Six Week AI · Ajit KrishnaFirst workflow
Everyone agrees AI matters. Nobody knows what to build first.
- Your CEO wants progress.
- Engineering has other priorities.
- Consultants want to start with strategy.
- Vendors want to sell another platform.
Meanwhile, your team is still losing hours every week to work AI could genuinely help with. The risk isn’t that AI can’t help - it’s spending months on the wrong workflow.
We help you choose the one workflow that’s worth building first, redesign it around what AI makes possible, and get it into production in six weeks.
If it’s not a fit, we’ll tell you on the call.
Find the right first workflow
Every department has one workflow quietly eating hours.
The first job isn’t to “use AI.” It’s to find the workflow where AI can remove real operational drag.
Finance / AP
Coding invoices by hand
Today
Hours matching lines to GL codes.
With AI
Coded for you; you handle only exceptions.
Marketing
Stitching campaign reports
Today
A day pulling numbers from five tools.
With AI
A drafted report in minutes; you edit.
Customer Success
Preparing for a renewal
Today
Hours digging through account history.
With AI
A one-page brief, ready before the call.
Why us
Your engineers can probably build it. That’s not the hard part.
Most AI projects don’t fail because engineering can’t write the code. They fail because nobody has clearly answered three questions:
- 1.How does the work actually happen today?
- 2.What does AI make possible that wasn’t before?
- 3.What has to be true before this can safely go live?
Understand the work.
Map the current workflow and capture the edge cases. Document how the work really happens - not how the process is supposed to happen.
Redesign around AI.
Don’t automate yesterday’s process. Redesign the workflow around what AI can now do - remove steps where possible, and change the shape of the work.
Define the build.
Identify the data, systems, permissions, risks, integrations, and success measures - and produce a build-ready implementation specification.
Get it live.
Your engineer can build it with our support, or we build it with access from your team. Either way, we validate the result, refine the workflow, and hand it over.
Ownership
Your first AI workflow shouldn’t become your first AI dependency.
The goal isn’t just to ship one useful workflow. It’s to leave your team with the understanding, documentation, and confidence to build the next one without starting from scratch.
We explain the decisions.
Your team understands why the workflow is designed this way.
We document the system.
No mystery prompts. No hidden architecture. No black box.
You own the capability.
The first workflow becomes the pattern for the second.
The real problem
Companies don’t have an AI problem. They have a workflow problem.
Most companies start by asking, “Where can we use AI?” That’s the wrong first question. The better one: which workflow is painful enough, repetitive enough, and valuable enough that changing it would actually matter? That’s where we start.
- Choose the right workflow
- Understand the messy reality of the work today
- Redesign it around AI, not old habits
- Define what must be true before anyone builds
- Measure business outcomes, not demo quality
How it works
One workflow. Three phases. Six weeks.
Diagnose & design.
- Discovery sessions.
- Current workflow map.
- AI-native future workflow + spec.
Deliverable
A build-ready blueprint.
Build.
- Your engineer builds with our support, or we build with your access.
- Build follows the signed-off blueprint.
- Weekly review of progress.
Deliverable
A working workflow.
Validate & hand over.
- Test real cases, refine edge cases.
- Measure output quality.
- Document and hand over.
Deliverable
An owned system + next-step plan.
There’s a walk-away point after Phase 1. Take the blueprint and stop before any build money is spent - you’ve lost nothing but two weeks.
Why I can help
Twenty years deciding what to build before building it.
AI is new. The hard part isn’t. For twenty years I’ve worked across product, design, and software delivery - understanding users, finding the right problem, defining what should be built, and getting systems shipped. Six Week AI applies that same product discipline to AI adoption.
01
Product thinking
02
Workflow design
03
AI expertise
04
Software delivery
Product leader on the platform team behind its workflow engine.
Consumer-scale product design, for millions of users.
Service-design rigor, at national scale.
Evaluating this with your CIO or engineering team? Technical considerations →
What it costs
No mystery about cost.
Start with a free conversation. If we both believe there’s a workflow worth building, we’ll move to discovery.
Start here · The fit call
Free
A 20-minute conversation
A chance to get to know each other, hear the challenge you’re facing, and get a sense of how I’d approach it. No prep, no pitch.
Book a callThe workshop
$2,500
A working session with your team
We map how the work really happens, find where the friction is, and score the one workflow most worth building. You leave with a plan, yours to keep either way.
Book a callThe build
$25,000
The six-week build, three phases
Diagnose, build, hand over - one workflow shipped on your data, fixed fee. Walk away free after Phase 1 if the blueprint isn’t worth building.
Book a callStill deciding?
The questions buyers actually ask.
Why would I not just ask my engineering team to build this?
You might. But engineering is usually not the bottleneck. The hard part is choosing the right workflow, redesigning it around what AI can now do, identifying the gaps before implementation, and producing a specification your engineers can confidently build from. Once that work is done, engineering becomes dramatically easier.
How do you handle our data and IP?
Everything stays in-house. Your data never leaves your systems, and the workflow, prompts, and tests are all yours. I work read-only and never touch production.
What do you need from our team?
One engineer, part-time, across the six weeks, plus about 8 to 12 hours total from the wider team for the discovery sessions and reviews.
What if the workflow turns out not to be buildable?
You find that out in Phase 1, at the walk-away point, before any build money is spent. You keep the blueprint and stop. The whole structure protects you from sinking the budget into a bet that wasn’t there.
More questions: which department to pick, models and tools, pricing, scope, and timing. Read the full FAQ →
Find the first workflow worth building.
Book a 20-minute call. We’ll talk through where your team is losing time, whether AI is actually a good fit, and what would need to be true before anything gets built. Whether we work together or not, you’ll leave with a clearer next step.