Connect and stay in control

Connect the software you already use: a handoff checklist

Plan a connection between two tools you already run so that records move once, in the right direction, with an owner, a failure route and a way to check it worked.

For: Buyers and managers with an established software stack who want the tools to work together without replacing them.

Format: Guide + integration-planning sheet
Published: September 9, 2026
Last checked: September 9, 2026
Written by: Dryvn AI editorial
Edited by: Dryvn AI editorial
Reviewed by: Not yet reviewed by a named person
Read as: Editorial (Dryvn's practical method)

Keep what works, connect what does not

Most businesses we talk to have already been through two or three rounds of new software. Each tool fixed one thing and created a new place to type. The calendar knows the appointment, the CRM knows the customer, the accounting package knows the invoice, and a person carries the details between them by hand. BDC's guidance on digital improvement makes a point we agree with: build on the proven core software you already have before adding more [S10]. The problem is usually the handoff, not the tool.

The research position behind connected handoffs is set out in The AI Fabric: loops, graphs and everyday coordination.

An integration is just a handoff done by software instead of a person: when something happens in one tool, a record moves to another. The reason integrations disappoint is that people design the movement without designing the decisions underneath it. This checklist is those decisions.

Eight questions to answer before you connect anything

01Which system is authoritative?For each kind of record, pick one tool that holds the truth. Customer contact details might live in the CRM; the invoice total lives in accounting. Every other tool holds a copy. If two tools both 'own' a field, you have already lost.
02What moves in each direction?Write the direction down per field. Customer name goes CRM to accounting, never back. Invoice status goes accounting to CRM, never back. Two-way sync sounds convenient and is where most conflicts come from.
03How is a record identified?A customer called 'J. Smith' in one tool and 'Smith, John' in another is two customers to a computer. Pick one identifier (an ID, an email, a phone number) that both tools carry, and decide what happens when it is missing.
04Who owns the handoff?One named person who is told when it breaks and has the authority to fix or pause it. 'IT' or 'the vendor' is not an owner.
05What permissions does it need?Grant the connection the least access that does the job. Reading appointments is not the same as creating them. Use the approval boundaries worksheet to sort what it may do.
06How are duplicates caught?Decide what the integration does when the record already exists at the destination: update it, skip it, or flag it for a person. 'Create another one' is the default in most tools and the source of most mess.
07What happens on error?The connection will fail: the destination is down, a field is missing, a permission expires. Decide where failed items go, who is notified, and how fast. A silent failure route is the most expensive kind.
08What is the manual fallback?When the integration is paused, how does the business keep working? Write the manual steps down and test them once. If there is no fallback, the integration has become a dependency you did not choose.

A worked handoff, and a broken one

Illustrative example · fictional

Quote approved in the CRM becomes an invoice in the books

A fictional mobile detailing business uses a CRM for quotes and an accounting package for invoices. The owner wants an approved quote to become a draft invoice without re-typing. Here is the handoff on paper before anyone touches a connector.

One row of the planning sheet, filled in
Field	Decision
Business event	Quote status changes to 'Approved' in the CRM
Source record	CRM quote, identified by the quote number and the customer's email
Destination	Accounting package, draft invoice
Data fields	Customer name, email, line items, amounts, quote number (as a reference)
Direction	CRM to accounting only. Invoice status does not flow back in version one.
Authorized action	Create a draft invoice. Never send it. Never create a customer if the email is not found; flag instead.
Owner	Dana (office), notified by email on any failure
Failure route	Failed items land in a 'needs attention' list Dana checks each morning; the quote stays approved in the CRM
Completion check	Weekly: count approved quotes and count draft invoices with a quote reference. They should match.

The key choices are the small ones. Draft, not sent, keeps a person in the loop on money. 'Flag instead of create' when the customer isn't found prevents duplicates. And the weekly count is how Dana knows it is working without trusting the tool's own dashboard.

Illustrative example · fictional

The same business, done the fast way

Imagine instead the owner turns on a two-way customer sync between the CRM and accounting because the connector offered it. Customers in the CRM have phone numbers as their main identifier; accounting uses email. The first sync creates a second copy of eighty customers in each tool. Dana, not knowing which copy is real, updates one address in the CRM. The sync pushes it across to the wrong copy. Two months later, the wrong customer gets a statement.

Nothing technical failed. The integration did exactly what it was told. What was missing was the decision about which system was authoritative and how a customer is identified. Cleaning up eighty duplicates by hand cost more time than the sync ever saved.

Fill in the planning sheet

One row per business event. If you cannot fill in the failure route and the completion check, you are not ready to connect that event yet, and that is useful to know before you pay for anything.

Worksheet

Integration planning sheet

Complete one of these per business event before configuring any connector or automation.

Business event: What happens that should trigger the movement
Source record: Which tool, which record, identified how
Destination: Which tool, which record type
Data fields: Exactly which fields move
Direction: One way, and which way
Authorized action: Create, update, flag; and what it must never do
Owner: Named person who is notified and can pause it
Failure route: Where failed items go, who sees them, how fast
Completion check: How you confirm it worked, and how often

Downloads · no email required

Integration planning sheet (blank)
One row per business event with the nine planning columns. Opens in any spreadsheet.
Download .csv
Integration planning sheet (fictional sample)
Three fictional rows showing the level of detail that makes a handoff safe to automate.
Download .csv

After it is connected

Run the completion check by hand for the first month. Counts should match. If they don't, find out why before adding the next event.
Keep the manual fallback written down and reachable. Vendors change APIs, permissions expire, subscriptions lapse.
Add one event at a time. The second integration is easier once the first one has an owner and a failure route that works.
Revisit the direction decision when a tool changes. A new feature in one tool often tempts a second 'owner' back into the picture.

Questions people ask

What should I check before automating a handoff between apps?: The eight questions above, in order: authoritative system, direction, identifier, owner, permissions, duplicates, error route, manual fallback. The handoff assessment helps you find which handoffs are worth the effort.
Should I use two-way sync?: Rarely in version one. Two-way sync requires that both tools agree on who owns each field and how conflicts are resolved. Start one-way, prove the counts match, then add a second direction for specific fields if you still need it.
Do I need an AI tool to connect my software?: No. Most record movement is fixed automation: when X happens, copy these fields there. AI becomes useful when the input is messy, such as reading a photo of a receipt or a free-text request. The tool categories guide explains the difference.
What if my tools don't offer a connector?: Then the manual fallback is the process, and the planning sheet still helps: it tells the person doing the handoff exactly which fields to move and how to check it. A documented manual handoff beats an undocumented automated one.

Where is work getting stuck between your tools?Find the handoffs in your business that still depend on copying information, remembering follow-ups or chasing people, and get one next action for each gap.
What should AI be allowed to do without asking?Decide, action by action, what an AI assistant or automation may do on its own, what it may do inside an approved scope, and what always waits for a person.
What should I automate first in a small business?Compare recurring work on six practical measures, sort each task into the right lane, and pick one bounded experiment with a baseline.
BlogThe AI Fabric: A Research Thesis on Loops, Graphs and Everyday CoordinationBrent develops a proposed AI fabric connecting people, context, permissions and ongoing work. The paper examines its research foundations, limits and a method for testing whether it reduces human coordination effort.
BlogWhat would have to change for AI to run more of the working day?A scenario analysis, not a forecast: five conditions that would need to hold before AI carries most of an ordinary business day, and what evidence would tell you each one is arriving or stalling.

Sources

Dates are when each source was last checked by the editor. Sources support specific claims; they are not endorsements.

S10How to complete a digital transformation in the age of AI · BDC · checked September 9, 2026Supports building on proven core software before adding new tools, and fixing the process before connecting systems.
S19Integrations: plug Dryvn into the tools you already use · Dryvn AI · checked September 9, 2026Dryvn's public page distinguishes Live, In rollout and Planned connectors; as of the checked date only Stripe and SMS/MMS are badged Live. Cited as an example of status labelling, not as evidence of any specific capability for your business.

R10 · Published September 9, 2026 · Next scheduled review December 9, 2026 · Teaches process management; not legal, warranty, safety or engineering advice. Examples are fictional unless stated. Part of the Dryvn resource library (17 resources).