AI has become the most exciting business conversation of the moment. Founders want AI in their products. Leaders want AI in their workflows. Sales teams want AI-written messages. Support teams want AI summaries. Everyone wants speed, automation, and intelligence. But there is a critical mistake many businesses make: they try to add AI before fixing the foundation.
Limesh Parekh explains this with a memorable analogy: AI is the tadka, not the dal. The dal is the base meal. The tadka adds flavor, aroma, and power. But tadka without dal is not a meal. In business terms, AI is an enhancement. The foundation is clean data, clear process, disciplined workflows, and a team that understands how the business actually runs.
For SMBs and SaaS teams, this is one of the most practical AI lessons available.
Why AI Cannot Fix Chaos
AI can summarize, predict, generate, classify, recommend, and automate. But it depends heavily on the quality of the inputs it receives. If customer data is scattered across spreadsheets, WhatsApp chats, personal notebooks, and employee memory, AI has no reliable base to work from.
A messy system produces messy automation. If lead stages are unclear, AI cannot accurately predict pipeline health. If customer notes are incomplete, AI summaries will miss context. If service history is not captured, AI cannot identify recurring issues. If the team does not update records, AI recommendations become weak.
This is why businesses must avoid treating AI like a magic layer. AI does not replace operational discipline. It amplifies whatever discipline already exists.
The Dal: Process and Data
Before adding AI, a business needs its dal: the core operating system.
That includes:
Defined customer stages.
Clean lead and customer records.
Documented follow-up rules.
Clear ownership of tasks.
Consistent reporting.
Structured service history.
Reliable product, pricing, and support information.
A team trained to update the system properly.
In CRM terms, this means the business must first know who the customer is, what stage they are in, what has been promised, what happened last, and what should happen next. Without that, AI has no meaningful context.
A company that cannot track follow-ups manually will not suddenly become customer-centric because it buys an AI tool. It must first build the habit of tracking.
The Tadka: AI Enhancement
Once the foundation is strong, AI becomes extremely useful. It can help salespeople prepare for calls, summarize customer history, identify inactive leads, draft follow-up emails, detect churn risk, and support managers with pipeline insights. It can help customer success teams review tickets, prioritize escalations, and personalize communication.
But notice the sequence. AI works best after the CRM or operating system contains structured data. Then AI can enhance the work already happening.
For example:
If call notes are captured, AI can summarize them.
If lead sources are tracked, AI can identify better channels.
If deal stages are updated, AI can flag stuck opportunities.
If support tickets are categorized, AI can find recurring product gaps.
If customer usage is measured, AI can predict churn risk.
The quality of AI output follows the quality of business input.
Humans With AI vs. Humans Without AI
Limesh’s framing also avoids the common fear-based question of AI versus humans. The better comparison is humans with AI versus humans without AI.
AI will not remove the need for judgement, empathy, customer understanding, or leadership. But it will strengthen teams that already have process discipline. A salesperson who listens well and uses clean CRM data can become faster and more prepared with AI. A support agent who understands the customer can respond better with AI-assisted summaries. A manager who reviews data regularly can make sharper decisions with AI insights.
The advantage goes to people and companies that know how to combine human context with machine assistance.
Why SMBs Should Not Rush Into AI Features
SMBs often feel pressure to adopt every new technology quickly. But rushing into AI without process clarity can waste money and create confusion. Teams may start using disconnected tools, generating content without strategy, or automating messages that do not match the customer’s actual need.
Before investing heavily in AI, SMB leaders should ask:
Is our customer data clean?
Do we have one source of truth?
Are our sales stages defined?
Do employees update information consistently?
Do we know which process we want to improve?
Can we measure whether AI is helping?
If the answer is no, the first investment should be in process and data hygiene.
AI in CRM: Practical Use Cases
Once the foundation is ready, AI can add real value to CRM workflows.
Lead prioritization: AI can help identify which leads deserve immediate attention based on source, behavior, industry, or past conversion patterns.
Follow-up assistance: AI can draft contextual messages based on the last conversation and next action.
Pipeline risk detection: AI can flag opportunities that have been inactive too long.
Call and meeting summaries: AI can convert conversations into structured notes and action items.
Customer support insights: AI can detect repeated complaints or common implementation problems.
Manager coaching: AI can help identify sales patterns, lost-deal reasons, and team training needs.
These use cases are powerful only when data is consistently captured.
Do Not Use AI to Hide Weak Process
Some companies use AI as decoration. They add AI labels to products, campaigns, or workflows without changing the underlying customer experience. That may create short-term attention, but it does not create long-term value.
AI should solve a real process problem. It should reduce time, improve accuracy, personalize service, reveal insights, or help teams make better decisions. If the business cannot clearly name the improvement, it may be adding tadka to an empty plate.
The Bottom Line
AI is not the main dish. It is the enhancement that becomes powerful only when the foundation is strong. For SMBs, that foundation is clean data, clear process, disciplined CRM usage, and trained teams.
Businesses should not ask, “How do we add AI?” first. They should ask, “What process do we need to strengthen so AI can actually help?” Once the dal is ready, the tadka can make it exceptional.
FAQs
Why is AI compared to tadka?
Because AI enhances an existing business process, just like tadka enhances dal. It cannot replace the base.
Can AI fix poor customer data?
No. AI can help organize or summarize data, but it cannot produce reliable insights from incomplete or inconsistent information.
What should SMBs do before adopting AI?
They should clean customer data, define workflows, create one source of truth, and train teams to update systems consistently.
Source Note: Based on The Thrive podcast episode featuring Limesh Parekh of Enjay IT Solutions: https://www.thethrive.in/podcasts/from-inr-2-crore-loss-to-crm-success-limesh-parekhs-bootstrapped-journey-from-bhilad/
