Loisa Review

Loisa's own site no longer describes the lead-qualification product some directories still list it as. As of this review, loisa.ai pitches an open-source, self-hosted "AI coworker" — an agent framework you run on your own machine that plugs into Slack, Telegram, WhatsApp, and email, builds a shared company memory, and works triggers and cron jobs to handle recurring tasks. The product is pre-launch: the homepage's only call to action is "Join the waitlist," and the GitHub link is marked "coming soon." This review covers what's actually live today, not what older listings describe.

Verdict at a glance

What kind of tool is Loisa today?

Loisa positions itself against two alternatives it names directly on its own site: hosted "AI employee" products (which it says are "priced like headcount — and closed") and pure do-it-yourself agent building (which it calls "honest" but "two weeks of glue before hello-world"). Loisa's pitch is the middle path — an open-source boilerplate with memory, tool access, triggers, and permissions already wired up, that a company deploys on its own machines rather than renting from a vendor.

That's a meaningfully different product than the earlier positioning some cold-outreach directories still carry for this slug, which described a paid, hosted SaaS that scored inbound product signups and drafted outreach automatically. Nothing on the current site describes signup scoring, lead qualification, or CRM sync as a core feature. "Outreach" appears only as one bullet among several generic agent tasks (weekly reports, follow-ups, approvals, status updates), aimed at general company operations rather than cold email specifically.

How it's meant to work

The site describes three steps once the product ships: point a coding agent (it names Claude Code as an example) at the cloned repo and let an installer handle setup; connect the company by inviting Loisa into Slack and handing it access to Telegram, email, and its tool library; then let it run — triggers, cron jobs, and heartbeats drive ongoing work like reporting and outreach drafts without a human kicking things off each time. None of this is independently testable right now: there's no public repo, no demo environment, and no self-serve signup — only an email waitlist.

What we could and couldn't verify

We could verify the current positioning, feature list, and pre-launch status directly from loisa.ai. We could not verify pricing, because none is published — there's no plan comparison, no checkout flow, and the GitHub repository the whole product depends on is marked "coming soon." We also found no independent reviews: no G2 or Capterra listing, no Trustpilot page, and no meaningful Reddit or forum discussion turned up for this product under its current positioning. That's expected for a pre-launch tool, but it means any "users say" section would be fabricated — so we're not writing one. Anyone considering Loisa today is evaluating a manifesto and a waitlist, not a shipped product.

Pros and cons

Pros

• Self-hosted, open-source model means your data and agent logic stay on infrastructure you control
• Clear, specific positioning against both hosted "AI employee" vendors and raw DIY agent-building
• Advertises broad integration surface (1,000+ tools) and native presence in Slack/Telegram/email
• Company-wide persistent memory is a genuinely useful idea if it's implemented as described

Cons

• Nothing is publicly shippable yet: no repo, no self-serve access, no live product to test
• No pricing published, so total cost of ownership (including self-hosting overhead) is unknown
• Zero independent reviews or track record exist to check the vendor's claims against
• Self-hosting an agent with broad tool access and company-wide memory raises its own operational and security questions that a pre-launch landing page doesn't address

Loisa alternatives

Because Loisa is positioning itself as general-purpose company AI infrastructure rather than a cold-outreach point tool, the honest comparison set is open-source agent frameworks and hosted "AI employee" products, not email tools:

• CrewAI — open-source multi-agent orchestration framework with an active community and existing production usage.
• LangGraph — open-source agent workflow framework from the LangChain team, widely used for building custom company agents today.
• n8n — open-source workflow automation with AI nodes; a mature, self-hostable option covering some of the same "connect it to everything" ground.
• Lindy — hosted, no-code AI agent/employee builder; the closed, "rented" alternative Loisa's manifesto argues against.
• Artisan — hosted AI sales-rep product; a named example of the "AI employee priced like headcount" category Loisa contrasts itself with.

Who should use Loisa — and who shouldn't

Good fit (once it ships): engineering-capable teams who want an AI agent living in Slack and email, are comfortable self-hosting and maintaining it, and specifically want to avoid recurring per-seat pricing from hosted AI-employee vendors.

Poor fit today: anyone who needs a working tool now — there is nothing to sign up for beyond a waitlist. Also a poor fit for teams that specifically want the signup-scoring, CRM-sync outreach product that older listings describe; that is not what the current site offers, and there's no indication it still exists in that form.