AI in your teams’ real work, not in a slide deck

The reality

Generative AI is new. Demos are easy; successful deployments are rare. Failures look alike: a tool bought with no specific problem to solve, a pilot that never reaches production, teams left out of the process, no measurement of results.

Few companies, and few leaders, have taken an AI transformation all the way. There’s no proven playbook yet to apply as is. What will work for you gets built with you, starting from your teams’ real work.

The method: the forward deployed engineer

A forward deployed engineer (FDE) is an engineer who works inside a client’s teams to turn a loosely defined problem into a system that runs in production. The role started at Palantir. OpenAI, Anthropic, and Google have adopted it widely since 2025, for a simple reason: in a company, the hard part isn’t the AI model, it’s deploying it into existing processes.

Monolithic Lab applies this model to transforming a small or mid-sized company:

• On site, in the real work. Use cases come from watching the teams’ processes, data, and tools, not from off-site workshops.
• Code, not slides. Solutions are built, deployed, and debugged on your data and your systems.
• With the domain experts. The person doing the work knows what actually helps.
• Judged on usage. A successful deployment is measured after go-live: time saved, errors avoided, service quality.
• Handed over. Your team can evolve what was built without outside help.

Read the article: what is a forward deployed engineer?

For the engineering team

AI is changing how software gets built. Agents write, review, test, and document a growing share of the code. The engineering team has to change how it works:

• coding agents built into the development cycle: writing, review, testing, documentation, migrations
• new skills: precise specs, breaking down work, reviewing and validating what agents produce
• guardrails: security, data privacy, code ownership, cost control
• organization: team size, roles, profiles to hire

For every other team

AI isn’t just for developers. The fastest gains are often elsewhere:

• customer support: triaging requests, drafting replies, keeping the knowledge base current
• sales and marketing: meeting prep, lead qualification, content production
• finance and admin: document extraction, reconciliations, reporting
• operations: tracking, scheduling, quality control
• HR: onboarding new hires, internal documentation

The adoption plan, built with your teams

With each team: repetitive tasks, the data involved, the pain points. A list of concrete use cases, not a grand vision. Each use case is scored on expected gain, risk, and feasibility. Three or four to start. Prototypes on your data, used by the people concerned, with a success criterion set in advance. What doesn’t work gets stopped fast. Go-live, training, usage rules, measured results. Internal champions in each team, a repeatable method, and a plan for the next use cases.

How to engage

• As part of a CTO engagement. AI transformation is part of both the fractional CTO and the interim CTO roles.
• As a fixed-price project. AI exploration, POC, or MVP, with scope, deliverable, and price set in advance. See fixed-price projects