Legal & Professional Services · SaaS

EchoBase AI: a $10K MRR tool for consultants that shut down

EchoBase AI automated analysis for consultants and reached about $10,000 a month, mostly from $299 enterprise plans. EchoBase AI shut down in 2024; here is what happened.

TL;DR: EchoBase AI, from the startup Artificial Workflow, analysed consultants' interview transcripts and suggested recommendations, and reached about $10,000 a month in recurring revenue, about 80% from enterprise clients paying $299 a month (founder-reported, Starter Story, June 2024). The founders said a cheap first developer cost them about $50K in wasted payments. Starter Story notes the business shut down in 2024.

Founders: Andrew, James · By · Updated

EchoBase AI at a glance
Revenue About $10,000 MRR (before shutting down)Founder-reported Source, 2024-06
Team size 3 founders, 2 employeesFounder-reported Source, 2024-06
Pricing Enterprise clients at $299 a monthFounder-reported Source, 2024-06
Founded 2023 (shut down in 2024)
Founders Andrew, James

Most case studies end with growth. This one ends with a shutdown. EchoBase AI, the main product of a startup called Artificial Workflow, reached about $10,000 a month in recurring revenue selling AI analysis to consultants (founder-reported, Starter Story). Starter Story later noted that the business, started in 2023, shut down in 2024. The founders’ own account shows what worked and what cost them.

How did 200 sales calls shape EchoBase AI?

The company started as an AI consulting business, building custom AI solutions for clients. Co-founder James had worked as a management consultant at EY and had seen how much time consultants spend on manual data processing and analysis. The team found clients through cold outreach on LinkedIn, which led to about 200 sales calls, mostly with consulting firms.

On those calls a pattern appeared: they kept building the same solution for different clients. So they packaged it as subscription software. EchoBase AI let consultants upload interview transcripts, analysed all of them with AI and suggested customised recommendations.

What did a cheap developer cost EchoBase AI?

They designed the first version in Figma using feedback from the LinkedIn calls, then hired a very cheap developer on Fiverr, who got the MVP over the line, just barely. It was enough to sign four or five enterprise clients. Then they discovered the app had no asynchronous processing: whenever one client was uploading or using it, every other client had to wait.

The rebuild took three months, two developers who were not the right fit and finally a development agency. The founders estimated about $50K in wasted payments to developers (founder-reported, Starter Story).

“The biggest lesson we learned was the importance of investing in quality development from the start.” — Andrew (founder-reported, Starter Story)

Which channels brought EchoBase AI paying users?

Not social media. After the rebuild they relaunched with posts on LinkedIn, Medium, Instagram and TikTok, and those brought almost no users. Two channels did the work. LinkedIn cold outreach, helped by James’s reputation in consulting, brought enterprise clients. SEO, built by targeting relevant keywords with blog posts and tool pages, brought almost 10,000 clicks in a single month.

How much revenue did EchoBase AI make?

In June 2024, EchoBase AI made about $10,000 a month in recurring revenue, about 80% of it from enterprise clients paying $299 a month (founder-reported, Starter Story). The team had three founders and two employees. Total revenue is not disclosed, and the company shut down later in 2024; the reason given was simply that it shut down.

StageWhat happened
2023AI consulting agency; about 200 sales calls via LinkedIn
MVPFiverr-built app; 4–5 enterprise clients; no async processing
RebuildThree months, two developers and a dev shop; about $50K wasted
June 2024About $10,000 MRR, 80% enterprise at $299 a month
2024Business shut down
Sourcefounder-reported, Starter Story

How does EchoBase AI compare with Harvey and Gojiberry AI?

Harvey sells AI to professional services firms too, lawyers instead of consultants, and raised hundreds of millions to handle security, compliance and long enterprise sales. EchoBase AI tried the same kind of buyer on a tiny budget, and the gap shows how expensive enterprise-grade software is to build properly.

Gojiberry AI followed a similar path from service to software: deliver the result by hand first, then automate it. Gojiberry AI invested in a CTO; EchoBase AI tried to save money on its first developer and paid for it later. The service-first idea worked in both cases; the build quality made the difference.

What are the lessons from EchoBase AI’s shutdown?

  • Turn repeated custom work into a product. Building the same thing for many clients is a strong signal.
  • Don’t cut corners on enterprise software. A cheap MVP cost about $50K and months of rebuilding (founder-reported, Starter Story).
  • Channels must match the buyer. Cold outreach and SEO worked; social posts did not.
  • Revenue is not survival. About $10,000 a month was not enough to keep the company going.

Frequently asked questions

How much revenue did EchoBase AI make?

About $10,000 a month in recurring revenue in June 2024, roughly 80% from enterprise clients paying $299 a month (founder-reported, Starter Story). Total revenue is not disclosed.

Is EchoBase AI still running?

No. Starter Story notes that Artificial Workflow, the company behind EchoBase AI, was started in 2023 and shut down in 2024.

What did EchoBase AI do?

It let consultants upload interview transcripts and used AI to analyse them all and offer customised recommendations, so consultants could focus on strategic work.

What went wrong in EchoBase AI's build?

A cheap Fiverr developer built an MVP without async processing, so one client had to wait while another used the app. Fixing it took two more developers and a dev shop, and about $50K in wasted payments (founder-reported).

Sources

  1. We Created AI Tools For Consultants And Make $10K/Mo – Starter Story, 2024-06

Labels: Founder-reported = stated by the founder or company; Verified = backed by a payment-processor screenshot, TrustMRR or a public platform count; Media-reported = published by a news outlet or third party. "Implied" monthly figures are annual figures divided by 12. We never estimate numbers ourselves.

Last updated: