers record their views is being shaped with design partners. Selection decisions stay with your team.

Publish a new rubric version for each cohort. Earlier versions are kept, and re-running a screen creates a new run without overwriting the old one, so you can compare cohorts on the criteria that applied at the time.

PDF first. PowerPoint follows once extraction quality is measured on real decks. Applicants can export a PDF from any presentation tool, and you can say so on the form.

Delsivo is English first at launch. Applications in other languages can be submitted, but extraction and interview assessment are designed and checked for English.
Plan fit

Design partner pilot, then Fund workspace

Start the pilot with one application window: portal, referral links, screening and the selection queue. Guest reviewers and interviews come with the first release in the Fund workspace. Prices are agreed with each design partner.

See pricing
Design partner pilots

Run your next application window on your own portal

Pilot with one window: we set up your branded portal and rubric with you, and your selection team keeps every decision.

Talk to us