Development guides / Business systems and workflow integration
Enquiries, quotes and orders: where should system integration begin?
Choose one process with a clear start and finish. Find where information goes missing, who takes the next step and how you know the work is complete. That gives you something concrete to test before making wider changes.

Why do spreadsheets become hard to coordinate?
A customer enquires through a site or email, staff copy the details into a spreadsheet, and another person creates the order after approval. Each step looks simple. Changes, absent owners and duplicate messages make it harder to know which record is current.
Before choosing software, follow one real enquiry through the people and tools involved. A common reference number, clear status and shared list of pending work may be the most useful place to start.
Improve the setup, connect tools or build a new system?
Check whether configuration changes in the existing CRM, spreadsheet or order system can solve the problem. Integrate tools when data needs to move between them. Consider a custom system for distinctive quotation rules, role permissions or cross-department approvals.
Decide how to identify the same customer or order, which system holds the accepted information and who handles failed transfers. Otherwise, automation can repeat existing mistakes and create duplicate records.
Improve the existing setup
When unclear fields, statuses and ownership cause most of the trouble, improve the current tool first. Establish a consistent record before deciding whether another connection is needed.
Connect two suitable tools
If both tools fit the business but staff repeatedly retype data, focus on the connection. Confirm supported read/write access, field mapping and recovery when a transfer fails.
Build a custom business system
When complex pricing, permissions or cross-department approvals do not fit existing tools sensibly, define one core workflow for custom development before expanding the system.
Map the data path of one enquiry
The example below illustrates an approach. It is hypothetical, not a client case study or a claim of delivered results.
Suppose a website enquiry becomes a quote and then an order. Create a request number, validate required data and assign an owner; create the order only after quote approval. A repeated notification updates the existing record. Missing information goes onto a follow-up list instead of being marked complete.
What happens when a notification repeats, a price changes or a connection fails?
Two notifications must not mean two orders
A notification can arrive twice after a temporary delivery failure. Use a reference and processing history to recognise work already done. A returning customer must still be able to place another order, so test the difference between a resent notification and a genuinely new request. A customer name alone cannot distinguish them.
Which system owns the price?
For each important field, decide which system’s value to use. Quotes can change before approval, but the approved version needs to remain traceable. If prices can change in both systems, agree the sync direction and what happens when values conflict. Later updates must not quietly overwrite an agreed price.
Someone needs to see failed work
Agree which failures can be retried automatically, which need more information and when to alert someone. Give staff a list showing the reason, original record and next step. A technical log alone does not help them finish the outstanding work.
Test failed transfers and retries too
- Prepare normal, incomplete, duplicate, cancelled and updated examples. Reconcile records and outcomes in both systems.
- Simulate a temporary connection failure. Confirm that the failure is visible, an alert is raised and a retry does not create a second order.
- Compare manual steps, missed records and exception-handling time for the same type of work before and after the change, then decide whether to expand.
Include data and maintenance in the estimate
The work depends on the tools’ connection options, data quality, approval rules and exceptions. If old records contain duplicates or missing values, allow separately for cleaning them up. Include subscriptions, usage fees and ongoing monitoring when comparing costs.
Pilot a workflow with enough activity to expose problems and a practical manual fallback. Before switching, agree when the old process stops, who reviews failures and how rollback avoids two systems creating the same order.
Pilot the change and measure completed work
Try one order type or a limited part of the process. Record handling time, repeated entry and missed work before the change, then compare similar work afterwards. Include time spent on exceptions. A successful data transfer does not necessarily mean the business task is finished.
Agree which new records enter the new process, how unfinished work continues and who takes over if processing stops. During parallel checks, the new system can compare results without creating live orders. This avoids both systems acting on the same request. Expand once the results have been checked.
Bring these examples to the first discussion
Prepare current tools, anonymised spreadsheet samples, fields, approval rules, actual daily workload and common exceptions. During scoping, agree historical migration, platform permissions, usage costs, monitoring and responsibility for later changes.
Questions before you start
If we already use automation tools, do we need development?
Check whether your existing tools can handle the rules and exceptions reliably. Simple field mapping may only need configuration. Complex permissions, conflicting data or retries that could repeat an action may require extra design and development.
Must we migrate all historical data first?
Not always. Decide which data the trial needs, how to continue unfinished work and where older records can be read. Treat a full historical migration as a separate task with its own checks.

