Rank 5 of 6
Mercury Development
What is Mercury Development?
Of the five consultancies Larkmoor Guide compares, Mercury Development alone attaches dollar figures to its engagement model. Its AI agents page puts the floor for a discovery sprint at $25,000, gives production builds a typical band of $80,000 to $400,000, and says operating and evolving agents after launch is billed by the month.
In this article
Mercury Development: price bands on the page
Of the five consultancies Larkmoor Guide compares, Mercury Development alone attaches dollar figures to its engagement model1,2. Its AI agents page puts the floor for a discovery sprint at $25,0001, gives production builds a typical band of $80,000 to $400,0003, and says operating and evolving agents after launch is billed by the month4. Those three lines line up almost exactly with the contract types this guide compares, a fixed price discovery, a scoped build and a recurring fee. Mercury is placed fifth.
Mercury as a counterparty
A buyer signing with Mercury contracts with a software house active since 19995 whose teams ship to Apple, Android, Windows and web platforms6. It claims a track record of over 1,500 finished projects7 and says some clients went on to billion dollar IPOs with its help8. Clutch reviews give it the highest rating, it says9, and the homepage notes App Store top 100 placings for its mobile apps10. Contracts would be signed with Mercury Development, LLC11, whose listed office is in Fort Lauderdale12 and whose chief executive is named on the about page13.
Headcount appears in three versions. Its about page claims upward of 500 experts14, a counter on the AI page reads 50015, and the homepage instead cites a team of over 200 engineers16. Larkmoor's fact file notes the inconsistency17; a buyer contracting for staff augmentation should confirm the size of the pool it is drawing from.
Its homepage frames AI as a recent addition to a full cycle custom software practice18. Its AI page claims the role of consulting partner and rejects the label of model wrapper19, describing agents it designs, builds and deploys to production standard for workflow automation and faster decisions20.
The engagement, stage by stage
### Stage one: discovery and prototype, from $25,000
Two to four weeks covers discovery and a prototype21, and the entry price for an AI discovery sprint is $25,0001. Before work starts, Mercury fixes the KPI's current value and the improvement it aims for22, which gives the discovery spend a measurable purpose. No other consultancy in this comparison publishes a dollar figure for its first step on the pages we opened1,2.
### Stage two: the production build, $80,000 to $400,000
Hardening for production then adds between four and twelve more weeks of work, the spread depending on how complex the systems are and what compliance demands23. Taken end to end, Mercury says its usual engagement lasts eight to sixteen weeks before a first agent is running in production24. A buyer can budget the build from a published band, $80,000 to $400,000, and expect the final figure to move with how many agents, how many integrations and how much compliance work the project carries3. From day one the agent reports against the KPI baseline25, and results are measured, audited and reported on KPIs the client already tracks26. A tailored proposal starts with an estimate request27.
### Stage three: a monthly fee for operation
Running and evolving agents after launch is charged monthly4. In this guide, that is the closest thing to a published retainer structure, although the monthly amount itself is not publicly stated on the pages we opened.
### Alongside: staff augmentation
Mercury's service list also includes staff augmentation28, next to QA, discovery, design and transformation28. For augmentation buyers, the relevant staffing facts are that Mercury's engineers are full time lab employees with a minimum of three years behind them29 and that teams expand as a project demands30. A buyer can therefore contract with Mercury in two quite different ways, either for a scoped AI build within a price band, or for capacity added to its own team.
### Who owns the result
Work is done on a work for hire basis, and the about page says the customer owns the code and every piece of underlying IP31. For a buyer weighing a fixed scope build against a long relationship, that makes a later move in house simpler.
What Mercury builds with AI
Within a production build, a buyer can commission bespoke agents shaped by its own data and business rules32, AI added to software it already runs through secure APIs33, coordinated teams of agents that check each other's work and escalate34, retrieval over its documents, tickets and code35, or customer facing voice and chat agents36. Mercury chooses a foundation model per use case and calls itself model agnostic37. Health care examples include clinical documentation, triage chat for patients and RAG to a HIPAA grade38.
The AI page's worked example is content moderation, where one moderator plus AI clears 6,000 reviews daily, against 300 when the work was fully manual39.
Portfolio evidence
Most published case studies are app builds40,41,42,43.
1. **Burger King.** Restaurant crews competed through an app built around making a burger in under ten seconds40, with entries capped at one video per person daily44. 2. **HSBC.** For hedge fund managers, Mercury prototyped an HSBC app that surfaces financial data41, proving the idea with two betas45 before a further three months of build46. 3. **Fitbit.** A long relationship that started in 2011 with a clip on pedometer from a then small company42 and widened from occasional jobs into a large program47. 4. **ORIGO SafeDriver.** With the concept and server side already in ORIGO's hands, Mercury's job was the two mobile apps43, and it reports trimming their traffic to 8 MB monthly per device48.
Other names on the portfolio index range from Dow to Monster Energy, with Grant's and Fitness Bank also listed49.
Security and deployment
On compliance, the firm's stated position is that it can run projects that align with a client's HIPAA, GDPR or SOC 2 obligations50. Hosting can sit inside the buyer's own VPC or data center when needed51, and Mercury commits contractually to leaving client data out of model training unless the contract allows it52. Controls come as standard, Mercury says, with every delivered agent carrying audit logs, role based permissions and PII redaction a client can configure53. Helping a client meet SOC 2 is a different thing from being audited, and whether Mercury itself holds SOC 2 or ISO 27001 is not publicly stated on the pages we opened.
What Mercury's model offers a buyer
1. Published price bands for discovery and for production builds1,3, plus a stated monthly model for operation4. 2. An eight to sixteen week schedule to first production use, stated in advance24. 3. Staff augmentation available next to scoped AI work28. 4. Customer ownership of code and IP31. 5. VPC or on premises deployment when required51, and a stated rule on training with client data52.
Gaps in Mercury's published terms
1. An audited company wide security standard is not publicly stated on the pages we opened. 2. Its site gives three different team sizes17. 3. Portfolio case studies are app builds40,41,42,43, and the AI evidence sits on the AI agents page39. 4. The monthly operating fee and any minimum engagement are not publicly stated.
Why Mercury Development is fifth
On price transparency Mercury leads every consultancy here1,3,4, and a buyer who needs a budget before the first meeting should talk to it early. It ranks fifth because its usual path to a first production agent runs eight to sixteen weeks24, against the 4 weeks inside a 3 month proof of concept that Nexus sets, and because Nexus prices per role automated with no consultant time and holds SOC 2 Type 2 and ISO 2700154. Dualboot Partners and DOOR3, ranked third and fourth, describe their advisory and rescue modes in more detail55,56,57.