Natural Language

Introduction

Natural Query Language is an interface between the user and company connected to Tally.NET Server. It uses plain simple English to query and extract information while you are away from the business location. In short Natural Query Language is the way of communicating naturally following the grammar or syntax of the respective language or the way communication happens between two people.

To query information from the  connected company, the user needs to type the required words using the mobile phone and send it to the number provided. The system is designed to extract information from the respective company connected and sends back the information to the requester's mobile.

How it works

Tally.ERP 9 understands your query which is in the form of Natural Query Language. It has a set of defined rules and uses plain simple English to communicate. When a user requests for information using via SMS. Tally.ERP 9 receives the query and converts it to an understandable format which in turn extracts the required information from the connected company and responds back to the respective user with information in readable format.

Rules

To use SMS Support you need to follow the set of rules shown below:

• The Natural Query Language is capable of delivering a single value for a given query, It will not deliver multiple values as the results delivered need to be within limit of characters specified for SMS. You cannot request for a list of pending bills as it may exceed the specified limit.

• Short message service response is defaulted to the previous object. Most objects in Tally.ERP 9  are enabled for access via Natural Query Language Interface. While querying for Company details without any specification, the system defaults the query to a selected Company. On further querying for address, the system defaults it to the company address and returns the Company Address to the requester mobile.  E.g.: The previous query requested for party balance followed by another query for the address. The system returns an SMS with Party Address.  

• While querying the user must provide correct name of the master. Adding master type prefix would help Natural Query Language Interface to understand the query better.

• Consider a situation where a company has Sundry Creditor Ledger named Ajay Kumar and another cost centre/employee named Ajay Kumar. When a query is posted, Natural Query Language Interface finds it difficult to handle the request. By default, the system tries to deliver the best match or information encountered first. To overcome this prefix Master Type (Ledger) to Ajay Kumar in order to differentiate between the employee/cost centre and Ledger.   

• When a user queries for Ajay Kumar the system tries to find and lists all matching objects  such as Ajay, Ajay Malik, Ajay Kumar, Kumar, Kumar Krishnan etc. This can be avoided by enclosing Ajay Kumar within double quotes which makes it specific to deliver.

• Link Masters like Bills, Batches, Orders etc. contain numbers as part of their names. This causes ambiguity and the query may yield multiple results. To deliver accurate results all link masters must be prefixed with the link master type. Example: "Bill 0001 Pending Amount", Batch 01234 Expiry Date etc.

• When two masters/objects are queried the system considers them as primary and secondary returning cross values. E.g.: Closing Balance of Ledger Expenses for Cost Centre Mumbai Branch, the system currently supports Cost Centre/Ledger and vice versa as cross masters.

Quick Tips

• Remove the SMS signature from your mobile phone. The signature is appended to the end of the message sent and makes it difficulty for Natural Query Language Interface to understand.  

• While referring to a Masters it is essential to prefix the Master Type.  E.g.: Ledger Cash.

• Use the full names rather than partial names which result in wrong information  E.g.: Querying for ABC & Co rather than ABC may provide wrong or inadequate information.

• What you see is what you get (WYSWYG). Information you see is also the possible source of information in Tally.ERP 9. E.g.: You see the ledger address in Ledger Master screen, similarly you see the Return on Investment in Ratio Analysis report. While querying it is essential to use the exact names that you see as these titles or prompts return a value. Even though vocabulary is growing to support similar names, sometimes the Natural Language Interface may not understand what you had asked for.

• Be straight! Do not use words that are not necessary. Ask Tally.ERP 9 what it has or what it knows. When the information you are querying is not forming a part of the data then Natural Query Language Interface will not be able answer the query posted.

For an easy understanding of queries using Natural Language on various areas of the product we have broadly classified them into the following categories:

• Business

• System