Operator-built AI playbooks
We’re operators who’ve already deployed AI across our own businesses — exec, sales, marketing, product, design, engineering. We bring those playbooks and tools with us, configure them for your team, and have your company using AI on real work in weeks, not quarters.
of organizations now use AI in at least one business function
McKinsey, State of AI 2025
are capturing real value from AI at scale
BCG, The Widening AI Value Gap, Oct 2025
of GenAI pilots fail. And the gap isn't the model — it's people and process.
MIT NANDA, The GenAI Divide, Aug 2025
The approach
Tools land. Slack channels light up. A few power users go feral. Everyone else watches, waits, and quietly carries on with the old workflow. Six months later, the dashboard says adoption is flat and nobody can quite explain why.
Kindling is built different on purpose. We’re operators who run an Applied AI Lab against our own work every day: models, agents, skills, plugins, rituals, and workflows tested before we ask your team to trust them.
So we arrive with the playbooks already built, the tools already configured, and a strong opinion about what is durable enough to install. Our job is to tune that toolkit to your context and stay until the new behavior is the boring default.
Operating credibility
Kindling exists because we kept running the same playbooks inside our own businesses — and friends kept asking how. We’re not consultants who happen to write about AI. We’re operators who happen to consult.
Kindling is a partnership of operators with companies in market. The playbooks you see here ran inside ours — across exec, sales, marketing, product, design, and engineering — before we ever shipped one to a client.
Operating companies
Claude and Codex skills in our planning loops. OpenAI workflows in our tooling. Plugins, agents, and evals tested against actual client-shaped work before they become recommendations.
Applied AI Lab
We don't whiteboard a 12-week strategy and hand off slides. We arrive with the agents already configured, the prompts already evaluated, and the rituals already proven. Your week-one deploy is our week-N + 1.
Time to first deploy
Where these playbooks have run
Logos and case studies are coming with our first published engagements. In the meantime, ask us about specific companies and functions on a call — we’ll show you the live work.
Reveal soon →What we bring
Before Kindling existed, we built and shipped AI inside our own businesses — across every function below. Every playbook on this list has been used to move real work, not just fill a deck. We arrive with the tools already configured. Your team starts adopting in week one.
Weekly reviews powered by AI-summarized data rooms, board prep that writes its own first draft, and a decision log that doesn't rot. Configured for your cadence in week one.
Pipeline reviews that write themselves, AI-drafted follow-ups that sound like your AE, and account research that arrives before the rep does. Built in our own GTM motion and battle-tested on real quota carriers.
Editorial workflows that turn one strong idea into a week of content, lifecycle messages that actually personalize, and analytics that explain themselves. Drop-in agents your team operates, not a black box.
User-interview synthesis that doesn't lose nuance, PRDs your eng team will actually read, and a roadmap doc that updates itself from Slack and the bug tracker.
Briefs that capture intent in a paragraph, exploration loops that 10× the options you can put on the table, and design QA that catches the boring stuff before a human has to.
Cursor and Claude Code configured for your codebase, PR review agents that catch the third thing you would have caught, and runbooks that turn outages into closed loops.
Every playbook ships with the prompts, the tooling, the rituals to run it, and the metrics to prove it.
The model stack
We’re model-agnostic on purpose. We pick the right model for each job — frontier when it earns its keep, smaller and cheaper when it doesn’t, fully local when the data can’t leave the building. Every playbook in the previous section runs on this stack, and we’ll guide your team on choosing too.
Production-grade work across the Claude family — Sonnet, Opus, Haiku. We ship with Claude Code in the codebase, build agents on the SDK, and engineer for prompt caching, tool use, and MCP integration.
We use the full OpenAI stack — Responses, Agents, voice and realtime — and know which model earns its keep on which job. Cost-aware production patterns, not benchmark cosplay.
When the data can't leave the building — finance, legal, healthcare, regulated workflows — we deploy open models inside your perimeter. Same playbooks, your hardware, no API call leaving your VPC.
Vendor-neutral by design. The right model for the job, the right deployment for your risk posture, and an exit plan from any of them.
Who we work with
We work with growing companies that have outgrown “just–use–ChatGPT” and aren’t about to hire a Chief AI Officer.
Engineering, product, and growth teams from 50 to 2,000 people. Usually already paying for Cursor or Copilot, usually still wondering why velocity hasn't moved.
What we typically deploy
Consulting firms, agencies, law firms, accounting practices, RIAs. Knowledge work is the entire P&L, which means AI changes the unit economics — if you actually deploy it.
What we typically deploy
Services rollups, regional brands, PE portfolios, multi-entity operators. AI leverage compounds across business units when one playbook can land in many places.
What we typically deploy
Not on this list? The honest answer is we’re probably still a fit. The playbooks generalize. Ask us.
What we do
We compose engagements from these four practices. Most teams start with the diagnostic and grow from there.
We meet your people where the work actually happens — in tickets, deals, designs, code, and meetings — then map where our current AI toolkit changes the math and where it doesn't.
We turn the diagnostic into the way Mondays actually run: standups that surface AI bets, weekly reviews that close the loop, and per-function playbooks your team can pick up and use without a deck.
Coaches and builders working alongside your team — in their tools, on their calls. We adapt the lab's strongest workflows to your business and stay until the new behavior is the boring default.
Your leadership sets the temperature. We work with founders and ops leaders on the harder questions: what work disappears, what new work emerges, and how to talk to your team about both — honestly.
The process
We don’t run discovery sprints from a war room. We sit next to your people until we can do their job.
We embed with the people doing the actual work. Sales calls, design reviews, support queues, financial close. Before we recommend a single tool, we see how Mondays really run.
We arrive already holding the playbooks. By the end of week two, the right ones are configured for your stack, your data, and your tone — and a first cohort is using them on real work.
We stand up the rituals, train your champions, and rewrite the metrics. Then we keep showing up — because change that ships in a memo doesn't survive contact with Q4.
Models change. Vendors churn. Teams reorg. We stay close enough to your operating cadence to keep the fire burning long after the launch dust settles.
Our manifesto
We measure success by what work moved, not how many seats lit up. If your sales team didn't write fewer emails this quarter, the rollout didn't work.
Most pilots exist so nobody has to decide. We design engagements that force the call: invest deeper, or kill it clean. Either is fine. Drifting is not.
Models, agents, plugins, and workflows are changing too quickly for most teams to track while running the business. We absorb the frontier so your team can adopt what survives.
If AI is going to take work from your team, say so. Plainly. Then help them go somewhere new. The companies that won't have this conversation will lose the trust war first, and the talent war right after.
A 30-minute call. No deck. We’ll ask three honest questions about your AI rollout, tell you what we’d do, and you can take the answers home for free.
Booking channel opens shortly