About Bad Clause
The old way is optional.“That is how we have always done it” is an expensive sentence. Bad Clause helps businesses replace the repetition they have learned to live with.
Find the bad clause
The point of view
They build relationships, make judgments, and solve problems that do not fit a dropdown. They should not spend the day moving the same information between tools.
We work at the intersection of process design and implementation: understand how the business works, simplify the handover, and use automation where it earns its place.
Sometimes that means AI. Sometimes it means a clear rule, a better form, or a person finally owning the next step.
The Bad Clause manifesto
01. Start with the business problem.
02. Keep people in control.
03. Make failures visible.
04. Measure the actual change.
05. Leave it understandable.
Why Bad Clause?
“Someone will update the sheet.” “We’ll follow up tomorrow.” “The team knows what to do.”
Those sentences become unwritten operating clauses: repeated, accepted, and expensive. We find the clause, rewrite the process around reality, and leave the operating terms clear.
The name is a reminder to question the rule everyone stopped noticing.
Founder-led. Direct by design.
Bad Clause brings together Gaurav’s work across product design, customer experience, and software implementation.
The approach starts with how people actually use a process, including the workarounds, delays, and judgment calls the official diagram leaves out.
Based in India. Built for remote collaboration. You work directly on the problem, the decisions, and the result.
Start a conversationThe practice behind Bad Clause
Bad Clause is an independent AI automation practice built around a simple position: the useful part of automation is understanding the work well enough to change it without breaking the business.
The approach draws on product design, customer experience, business development, and software delivery. That combination matters because the workflow has to make sense to the person doing the job, the customer receiving the outcome, and the team responsible for the systems.
People work around software when the official process is slow or incomplete. We examine the emails, spreadsheets, reminders, copy-paste steps, and informal handovers that keep the work moving before deciding what the future process should be.
The business needs to understand outcomes, risks, cost, and ownership. The implementation needs precise events, fields, permissions, states, and failure behaviour. Bad Clause connects those two conversations without hiding behind AI vocabulary.
Classification, extraction, retrieval, drafting, and bounded reasoning can be useful. Fixed routing, calculations, validation, and approvals are often better handled by explicit logic. The design should make that difference visible.
A compelling demonstration can ignore messy data, access expiry, duplicate events, user corrections, and the person who receives an alert at 9:00 on Monday morning. Our standard is whether the workflow can be operated after the presentation ends.
Recommendations are not tied to selling one automation platform or AI model. Tools are selected during scoping based on the workflow, existing systems, access, risk, maintainability, and realistic operating cost.
Bad Clause will not manufacture client results, guarantee revenue, or describe a prototype as a production system. Evidence begins with a baseline and becomes credible only after the workflow operates in the real environment.
Enough circling back.
One conversation. One workflow. A clearer way forward.
Let’s kill the busywork