Hire enso to go-to-market in places only agents can do the job. We let AI agents growth hack your way to customers through the channels your buyers already use - Google, Reddit, LinkedIn, Wikipedia and ChatGPT - so you find demand before competitors notice.
What enso does
enso ships five always-on AI agents that replace fragmented growth tooling: Agentic SEO wins AI Overviews, GEO answers and classic search; Agentic SDR books meetings with in-market accounts on autopilot; Agentic Community lives in Slack, Discord, Reddit and Circle and surfaces buying-signal threads; Agentic Newsletter ships a weekly newsletter your team does not have to write; Agentic Social posts daily across X, LinkedIn, Threads, Bluesky and Instagram. The enso engine orchestrates all five and learns from every experiment run in the lab.
Pricing
enso starts at $49 per month for a single agent surface. Most teams deploy the full stack of five agents on the Growth plan; custom pricing is available for multi-brand and enterprise rollouts. See the pricing page for current tiers.
Compare
Teams use enso instead of HubSpot Workflows, Outreach, Apollo, Clay, Jasper, MarketMuse, and human-run growth agencies. See the comparison page for head-to-head capability tables.
enso is led by Mickey Haslavsky, Omry Hay and Dani Shvarts. The team is based in Tel Aviv and writes blog posts documenting live experiments on platform attention growth hacks.
The Wikipedia citation backdoor that put us inside 47 articles in 90 days.
We discovered an unguarded passage inside Wikipedia's citation castle - the moderated, trust-weighted source of half the open web. 312 source candidates, 47 surviving citations, a 9x lift in downstream answer-engine mentions.
TL;DR: Wikipedia is the most-cited corpus inside every major LLM and answer engine. Most growth teams treat it as a no-go zone because of its editor immune system. We found the door: shape your owned content as a citable secondary source, not a brand asset. 47 of 312 source candidates survived 90 days of editor scrutiny, and downstream AIO / Perplexity mentions of those URLs lifted 9.0x.
Watch the field report - the Wikipedia castle experiment in 90 seconds.
· · ·
The hypothesis
Wikipedia's castle is guarded by volunteers, bots and a 20-year-old policy manual. The front gate (paid placement, brand-page edits, ref-spam) is welded shut. But the castle has a service entrance: articles that need better sources. Every "citation needed" tag is an open door for a reference that meets WP:RS.
If we publish independent, methodology-first research that fills a real factual gap, then editors will place it themselves - and the citation will compound into every answer engine that trains on Wikipedia.
The setup
◐Article pool. 1,140 Wikipedia articles in our domain space with at least one open {{Citation needed}} tag, sorted by monthly pageviews.
◐Source candidates. 312 long-form research pages we published, each filling exactly one factual gap with original data, transparent methodology and a third-party author byline.
◐Door. Edits proposed through Talk pages and WP:RFC, never direct refbombing. Half the placements made by independent editors after the source was indexed in Google Scholar.
◐Metric. 90-day citation survival, plus downstream cited-mention rate in AIO, Perplexity, ChatGPT Search, Claude, and Gemini.
Source archetype → 90-day survival% of placed citations still live
Original dataset + methodology
92%
Independent expert byline
84%
News-style summary post
41%
Product or brand page
6%
Editors keep what looks like a secondary source and prune anything that smells like marketing - independent of who proposed the edit.
What we saw
Of 312 source candidates, 64 were placed within 30 days and 47 survived the full 90-day window. The reverted 17 fell to one of three failures: promotional tone, single-source claim, or an editor flagging the byline as affiliated.
The downstream signal was the headline. Pages cited by even one Wikipedia article were quoted by AI Overviews 9.0x more often than identical pages that never got placed - and Perplexity citations rose 7.3x. Wikipedia is the trust spine the answer engines lean on.
Sources with a methodology section longer than 600 words survived at 89% - versus 12% for sources without one.
Signal contribution to 90-day survivalshare of variance explained
Transparent methodology
29%
Independent author byline
24%
Scholar / news indexing
18%
Neutral, non-promo tone
12%
Backlink authority
9%
Domain age
8%
The top four signals are editorial decisions, not link-building. Authority and domain age still matter - they just stopped being the gate.
What ships next
We are packaging the source-archetype playbook inside Agentic GEO Engine and rolling it across a 2,400-article candidate pool across nine non-English Wikipedias, with parallel tracking on every major answer engine.
