Products & Tools

Lever launches shared business context for AI assistants

Form & Function's Lever links operational data with company definitions for staffing, revenue and profitability questions. One client found $1.8 million in potential capacity.

Form & Function Launches Lever to Give AI Assistants Shared Business Context
Form & Function Launches Lever to Give AI Assistants Shared Business ContextAS_Photography / pixabay
By Marcus Bennett6 min read

Why it matters

  • Form & Function Consulting launched Lever on November 12.
  • One client identified approximately $1.8 million in potential incremental annual revenue capacity without adding employees.
  • Form & Function tested five business questions five times each against raw BigQuery data.
  • The Malloy-built semantic model performed better on questions requiring definitions and table relationships.
  • Form & Function did not disclose pricing or detailed deployment requirements.

By Marcus Bennett

Form & Function Consulting launched Lever on November 12, and the company reported that one client identified approximately $1.8 million in incremental annual delivery capacity without adding employees.

Form & Function described that figure as a potential operating gain identified through analysis, not revenue the client had already collected. The distinction sets the evidence standard for a product sold around decision support rather than completed financial results.

Form & Function says Lever gives professional services firms a consistent view of operations across disconnected sales, delivery and finance systems. The platform combines operational data with company-specific business definitions, then lets leaders ask about staffing, revenue forecasts and project profitability through conversational interfaces.

The commercial stakes are concrete, according to the launch announcement. A sales system can record a promising deal while the delivery team sees insufficient capacity to fulfill it. Merely connecting both databases does not tell an AI assistant how to interpret that conflict in business terms.

This is where Lever differs from another financial chatbot. Form & Function built its proposition around work performed before answers appear, using a context layer (a shared description of definitions, operating assumptions and relationships between records) to make separate systems comparable.

What problem does the context layer solve?

Form & Function's launch announcement says Lever starts with information from customer relationship management, professional services automation and finance systems. It also gathers documents and the knowledge employees use to interpret those records.

The Lev chatbot supports stakeholder interviews during onboarding. Form & Function says those interviews capture "how a firm actually operates," adding operating knowledge that structured records may omit.

That onboarding then feeds a context layer describing how work is "sold, staffed, delivered and measured," according to the company. The layer aims to connect questions that otherwise fall between departments.

Form & Function highlights three operational decisions:

  • Whether the sales pipeline supports a hiring decision.
  • Which engagements offer the best margins.
  • How a delayed investment would affect projected cash flow.

Form & Function says similarly named metrics can conceal different assumptions. Available capacity might exclude time spent selling. A revenue forecast might depend both on a deal's likelihood of closing and on when the resulting work can be delivered.

Definitions determine the answer. According to Form & Function, a plausible response based on the wrong definition can be more misleading than an obvious error, because the answer remains fluent while carrying the wrong business assumption.

Form & Function's framing shifts the value proposition for enterprise AI. Connection to data establishes access. Shared definitions establish meaning. The second step requires agreement about terminology, assumptions and relationships before a model can reason across departments.

Which AI interfaces can use Lever?

Form & Function's Lever product page shows Lev answering questions, generating visual output and working inside Slack. The page also shows access through Claude and ChatGPT using Model Context Protocol (MCP), a protocol used to connect those AI clients.

Form & Function presents this arrangement as a shared foundation for several interfaces. Rather than requiring every chatbot or AI client to reconstruct the firm's business logic separately, Lever supplies the same underlying interpretation of operations.

That architecture lets a firm preserve one interpretation while changing the interface used to interrogate it, according to Form & Function. The product page shows Lev operating in Slack while Claude and ChatGPT can access the same shared context layer.

Form & Function positioned the launch within operating-partner advisory engagements and emphasizes consultant-assisted onboarding. The current product page also describes standalone availability, so the product has both an advisory path and a direct software offering.

Pricing and detailed deployment requirements remain undisclosed. Form & Function says the announcement should not be read as a self-service rollout with a published subscription price. Buyers therefore have no public subscription figure from the launch material.

The company also points to Collate's Ontology Studio announcement as a parallel sign of businesses moving toward shared business meaning. Form & Function characterizes Lever as a narrower approach centered on professional services operations and consultant-assisted onboarding, rather than a generic layer every organization can adopt unchanged.

Form & Function uses that comparison to place Lever inside a broader enterprise AI transition: models can read more systems, but organizations still need a way to encode what those systems mean. For professional services firms, the immediate market test is whether a shared operating model can shorten the path from records to staffing, margin and cash-flow decisions.

What do the early results establish?

Form & Function reports that one client identified additional delivery capacity sufficient to support approximately $1.8 million in incremental annual revenue without adding employees. The release also describes an improvement in projected gross margins, a profitability measure based on revenue remaining after direct costs.

Form & Function treats those figures as potential operating gains identified through analysis. They do not establish that the client collected $1.8 million in new revenue, and the launch announcement provides no basis for treating the projected margin improvement as a realized result.

The company also offers a technical comparison intended to show how its semantic approach differs from raw data access. It tested five business questions five times each, contrasting raw BigQuery data with a semantic model built using Malloy.

A semantic model is a structured representation of business definitions and relationships between tables. In Form & Function's comparison, the structured approach performed better on questions that required those definitions and relationships. The test covered 25 runs across five questions, according to the company.

Form & Function presents the comparison as an illustration of the mechanism, not a general accuracy guarantee. It is a small vendor-run test. It does not establish perfect accuracy across arbitrary questions, changing datasets or every customer's deployment.

That limit matters because service businesses revise contracts, staffing policies and delivery processes. Predefined metrics can reduce ambiguity, but Form & Function acknowledges that they do not eliminate the need to maintain definitions and review consequential decisions.

Where does the implementation work go?

Form & Function's central argument is straightforward: connecting a model to data is only the beginning. Someone must decide what the data means, resolve conflicting definitions and keep that interpretation current as contracts, staffing policies and business processes change.

The work therefore moves partly from prompt writing toward maintaining business definitions. That maintenance has direct decision consequences. A capacity definition that omits selling time, or a forecast that ignores delivery timing, can steer a hiring or investment decision even when the underlying records remain accurate.

For a professional services firm, Form & Function says the potential benefit is a shorter path from operational records to a decision about capacity or profitability. The platform supplies a common basis for producing and examining that decision; clients still need to review consequential choices.

The launch material also leaves the commercial comparison open. Form & Function discloses neither pricing nor detailed deployment requirements, even as it describes consultant-led operating-partner engagements and standalone availability. Prospective clients must assess implementation scope and cost separately from the product's analytical claims.

Form & Function identifies two lasting tests: whether shared context remains reliable as the business changes and whether identified opportunities become measurable results rather than more persuasive answers. The next proof point is whether the approximately $1.8 million in potential annual revenue and the reported projected gross-margin improvement become booked revenue and sustained margin gains. Until then, the figures remain identified operating opportunities.

Source: unite.ai

Share this article:

More from Marcus Bennett

Marcus Bennett

Show full bio

Senior reporter covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.

791 articles

Related articles

  1. Databricks Plugs Genie Business Context Into ChatGPT, Claude, Copilot and Grok
  2. Microsoft Copilot adds Esker checks for supplier payment risks
  3. Delphina raises $7.5M to scale its AI context system
  4. Australia's loan AI talks. Humans still make the decisions.
  5. Morgan raises AUD $3.65 million seed for AI insolvency software

« Previous articleNext article »