Skip to content
GullySales

Every enquiry is read, ranked and qualified before your seller picks up the phone.

Gully Sales writes the qualification criteria your business actually uses, puts a model to work reading each enquiry against them, and keeps a person in charge of what the model decides.

  • A written qualification summary on every enquiry, before the first call.
  • A score that learns from the enquiries your team went on to close.
  • A human review step, so no lead is turned away by a machine alone.
Book a free auditTalk to us

Gully Sales Private Limited works with small and medium businesses across India, and every model decision stays visible to your own team inside your CRM.

In one paragraph

What is AI Lead Qualification and Scoring Services?

AI lead qualification and scoring reads each enquiry, its wording, source and behaviour, judges it against criteria you have agreed, and ranks it for your team. Gully Sales defines those criteria, builds the model into your CRM, writes the response and handoff standards around it, keeps a person reviewing its decisions, and retrains it on what you closed.

The problem

Qualification lives in one person's head, and your response time shows it.

Enquiries reach you in whatever form the buyer feels like using: a form with three lines of text, a WhatsApp message with a photograph attached, a missed call, a reply to an email campaign. Somebody has to read each one, decide whether it deserves a seller's time, and act before the buyer goes elsewhere. In most growing businesses that somebody is already doing three other jobs. So the reading gets shallower as volume rises, the judgement changes with whoever is on duty, and the enquiry that mattered is recognised a day late.

You will recognise it as

  • The only way anyone can judge an enquiry is to call it, so every enquiry costs a call whether it deserved one or not.
  • Two people read the same enquiry and reach different conclusions, and neither can point to a written rule.
  • Enquiries that arrive at night or over a weekend are read when the office reopens.
  • What was asked and answered on a qualifying call stays in a notebook, a chat thread or somebody's memory.
  • Leads are marked unqualified with no reason recorded, so marketing keeps sending more of the same kind.
  • Your longest-serving seller qualifies well, nobody can explain how, and the skill cannot be handed to anyone else.

What it costs the business

  • Expensive selling hours go on discovering, one call at a time, what a careful reading could have told you in seconds.
  • Serious buyers judge you by how long you took to reply, and that judgement is formed before you ever speak to them.
  • Marketing receives no honest signal about which enquiries were worth having, so spend keeps following volume.
  • Qualification quality rises and falls with who is on duty, and drops whenever that person is on leave or resigns.

Why it persists. Qualification is judgement work, and judgement is hard to write down. The criteria sit in the experience of one or two people who are too busy selling to document them. AI, meanwhile, looks like a project for a large company with a data team, or like a closed box that will start rejecting good enquiries without explanation. So the sensible-sounding decision is to keep reading by hand, and the cost stays hidden, because nobody counts the enquiries read too late or read too quickly.

If it stays unresolved. Volume is the thing that breaks it. Every additional enquiry makes the reading shallower and the queue longer, so the moment a business most needs sound qualification is the moment it does it worst. Teams in this position usually add people to the problem, which multiplies the inconsistency instead of removing it, and the enquiries nobody read are never counted.

What changes

What changes when every enquiry is read properly, every time.

In the first weeks

  • Each enquiry carries a short written summary of what was asked, by whom, and how ready it looks.
  • Your team sees a rank and a stated reason on the record, not a raw list in arrival order.

In how the work runs

  • A written response standard, so nobody has to decide afresh who replies and how quickly.
  • One set of qualification questions, asked the same way whoever is on duty.
  • Qualified leads reach sales with the meeting scheduled and the context already attached.

In sales and marketing

  • Selling hours move towards enquiries that resemble the customers you already serve well.
  • Enquiries arriving outside office hours are acknowledged and read without waiting for Monday.

In what management can see

  • A stated disposition on every enquiry, so you see why leads were declined, not only how many.
  • A weekly view of qualification volume, response speed and how often the model was overruled.

Over the longer term

  • Qualification knowledge held in a written standard and a trained model, not in one person's memory.
  • A model recalibrated against closed business, so it ages with your market instead of against it.

Gully Sales controls the criteria, the model, the standards, the CRM build and the review that keeps it honest. Whether more enquiries become orders also depends on your offer, your pricing, your capacity to follow up and market conditions.

Who it is for

This fits businesses with more enquiries than reading time.

The businesses it suits

  • Businesses receiving enough enquiries each month that nobody can read all of them carefully.
  • Teams whose leads arrive as free text across web forms, WhatsApp, email and phone.
  • Companies with a CRM in daily use, where enquiries are recorded rather than kept in inboxes.
  • Owners who want qualification to survive a seller resigning or taking leave.
  • Businesses holding a record of past enquiries and what happened to them, even an imperfect one.
  • Teams already working to a written sales process who want the front of it to keep pace.

What usually prompts the call

  • Enquiry volume has grown faster than the number of people reading them.
  • A serious enquiry was answered late, and the reason was simply that nobody had read it.
  • Marketing and sales disagree about lead quality and have no shared evidence either way.
  • An experienced seller has left, taking the qualification judgement with them.
  • You want a first use of AI in the business that produces a result you can measure.

What Gully Sales does

The work, component by component.

Qualification criteria

We write down what makes an enquiry sales-ready for your business: the fit signals, the disqualifiers, the questions that must be answered before a seller is involved, and the difference between an enquiry worth a call today and one worth nurturing. The criteria are agreed with the people who sell, and tested against enquiries you have already won and lost.

Why it matters:
A model can only judge against a standard someone has stated. Written criteria also settle the marketing-and-sales argument before it starts.
You receive:
A qualification criteria document with definitions, disqualifiers and worked examples.
Business value:
Everyone judges the same enquiry the same way, whether it is read by a person or by the model.

Response standards

We set how quickly each type of enquiry is acknowledged, on which channel, by whom, and what happens outside office hours, on holidays and when the named owner is unavailable. The standard covers the first acknowledgement, the qualifying attempt, how many attempts are made and when an enquiry is parked rather than chased.

Why it matters:
Speed of first response is one of the few things about an enquiry you fully control, and buyers read it as a signal about how you will serve them.
You receive:
A written response standard with channels, owners, timings and escalation.
Business value:
Enquiries stop waiting to be noticed, including the ones that arrive at inconvenient hours.

AI enquiry reading and scoring

A model reads each enquiry the way an experienced person would: the wording of the message, the source and campaign behind it, the company that can be established from the email domain, the pages viewed, and your own history with that contact. It writes a score, a short summary and the reasons for both onto the lead record.

Why it matters:
The information that decides an enquiry is mostly unstructured text, which a rule-based score cannot read.
You receive:
A configured scoring and summarisation model, with its prompts, thresholds and reason codes documented.
Business value:
Your team starts each conversation already knowing what the enquiry says and how it compares with past business.

Qualification scripts

Question sets for each enquiry type, written so a junior team member can run them: the opening, the qualifying questions in order, how to handle the objections that appear at this stage, and how to close the call towards a meeting or a courteous decline. Scripts are drafted separately for phone, WhatsApp and email, because the same words do not work on all three.

Why it matters:
A model can rank an enquiry, but a person still has to ask the questions that confirm or overturn the rank.
You receive:
Channel-wise qualification scripts and objection responses.
Business value:
New joiners qualify at a workable standard from their first week rather than their sixth month.

Scheduling and meeting confirmation

The path from a qualified enquiry to a confirmed meeting: how slots are offered, who owns the calendar, how the invitation and reminders are sent, and what happens when a prospect does not appear. The steps sit inside your CRM, so the meeting, the notes and the lead stay on one record.

Why it matters:
A large share of qualified leads are lost between the qualifying call and a meeting nobody confirmed.
You receive:
A scheduling and confirmation workflow, with reminder templates and a no-show routine.
Business value:
More of the leads your team qualified turn into meetings that actually take place.

Handoff and disposition

A defined handoff record: what the seller receives, in what form, and what marks acceptance or return of a lead. Alongside it, a short list of disposition codes so every enquiry ends in a stated outcome, qualified, nurturing, not now, out of scope or no contact made, with a reason attached to each.

Why it matters:
Without a stated outcome on every enquiry there is nothing to learn from and nothing honest to report.
You receive:
A handoff standard and a disposition code list configured in your CRM.
Business value:
Leads stop dying quietly between two teams, and every enquiry leaves a trace worth reading.

Feedback loop and recalibration

A review rhythm where the model's calls are compared with what happened: which high-scoring enquiries closed, which low-scoring ones were later won, and which summaries misread the buyer. Criteria, thresholds and prompts are corrected on that evidence, and every change is recorded so you can see how the system moved and why.

Why it matters:
A qualification model that is never corrected drifts away from your market within months.
You receive:
A calibration review pack and a change log of every adjustment made.
Business value:
The system gets closer to your business over time instead of quietly going stale.

What you will have at the end.

  • A qualification criteria document, with fit signals, disqualifiers and worked examples from your own enquiries.
  • A written response standard covering channels, owners, timings, after-hours cover and escalation.
  • A configured scoring model inside your CRM, with thresholds and reason codes written to every record.
  • An enquiry summary generated onto each lead, in a format a seller can read at a glance before calling.
  • Channel-wise qualification scripts for phone, WhatsApp and email, with objection responses.
  • A scheduling and confirmation workflow, including reminder templates and a no-show routine.
  • A handoff standard defining what sales receives and what counts as acceptance or return of a lead.
  • A disposition code list, configured as CRM values rather than free text.
  • A qualification dashboard showing volume, response speed, accepted leads and reasons for decline.
  • A calibration pack comparing the model's calls with closed business, plus a change log.
  • A training session and a one-page reference for the people who qualify enquiries.
  • An anonymised sample of a scored enquiry, so you see the output before the process goes live.

How it runs

The engagement, step by step.

  1. 1

    Assessment

    We take a sample of your recent enquiries across every channel and read them ourselves, alongside what happened to each one. We look at where enquiries land, how quickly they were answered, what was recorded, and how many were judged on evidence rather than on feel. The output is a picture of your qualification as it stands, in numbers you can check.

    You provide:
    Access to your CRM or lead records, a sample of enquiries, and an hour each with the people who qualify today.
    We produce:
    A qualification assessment: response speed, data completeness, disposition quality and the gaps between them.
    Done when:
    You have agreed the picture of where enquiries are being lost.
  2. 2

    Criteria and standards

    A working session with the people who sell, to state what makes an enquiry sales-ready in your market. We turn that into written criteria, disqualifiers, response standards and disposition codes, then test them against real enquiries from the assessment sample so the rules meet reality before anything is built.

    You provide:
    Sales and marketing participation in the session, and a decision-maker who can settle disagreements.
    We produce:
    Agreed qualification criteria, disqualifiers, response standards and a disposition code list.
    Done when:
    Two people can read the same past enquiry and reach the same conclusion.
  3. 3

    Model build and calibration

    We configure the model against your criteria and run it over historical enquiries whose outcome is already known. Its calls are compared with reality, prompts and thresholds adjusted, and the cycle repeated until it agrees with your experienced people often enough to be useful. Where it disagrees, the reason is recorded rather than smoothed over.

    You provide:
    Historical enquiries with outcomes attached, and time from an experienced seller to judge disputed cases.
    We produce:
    A calibrated scoring and summarisation model, with prompts, thresholds and reason codes documented.
    Done when:
    The model's calls on past enquiries have been reviewed and accepted by your team.
  4. 4

    Build inside your tools

    The model, the scripts, the scheduling workflow, the handoff record and the disposition codes are built into the CRM and channels your team already opens. Nothing runs in a separate tool people have to remember, and every field the system writes is visible on the lead record where a seller can question it.

    You provide:
    Administrator access to your CRM and connected channels, and a nominated internal owner.
    We produce:
    A configured qualification workflow in your CRM, with dashboards, templates and permissions in place.
    Done when:
    A test enquiry travels from arrival to scheduled meeting without anyone rescuing it by hand.
  5. 5

    Supervised pilot

    The system runs on live enquiries with a person checking every model decision before it is acted on. Disagreements are logged and used to correct criteria and prompts. Only when the review shows steady agreement do we reduce checking to a sample, and declines remain open to inspection after that.

    You provide:
    A reviewer for the checking period, and honest feedback whenever the model reads an enquiry wrongly.
    We produce:
    A pilot log of decisions, overrides and corrections, and a revised model built from it.
    Done when:
    Your reviewer is satisfied that the model is not turning away enquiries you would have kept.
  6. 6

    Handover and review rhythm

    We train the people who will run it, hand over the documentation, and set the review rhythm: what is checked, by whom, and how a change is proposed and recorded. If you want us to keep running the calibration, that is a separate ongoing arrangement rather than something the design assumes.

    You provide:
    The team members who will own the process day to day, and time for the training session.
    We produce:
    Training, a written runbook, the calibration schedule and the change log format.
    Done when:
    Your team can run, correct and explain the system without us in the room.

Ways to work with us

Ways to work with us on qualification.

Qualification assessment

A standalone review of how enquiries are read and answered today, with the gaps named and a recommended order of work. Some businesses stop here and act on the findings themselves.

Build and supervised pilot

The full sequence: criteria, response standards, scripts, model build, CRM configuration and a supervised pilot, ending with your own team running it and able to change it.

Ongoing calibration

A continuing arrangement where we compare the model's calls with closed business, correct criteria and prompts, and report on qualification quality each period.

Qualification desk support

Where nobody is available to read enquiries at all, our team runs the qualifying contact to your standards while your internal owner is recruited or trained.

Why Gully Sales

What you are actually choosing when you choose us.

We sell for a living, not only advise.

Gully Sales builds sales and marketing systems for Indian SMBs, so the criteria we write are ones a seller can apply on a live call, not a framework borrowed from a large enterprise playbook.

The model is built on your evidence.

Nothing is switched on until it has been tested against your own past enquiries and their outcomes. Where your history disagrees with the theory, the history wins and the model is changed.

A person stays in charge.

Every score carries a reason, every decline stays visible, and overrides are logged. We do not put a machine between you and a buyer without a way to see what it decided and to overturn it.

It is built inside the tools you already use.

The scoring, the scripts and the handoff live in your CRM and your existing channels. There is no extra screen for people to remember, and no dependence on us to read the output for you.

You own everything we write.

Criteria, prompts, thresholds, scripts, dashboards and the runbook are handed over as your documents. If the engagement ends, the system keeps working and your team can change it.

Where it applies

The same service, in different businesses.

Industrial and engineering supply

The situation:
Enquiries arrive from the website, from marketplaces and by phone, mixing genuine project buyers with students, job seekers and traders collecting price lists.
How it applies:
Criteria separate a project requirement from a price request, and the model reads the message and its source so only real requirements reach the sales engineer.
Likely benefit:
Technical selling time goes to requirements instead of to filtering an inbox.

Healthcare clinics and hospitals

The situation:
Patient enquiries come through forms, calls and messages at all hours, and the front desk is busiest exactly when enquiries peak.
How it applies:
Response standards cover after-hours acknowledgement, scripts keep the questioning consistent and appropriate, and urgent wording is flagged for immediate human attention.
Likely benefit:
Patients receive a prompt, consistent first response, and staff attention goes where it is needed most.

Interior design and home services

The situation:
A site visit costs most of a working day, and a short web form rarely reveals whether the enquiry is a funded project or an early idea.
How it applies:
Qualification questions establish scope, timeline and who decides before a visit is offered, and scoring ranks enquiries by how closely they match completed projects.
Likely benefit:
Fewer wasted site visits, and design time reserved for enquiries that are ready to move.

B2B services and consulting

The situation:
Many enquiries come from individuals and very small firms outside the service range, and each still takes a conversation to identify.
How it applies:
Disqualifiers are written explicitly, applied with the reason recorded, and out-of-scope enquiries receive a courteous standard reply rather than silence.
Likely benefit:
Senior consultants meet fewer unsuitable prospects, and unsuitable enquiries still get a decent answer.

Education and training institutes

The situation:
Admission enquiries arrive in seasonal waves, and the counselling team cannot read them all carefully during the peak weeks.
How it applies:
Enquiries are scored on programme fit and readiness, counselling slots are scheduled automatically, and disposition codes record why an enquiry did not proceed.
Likely benefit:
Counsellor hours in the busy weeks go to candidates closest to enrolling.

Proof

Work we can point to.

Premier Marketing

The problem:
Enquiries were arriving from both domestic and industrial buyers, and the handling of each one after it landed needed tightening before reach could turn into sales activity.
What we did:
Gully Sales built a complete website for Premier Marketing and worked on what happened to an enquiry once it arrived, across the domestic and industrial sides of the business.
The result:
The published case study reports enhanced reach, streamlined lead handling and accelerated sales across those sectors.
Read the case study

Questions buyers ask

Before you enquire, the answers you will want.

What makes an enquiry or meeting sales-ready?

Sales-ready means the enquiry meets criteria your own team has agreed, not a general industry definition. Usually that covers fit, meaning the kind of business, the requirement and the geography you serve, and readiness, meaning a stated need, a rough timeline and a person who can decide or influence. We write these with your sellers, test them against enquiries you have already won and lost, and record the disqualifiers just as explicitly as the qualifiers.

How is this different from rule-based lead scoring?

Rule-based scoring ranks structured data: source, industry, pages viewed, form fields. This service adds reading. A model interprets the unstructured part of an enquiry, the wording of the message, the notes from a call, the thread of an email, and produces a summary and a judgement alongside the score. The two work well together, and many businesses start with written rules and add model-based reading when enquiry volume outgrows them.

How long before the model reads live enquiries?

It depends on four things: how many channels your enquiries arrive through, whether a CRM is already in daily use, how much enquiry history exists with outcomes attached, and how quickly your team can meet for the criteria session. The assessment is short. The build waits on your decisions more than on us. The supervised pilot runs until your reviewer is satisfied with the model's calls, which is a judgement rather than a date we can fix in advance.

What history do you need to calibrate the model?

Access to your CRM or wherever leads are recorded, a sample of recent enquiries across every channel with what happened to each, an hour or two from the people who qualify today, and a decision-maker who can settle disagreements about criteria. During the pilot you need one reviewer willing to check the model honestly. If your records are patchy we work with what exists and say so plainly rather than pretending otherwise.

How do we check the model is qualifying correctly?

Against your own starting position, recorded before we build anything. We report data completeness, time to first response and to the first qualifying attempt, conversion at each lifecycle stage by score band, how many enquiries flow through without manual rescue, how often your reviewer agrees with the model, and hours saved in reporting. Conversion figures are read only after a full sales cycle, because a lead qualified this month closes later.

What is outside the scope of qualification work?

We do not generate the enquiries; demand generation is separate work. We do not run your sales conversations after handoff, write proposals or negotiate. We do not replace your CRM, though we will tell you if the one you have cannot support the process. We do not supply purchased contact lists. And the model does not message your customers by itself: drafting is assisted, sending stays with a person unless you decide otherwise.

Do we have enough data for this to work?

Often more than you think, but it is worth checking before committing. The model needs written criteria more than it needs volume, and calibration needs a set of past enquiries with known outcomes, won, lost, or never genuine. If your history is thin we start with the criteria and the response standards, which help straight away, and add scoring once enough outcomes have accumulated to judge it fairly.

Will the AI speak to our customers directly?

Not unless you decide it should, and that is a separate decision taken with you rather than an assumption in the design. In the standard build the model reads, summarises, scores and drafts. A person makes the call, asks the qualifying questions and sends the reply. Where you want automated acknowledgements, we write them, you approve them, and they are clearly identified so a buyer is never misled.

4 more questions

What happens when the model reads a lead wrongly?

It will, and the process expects it. Every score carries a reason, so a reviewer can see the logic and overturn it. Overrides are logged rather than corrected quietly, and that log is what we use in calibration to fix criteria, thresholds or prompts. Declined enquiries stay visible and searchable instead of disappearing, so a wrong decline can be found and recovered while the buyer is still in the market.

Which CRM and tools does this work with?

Whatever you already use, provided it can hold custom fields, accept data from your channels and be reported on. Most Indian SMBs we work with run a mainstream CRM alongside WhatsApp, email, a website form and phone. If you have no CRM, that comes first as separate work. Putting a scoring model on top of scattered records produces confident answers resting on nothing anyone can check.

Is our customer data safe when a model reads it?

This is settled before anything runs. We agree which fields the model may see, where data travels, what is retained and for how long, and it is written into the build documentation rather than left to assumption. Access stays limited to the people and systems that need it. If your sector carries particular obligations, such as health or financial records, say so at the assessment and the design accounts for them.

What happens if we stop working with you?

You keep everything. The criteria, response standards, scripts, prompts, thresholds, dashboards and runbook are your documents, built inside your own CRM. Your team is trained to run and change them during handover, and the change log shows what was adjusted and why. The system does not need us to keep operating; the calibration reviews simply move to your people instead of ours.

Talk to us

Let us read a week of your enquiries with you.

The first conversation is a discussion, not a pitch. We look at how enquiries reach you, what happens to them and what your records show, then say honestly whether model-assisted qualification suits your volume or whether something simpler should come first.

  • No obligation and no sales script
  • A reply from someone who does the work
  • Your details are never sold or shared

Your details are used only to respond to your enquiry. We do not sell or share them, and anything you show us about your leads, customers or systems stays confidential.

Protected by reCAPTCHA — Google’s privacy policy and terms apply.

Get a free audit of how you sell, and a scored report of where the work is.

Book a free audit
Book a free audit