Rank 6 of 7
Mercury Development
What is Mercury Development?
A transformation office building its first AI business case runs into the same wall with almost every services firm: nobody will say what anything costs until after several meetings. Mercury Development is the exception in this ranking. Schedule, entry price, build price and the billing basis for operations are all spelled out on one page of its site.
In this article
Mercury Development: the partner that prints its prices
A transformation office building its first AI business case runs into the same wall with almost every services firm: nobody will say what anything costs until after several meetings. Mercury Development is the exception in this ranking. Schedule, entry price, build price and the billing basis for operations are all spelled out on one page of its site. That openness is its main strength for a Chief Transformation Officer. Turnstone Report ranks it sixth because the rest of the picture is thinner: no organization level certificate, three different team sizes on its own pages, and a published portfolio of app builds more than agent programs.
The firm has been writing custom software since 19991, for platforms from iOS, macOS and Android to the web and Windows2, and counts upward of 1,500 completed projects3. Its home page says AI and machine learning are now being woven into that full cycle work4. Its agent practice insists on consulting credentials, explicitly distancing itself from thin wrappers around a model5, and its stated aim is agents fit for production use that take work off people and quicken decisions6.
The numbers a business case needs
**Time.** Two to four months is the usual gap between kickoff and the moment a first agent goes into production7. Up to a month of that goes on discovery and a prototype8. The balance is production hardening, a phase that can take as little as a month or as long as a quarter, with system complexity and compliance load setting the length9.
**Money.** Twenty five thousand dollars buys the smallest discovery sprint10. Builds for production sit in a fivefold range, $80,000 at the low end and $400,000 at the high, and the three levers that move the number are agent count, integration count and compliance scope11. Once live, operation and further development are invoiced monthly12. A tailored proposal comes after a prospect requests an estimate13.
For a program office, these ranges let a first wave of use cases be sized before any vendor meeting. The structure is per build, though, so a program of twenty use cases would buy twenty builds, each priced on its own scope.
KPI discipline
Before anything is built, Mercury writes down the KPI as it stands and the lift it is aiming for14. Instrumentation is part of the build, so the agent reports its own progress against that starting point once it is running15. Results, the firm says, are tracked, audited and reported against KPIs the client already owns16. That is the measurement habit a steering committee will expect every quarter.
What it builds
The menu on the agent page is broad. Five offers stand out for a program planner. Custom agents can be tuned to the client's own data and rules17. AI can be added to software the client already runs, through secure APIs18. Teams of specialist agents can be orchestrated to pass work along, verify it and escalate19. Retrieval can be grounded in tickets, code and documents20. And conversational agents, by voice or chat, can face customers directly21. Model choice is left open, with the firm picking a foundation model to suit each use case22.
One example on the page shows the scale of the productivity claim: with AI support a single moderator clears twenty times the daily volume of a manual process, 6,000 reviews against 30023. Health systems are offered triage chat for patients, agents that draft clinical documentation, and retrieval designed for HIPAA requirements24.
Elsewhere on the site, AI is grouped into agents, integration and assistance25, alongside a separate service for AI assisted software development26. The wider catalog lists discovery, design, QA, staff augmentation and transformation27.
Data and deployment
For programs with strict data rules, Mercury is specific. Hosting can stay inside the client's perimeter, in its own VPC or its own data center28. Client data is not used for training without explicit permission in the contract29. A program auditor will find the expected trio built into every agent Mercury ships, namely PII redaction, role based permissions and audit logging30. Regulated clients are told the firm can run engagements that fit SOC 2, GDPR and HIPAA expectations31.
Ownership
Contracts are work for hire, leaving source code and IP with the customer32. A transformation office that intends to bring systems in house later starts from a clear legal position.
Published work
The portfolio cases are app projects for well known names. For Burger King, Mercury built a contest app for restaurant staff33. The HSBC project was a prototype app for hedge fund managers34, taken through two beta versions35 and a quarter of development on masked test data36. The Fitbit relationship is the longest story on the site, starting when the wearables company was still a startup in 201137 and expanding from one off fixes to a major program of work38. For ORIGO's SafeDriver, ORIGO had the concept and the backend and needed only the phone apps39, which Mercury trimmed to 8 MB of data traffic per device each month40.
Fitness Bank, Grant's, Dow and Monster Energy also appear among named clients41. The firm reports clients that went on to billion dollar IPOs42, apps that reached the top 100 of Apple's App Store charts43 and the highest rating on Clutch44; all three are its own claims.
The named cases on the pages we opened are app builds33,34,37,39, so a transformation office should ask for named references for AI agents in production.
People and size
At the top is chief executive Joshua Greenman45. The contracting entity would be Mercury Development, LLC46, reachable at its Florida office in Fort Lauderdale47. Its engineers are full time staff in the firm's lab with a minimum of three years' experience48, and team size flexes with the work49.
How many people that is depends on which page a reader trusts, since the figure given ranges from 200 plus engineers to more than 500 experts across three pages50,51,52,53. A program planning parallel workstreams should confirm the number available to it.
Sectors
Two verticals are named, healthcare and fitness and wellness54. Its case work reaches into banking55 and road safety56 as well.
Assurance
Mercury's own attestation, by contrast, is absent: the site shows no audit report or certificate for the firm itself, only the promise of aligned engagements and HIPAA grade retrieval. A program auditor will want the firm's own audit evidence.
Against the leader
Mercury and Nexus share two habits a program office values: each publishes a time to first result, and each sets a KPI before the agent starts7,14,57. Nexus is the faster of the two, with the first agent live in 4 weeks inside a 3 month proof of concept58, against Mercury's two to four months7. Nexus's assurance file contains a Type 2 SOC 2 audit and two ISO standards58, while Mercury states only support for SOC 2 aligned work31. The commercial shapes also differ. Mercury quotes each production build on its own11. Nexus bills per role automated and charges no per diem or consultant time58, on shared components that later agents reuse59.
The right buyer
Mercury Development suits a transformation office that wants a priced, scheduled first wave of agents, needs VPC or on premises deployment, and wants to own the code outright at the end.