2026-09-05 The first real calls have been placed: five ended normally, six failed. Here is what broke.

Comparisons

Every comparison, including the rows we lose

CallAgent competes in three different markets, and somebody arriving from one of them has never heard of the vendors in the other two. The pages below are grouped exactly that way.

Every comparison is written under the same rule: a sentence in both columns, no ticks, everything we say about the other product read off their own public pages, and the rows we lose left in the table and marked.

§01 The rule

What makes a comparison worth reading

A comparison page written by the seller is read by somebody who expects it to be rigged. The only useful reply is to say what would make it not rigged, and then do exactly that.

  • Sentences, not ticks

    A tick is a claim with the qualification removed. Every row has a sentence in both columns, and a cell lifted out of the table on its own has to still be true.

  • The losing rows stay

    Every page marks the rows the other product wins and closes with a list of situations in which you should not pick us.

  • The source is theirs

    Everything we assert about a competitor is read off their own site, with the month it was read printed on the page. Where their site does not say, the cell says we could not verify it.

§02 Comparisons

Three markets, three lists

A product that is a voice agent, a phone switch and a contact centre at the same time gets compared with very different things depending on who is looking for it. Pick the list you arrived from.

Voice AI platforms

Products for developers and for teams building the agent themselves. Here the comparison is about the per-minute price, which providers are open, and where the thing runs.

Contact centre and call centre

What a company replacing an entire call centre evaluates. Here the comparison is about human operators, the contract, the references and the certifications — and here is where we lose most.

Phone systems

What a company shortlists when it is looking for a telephone system first and AI second. Here the comparison is about SIP trunks, extensions and the price per year.

There are 10 comparisons written so far. Missing a vendor you are evaluating? Tell us and we will add it, under the same rule.

§03 Constant

What we lose in every one of them

Four rows repeat on every page, whoever the other product is. They are collected here so you do not have to open ten pages to find them.

The status page, with the figures

  • Real traffic. Eleven real calls in total, six of which failed, and no customer served. Every product on the lists above has more than that.

  • Certifications. None: no SOC 2, no ISO 27001, no PCI, and no penetration test by a third-party firm.

  • Scale. We publish no concurrency limit because we have never run concurrent volume, and that is not the same thing as "unlimited".

  • Support and implementation. There is documentation, there are guides, there is code. There is no implementation team, no support contract and no guaranteed availability.

§04

The vendor you are evaluating is missing

If the product you are comparing us with has no page here, tell us which it is. We read it off their own site and write the table under the same rule, including the rows we lose.

Notes

  1. 1 Every comparison page cites its sources in its own footnotes: that vendor's public pages, with the month they were read. We do not use third-party summaries of a competitor.
  2. 2 Our own figures come from the same place on every page: the status page for calls and latency, the pricing page for the €0.10 per minute, the registries in this repository for providers and connectors.