How to Run a Sales Tech Stack Audit
Your CRM holds the pipeline. Your calendar books the meetings. Your proposal tool sends the quotes, your e-signature app closes the contracts and your accounting platform raises the invoices. Yet the team may still struggle to say which deals are genuinely progressing, why some leads wait days for a reply, or how reliable this quarter's forecast really is.
A sales tech stack audit examines the systems behind those outcomes. It reviews your applications, data, processes, integrations, people, AI tools and automated workflows to establish how sales actually operates, from the first enquiry through qualification, proposal, signature and handover to delivery and finance.
The timing matters. CRMs such as Salesforce, HubSpot and HighLevel now bundle booking, calling, sequences, quotes, e-signature, payments and AI assistants, so many separate sales tools can be consolidated. At the same time, AI note-takers, AI email drafting and AI lead scoring are changing how much admin a salesperson needs to do, provided the underlying CRM data is reliable.
More than a subscription list
A tool can work perfectly while the process around it fails. A booking page can schedule a meeting successfully, for example, without linking it to the right deal, reminding the prospect or telling the salesperson what the lead enquired about. That is why the audit looks at how tools are used, connected and owned, not just what you pay for.
The seven areas of a sales audit
A complete audit examines seven connected areas:
Systems: the CRM and core platforms, and the sales data each one controls.
Processes: how leads move from enquiry to closed deal, and how deals hand over to delivery and finance.
Apps: functionality, usage, cost and overlapping capabilities across the sales toolset.
Integrations: how meetings, calls, quotes, contracts and payments flow between tools and into the CRM.
People: ownership, lead assignment, CRM adoption, skills, access and handovers.
AI: where AI supports research, calls, emails, scoring and forecasting, and how outputs are reviewed.
Automation: the triggers, conditions and actions behind follow-up, reminders, proposals and handovers, and the manual work still waiting to be automated.
For an SME, the result should be practical: understand what you already have, find what is missing, remove unnecessary duplication and prioritise the improvements that will win more business. It is also the foundation for effective sales automation, because you need a clear process before you add more workflows.
This guide is part of our tech stack audit series. For the framework across every department, start with our complete guide to running a tech stack audit; for the earlier stages of the customer journey, see the marketing tech stack audit.
Step 1: Define sales outcomes, scope and baseline
Start with a business problem rather than a preferred tool. "Improve our sales technology" is hard to evaluate; "make sure every qualified lead is contacted within an hour and every open deal has a next action" gives the audit a specific, testable outcome.
Choose your objectives
Useful sales objectives include:
Reducing the time between enquiry and first meaningful sales contact.
Making sure every open opportunity has an owner and a scheduled next action.
Improving conversion between pipeline stages and overall win rate.
Shortening the time from qualification to proposal, and from proposal to signature.
Making the forecast more accurate and less dependent on individual judgement.
Removing manual data entry, so salespeople spend more time selling.
Speeding up the handover from a won deal to onboarding and invoicing.
Cutting overlapping sales tools and unused subscriptions.
Then choose the scope. You might review the whole sales function, or start with a single journey such as inbound enquiry to booked meeting, or proposal to signed contract and invoice.
Decide whether the review covers inbound sales, outbound prospecting, account management or several of these. Each sales motion may need its own process, stages and measures.
Establish a baseline
Record the current position before making changes, so every improvement can be proven:
Time to first meaningful sales contact: reveals routing and response delays.
Share of open deals without a next action: reveals follow-up gaps.
Stage-to-stage conversion and win rate: shows where deals stall or are lost.
Average sales cycle length: from qualification to won or lost.
Proposal turnaround time: from qualification to proposal sent.
Forecast accuracy: forecast versus actual revenue for recent periods.
CRM data completeness: the share of deals with required fields, values and close dates.
Time from signature to onboarding and invoice: reveals handover delays.
Hours spent on sales admin and reporting: the cost of manual work.
Sales software, AI and integration costs: the commercial baseline.
Define the period, records and exclusions behind each measure, and label estimates. If you cannot measure something on this list today, that gap is a finding in its own right.
Be precise about each calculation. Win rate as won deals divided by won plus lost deals answers a different question from customers divided by all incoming leads. Compare like-for-like services, lead sources and sales cycles, and remember that recent opportunities may still be open rather than lost.
Check: can you state each baseline figure and how it was calculated? Expected outcome: two or three agreed objectives, a defined scope and a baseline to measure against.
Step 2: Inventory your sales apps and functionality
Organise applications by sales function rather than by supplier, because grouping by function exposes gaps and overlap far more clearly. Include platforms owned by other departments if sales depends on them, such as the marketing platform, accounting system or project management tool.
Cover the full sales stack
Work through each function in turn, so nothing is missed:
CRM and pipeline management: contacts, companies, deals, stages, tasks and activity history.
Calendar scheduling and meetings: booking pages, round-robin assignment, reminders and video conferencing.
Email and communication: one-to-one email, shared inboxes, SMS and messaging apps, and whether conversations are logged to the CRM.
Calling and call tracking: VoIP, dialers, call recording and call attribution.
Sales sequences and outreach: multi-step follow-up across email, phone and social.
Sales intelligence and enrichment: company and contact data, buying signals and LinkedIn tools.
Proposals, quotes and CPQ: proposal templates, pricing rules, approvals and product catalogues.
E-signature and contracts: signing, contract storage and renewal dates.
Sales enablement: pitch decks, case studies, battlecards and demo videos, and how their use is tracked.
Live chat and conversational sales: website chat, chatbots and messaging channels.
Payments: deposits, payment links and subscriptions taken at the point of sale.
Reporting and forecasting: pipeline dashboards, forecasts and commission data.
AI and integration: AI note-takers, AI assistants and workflow platforms.
Build one master register
Record each application once and associate it with every function it serves. For each, document the owner, active users, plan, cost, renewal date, usage level (Basic, Moderate or Advanced), data held, live workflows, limitations and dependencies. Include spreadsheets used to track deals or commissions, free tools and subscriptions bought on personal cards.
Separate functionality from usage
Classify each relevant capability as:
Used: supporting a live sales process.
Available but unused: included in your plan but not configured or adopted.
Upgrade required: available at extra cost.
Unavailable or unsuitable: unable to meet the requirement.
The "available but unused" list is where the quickest wins hide: many teams pay separately for booking, sequences, quotes or e-signature that their CRM already includes. Check the detail before replacing a specialist tool, though, because pricing logic, approvals, reporting and limits all matter.
Treat low usage as a reason to investigate, not proof that a tool is unnecessary. A specialist quoting feature may be used infrequently but still be essential for complex deals.
Check: compare the register against card statements and the accounting system. Expected outcome: one complete register, with every tool's owner, cost, usage and unused features recorded.
Step 3: Define your sales systems and data ownership
Sales data often ends up split between the CRM, inboxes, calendars, proposal tools, spreadsheets and the accounting platform. This step decides which system is the source of truth for each type of sales data, and how records relate to one another.
Decide which system controls each type of data
A workable arrangement for most SMEs looks like this:
Contacts, companies and deals: the CRM, with every other sales tool reading from or writing to it.
Pipeline stages and deal values: the CRM, using agreed stage definitions.
Activity history: emails, calls and meetings logged automatically to the CRM record.
Products, services and price lists: one maintained catalogue, used by both quotes and invoices.
Quotes and proposals: the proposal or CPQ tool, with status and value written back to the deal.
Signed contracts: the e-signature or document system, with signed status and a link on the deal.
Invoices and payments: the accounting platform, with payment status shared back to the CRM. See our finance automation services.
Set record and relationship rules
Define how contacts relate to companies and deals. An email address helps identify a person, but one company may have several contacts and several deals, and one contact may be involved in more than one opportunity. Store each system's record ID against the others, so a quote, contract and invoice can always be traced back to the right deal.
Preserve the original lead source on the deal, so sales outcomes can be reported back to marketing.
Define how duplicates, merged records and conflicting updates are handled, and preserve useful history when records are corrected. A CRM automation and integration review can help align the record structure with how you actually sell.
Check the information needed to manage a deal
For each active opportunity, the team should be able to see:
The company and the relevant contacts, including decision-makers.
The enquiry source and the service or product of interest.
The sales owner and current stage.
The requirement, scope and qualification outcome.
The proposed value, currency and what the value represents.
The expected decision date and the evidence behind it.
The next action, its due date and who is responsible.
The current proposal, agreement and communication history.
The reason for loss or deferral, where applicable.
Spot the warning signs
These symptoms are worth investigating, as they often point to unclear data ownership:
Deals are tracked in a spreadsheet alongside the CRM.
Prices in quotes differ from prices on invoices.
Customer details are re-typed between the CRM, proposal tool and accounting platform.
Sales and finance report different revenue figures for the same month.
Check: pick ten recently won deals and trace each through quote, contract and invoice. Expected outcome: one documented source of truth for each type of sales data, with record IDs linking every system.
Step 4: Map your sales processes
Follow real deals through the stack. Interview the salespeople, managers and support staff involved, and compare how they actually work with the documented process. Map the main sales journey first, then the management and handover processes around it.
Use the six-question framework
For every step, answer six questions:
How: the steps involved, and whether each is manual or automated.
What: the data required, such as budget, timescale, decision-maker, service and deal value.
Why: the purpose of the step and the benefit of automating it.
Who: the salesperson, manager, customer or system responsible.
When: the event or timing that starts the step.
Where: the application in which the work happens.
Map the sales journey end to end
Ask these questions at each stage of a recent won deal and a recent lost one:
Lead received: an enquiry arrives from marketing, a referral or outbound activity. Is it assigned to the right person straight away, with its source and context?
First contact: the salesperson responds. How long did it take, and was the contact logged?
Qualification: the lead is assessed for fit, need, budget and timescale. Are the criteria and the outcome recorded consistently?
Discovery meeting: a meeting is booked and held. Are the booking, notes and next steps linked to the deal?
Proposal: a quote or proposal is prepared and sent. Is it built from approved templates and prices, and is its status tracked?
Negotiation and approval: terms and discounts are agreed. Are approvals recorded, and does a revised proposal replace the old one?
Signature and close: the contract is signed. Does the deal update automatically, with the signed document attached?
Handover: delivery and finance take over. Do they receive the scope, contacts, files and billing terms without re-keying?
Where deals were lost, check whether the reason was recorded and whether the lead was returned to a nurture sequence. Our guide to lead management automation covers the early stages in depth.
Include a stalled opportunity as well as a won and a lost deal, and map inbound enquiries, outbound prospecting and proposals separately where their rules differ. Capture the conversations, spreadsheets, approvals and checks that happen outside the CRM.
Define a complete handover
Follow each won deal far enough to see whether delivery and finance receive what they need: the signed scope, the approved proposal version, the billing schedule, customer contacts, deadlines, exclusions and any commitments made during the sale.
Assign someone to accept the handover, because a notification alone does not prove delivery has enough information to start. Keep commercial acceptance separate from operational readiness: a deal can be won while onboarding is still blocked by missing information or an unpaid deposit.
Map pipeline management and renewals
Document how managers review the pipeline, how the forecast is produced, and how renewals, upsells and repeat business are identified. Ask where the forecast actually lives, who updates it and how often it is overridden by hand.
Mark every step where someone copies data between systems, chases an approval by email or rebuilds a report manually; these become the shortlist for automation and AI in Steps 9 and 10.
Check: walk one won deal and one lost deal through the map with the people involved. Expected outcome: agreed process maps with every manual step and handover flagged.
Step 5: Audit your pipeline and CRM data quality
A CRM is only as useful as the data in it. If deals lack values, close dates or next steps, the pipeline view, forecast and every automation built on it become unreliable. This step tests the quality of the sales data itself.
Review your pipeline stages
For each stage, check that it has a clear definition, entry and exit criteria, and the fields required to move forward. Stages that describe activities ("call booked") rather than buyer progress ("needs confirmed") often make forecasts less reliable. Agree how long a deal can sit in each stage before it is flagged.
Each stage should describe meaningful buyer progress, with evidence that its conditions have been met. Adapt these examples to your sales model:
New enquiry: a valid enquiry has been recorded and assigned.
Qualification: suitability and the basic requirement have been assessed.
Discovery: the conversation has taken place and findings are recorded.
Proposal: an approved proposal has been sent.
Decision: the customer is evaluating an offer and the next decision step is known.
Won: your agreed commercial acceptance criteria are met.
Lost or deferred: the outcome and reason are recorded, with any suitable future action.
Avoid moving a deal forward simply because an email was opened or a reminder was sent, and keep task completion separate from buyer progress: a salesperson can complete several activities without the customer moving any closer to a decision.
Run a data quality check
Review a sample of open and recently closed deals and look for:
Deals without an owner, value, expected close date or next action.
Stale deals with no activity for an agreed period.
Close dates repeatedly pushed back without explanation.
Duplicate contacts, companies or deals.
Won or lost deals without a recorded reason.
Calls, emails and meetings that were never logged to the deal.
The last point is often the biggest: if activity lives in personal inboxes and calendars, managers cannot see what is really happening, and anyone covering a colleague's deals starts from nothing.
Decide what should be required
Make the essential fields mandatory at the right stage, not all at once at the start, so salespeople are not forced to invent data. Where possible, capture information automatically, from forms, enrichment tools, booking pages and call notes, rather than asking people to type it.
Check: export open deals and count those missing an owner, value, close date or next action. Expected outcome: agreed stage definitions, required fields by stage and a clean, current pipeline.
Step 6: Map integrations, triggers and actions
Document each distinct workflow rather than simply noting that two applications are connected. A calendar-to-CRM connection might create meetings, update deal stages, send reminders and notify the salesperson, and each of those operations needs clear rules.
Test lead routing and the marketing-to-sales handover
An enquiry arriving in the CRM does not mean it has reached the right person or received attention. Submit controlled test enquiries through each important route and follow them to the expected outcome. Check that service interest, source information and communication preferences arrive where available, and that assignment rules handle existing customers, territories, specialist services and unavailable salespeople. Define what happens when no rule matches.
Then agree the handover with marketing:
What makes an enquiry ready for sales review, and who accepts it?
What response target applies?
How are unsuitable or duplicate enquiries classified?
What happens to valid prospects who are not ready to proceed?
Which outcomes are returned to marketing?
Connecting these responsibilities with marketing automation preserves context across the customer journey.
Describe each workflow in the same way
For every sales workflow, record its trigger, the checks it performs and the actions that follow. Typical examples:
New enquiry: triggered by a form, chat or call. Checks for an existing contact or open deal and applies assignment rules. Creates or updates the deal, assigns an owner, sends an acknowledgement and creates a response task.
Meeting booked: triggered by a booking page. Matches the contact and deal, then logs the meeting, moves the deal to the right stage and schedules reminders.
Proposal sent and signed: triggered by status changes in the proposal or e-signature tool. Matches the deal using a stored ID, then updates the stage, value and documents, and stops proposal chasers once signed.
Deal won: triggered by a verified signature or deposit. Confirms the agreed scope and billing terms, then creates the onboarding project, notifies delivery and prepares the invoice in the accounting platform.
Call logged: triggered when a call ends. Matches the phone number to the contact, then records the call, recording and outcome against the deal.
For each, record the source, destination, fields transferred, integration method, timing, owner and date of the last successful test.
Test each core connection against what it should achieve:
Email to CRM: messages attach to the correct contacts and deals.
Calendar to CRM: bookings, changes and cancellations update correctly.
Calling platform to CRM: calls are logged against the intended record.
Proposal platform to CRM: the proposal version and status match the correct deal.
CRM to accounting: customer and invoice references map correctly.
CRM to project management: approved scope and delivery requirements transfer accurately.
Choose the right build method
A native integration may be enough. More complex flows may need a Zapier, Make or n8n workflow, a marketplace app or a custom connection. Compare the exact triggers, fields, costs and limits rather than choosing from application names alone. Our Zapier and HubSpot automation guide shows common patterns.
Test reliability and stop conditions
Test incomplete information, repeated events, revoked access and partial failures, such as a project created but the invoice step failed. Then check what happens when a prospect books, replies, declines or signs mid-sequence: each workflow should respond to their current situation. Look for competing automations too, where two systems both send the same reminder or create the same task.
Check: submit a test record for each workflow and confirm the outcome in the destination system. Expected outcome: every workflow has a named owner, a recent successful test and an alert when it fails.
Step 7: Review sales reporting, forecasting and feedback to marketing
The audit should establish which sales outcomes you can measure reliably and where uncertainty remains. Activity, pipeline, forecast and revenue each answer a different question, so keep them separate.
Check your core sales reports
Confirm you can produce each of these from the CRM without manual spreadsheets:
Pipeline value by stage, owner and expected close month.
Stage-to-stage conversion, win rate and average sales cycle.
Response time to new leads, and deals without a next action.
Win and loss reasons, by source, service and salesperson.
Activity levels alongside outcomes, so effort is not mistaken for progress.
Test the forecast
Compare recent forecasts with actual results. Check whether deal values and close dates are realistic, whether stage probabilities reflect real conversion rates, and how often the forecast is adjusted by hand outside the CRM. A forecast built on the clean data from Step 5 is far more reliable than one rebuilt in a spreadsheet each month.
Weighted pipeline is usually calculated by multiplying each deal's value by its stage probability and adding the results, so it is only as useful as those probabilities. Compare them with actual conversion history by stage and, where data allows, by service, lead source and deal type, and avoid treating a salesperson's confidence score as a calibrated probability. Keep dated forecast snapshots, so you can compare what was expected at the time with what closed, including timing slips and lost deals.
Flag deals that distort the forecast:
Close dates repeatedly pushed back.
No next action or recent meaningful activity.
Missing decision-makers or unresolved qualification.
Proposals that have expired.
Values unsupported by the current offer.
Activity counts can support coaching, but more calls or emails do not automatically mean stronger opportunities.
Agree revenue definitions
State whether "sales" means signed contract value, invoiced revenue or cash collected, as these are different measures. The accounting platform should own invoiced and collected figures, with the CRM showing contracted value and payment status.
Where relevant, also separate one-off contract value from recurring annual value and expected margin. Connect these outcomes through finance automation while keeping their definitions distinct.
Feed outcomes back to marketing
Record the original lead source, qualification outcome and won value on every deal, so marketing can see which campaigns produce customers, not just enquiries. Where supported and appropriate, these outcomes can also be sent to ad platforms to improve bidding; our sister agency One PPC can audit that set-up in detail. A single Looker Studio dashboard built on agreed definitions removes most monthly disagreements between sales, marketing and finance, and is a core part of revenue operations.
Our marketing tech stack audit covers the marketing side of this loop, including lead capture, conversion tracking and advertising feedback.
Check: compare last quarter's forecast with actual signed and invoiced revenue. Expected outcome: core reports produced from the CRM, agreed revenue definitions and a forecast you can explain.
Step 8: Review people, ownership and governance
A sales stack needs clear responsibility as well as technical connections. Sales is particularly exposed when salespeople leave, because their deals, contacts and conversations may sit in personal inboxes, calendars and phones rather than the CRM.
Clarify who is responsible for what
Name an owner for every major sales process and every integration, then agree, in writing:
Lead assignment: how new leads are allocated by territory, service, size or round-robin, and what happens when someone is away.
Response standards: how quickly new leads must be contacted, and who covers if the target is missed.
Marketing-to-sales handover: what counts as a qualified lead, and how rejected leads are returned with a reason.
Pricing and discount approval: who can approve non-standard pricing or terms.
Sales-to-delivery handover: what information delivery and finance need before work starts.
CRM administration: who maintains stages, fields, users and automations, and authorises changes.
Software purchasing: who checks existing capabilities before a new sales tool is approved.
Check CRM adoption and access
Compare who has a CRM licence with who actually logs activity. Low adoption usually signals a training need, a poor fit or too many required fields, so interview users before deciding which applies. Remove former staff promptly, reassign their open deals, and make sure contacts and conversations belong to the business rather than to individuals.
Watch salespeople complete common tasks and ask what they keep outside the CRM, and why. Workarounds may reveal missing functionality, too many fields, poor mobile usability or unclear expectations; training helps only when the underlying process is right. Check that shared logins and agency access do not hide who did what. When a salesperson leaves, explicitly reassign their open deals, tasks, mailbox connections and any automations that run under their account.
If calls are recorded, confirm prospects are told and that recordings are stored and retained in line with UK GDPR.
Review skills and coaching
Note where the team needs support, such as using sequences, logging calls, building proposals or working with AI assistants. Recorded calls and AI summaries can also support coaching, provided this is handled transparently.
Check: compare CRM licences with active users and list every admin. Expected outcome: written assignment and handover rules, current access only, and a named owner for every process and integration.
Step 9: Audit your sales automation
Sales automation removes admin, speeds up response and makes follow-up consistent. Many teams already run workflows across their CRM, booking tool, proposal software and Zapier or Make, built over several years by different people.
Inventory every sales automation
List every workflow, including lead assignment, follow-up tasks, sequences, meeting reminders, proposal chasers, stage updates and won-deal handovers. For each, record the owner, trigger, conditions, actions, monthly volume, cost in tasks or operations, last error and whether it is still needed.
Check how each workflow behaves
Each automation should have:
A clear trigger: such as a new enquiry, a stage change or a signed proposal.
A search before acting: so it updates existing deals rather than creating duplicates.
Sensible conditions: respecting consent and excluding existing customers or open deals where appropriate.
Stop conditions: so sequences stop when a prospect replies, books, declines or signs.
Monitoring: so someone is alerted when it fails.
Overlap is common: a prospect may receive a booking confirmation from the calendar tool, the CRM and a Zap. Assign one owner to every outcome and switch off the rest.
Inspect pricing and proposal controls
Check who can change prices, approve discounts and issue non-standard terms. Verify that proposals draw on the current service descriptions and approved price list, that a revised offer supersedes the old one, and that signing an outdated version cannot start the wrong workflow. For complex quotes, review dependencies between products, quantities, discounts and approvals: automating an incorrect pricing rule spreads the error faster.
If a later step fails, retrying should not create a second project or invoice. Record how completed actions are recognised and who handles unresolved exceptions.
The sales automations worth having
Once existing workflows are reliable, these typically deliver the most value for SMEs:
Speed-to-lead: instant acknowledgement, a booking link, assignment and an alert to the salesperson.
Follow-up tasks and reminders: every open deal gets a next action, with an alert to the manager when it becomes overdue.
Sequences: approved multi-step follow-up that stops automatically on engagement.
Meeting reminders and no-show follow-up: reducing missed calls and recovering the ones that slip.
Proposal generation and chasers: drafts built from deal data, with reminders until signed or declined.
Stale deal alerts: flagging deals with no activity or a passed close date.
Won-deal handover: creating the onboarding project and invoice, and notifying delivery and finance.
Renewal and upsell prompts: reminders ahead of contract end dates.
For a quick capacity estimate, multiply monthly task volume by minutes saved per task and divide by 60. Saving five minutes on 240 deal-admin tasks a month releases 20 hours, before monitoring and exceptions; this is an illustration, not an automatic payroll saving.
Our guide on how to boost sales with automation has more examples, and our sales automation services cover the build.
Check: list every active sales automation with its owner, last run and error rate. Expected outcome: no orphaned or duplicate workflows, alerts on every critical automation, and a scored shortlist of the next workflows to build.
Step 10: Audit AI across sales
AI can remove a large share of sales admin and help salespeople prepare better, but only when it works from reliable CRM data and its outputs are checked. An AI audit establishes what is in use, whether it is safe, where it overlaps and where it could add more value.
Inventory every AI tool and feature
Sales AI usually sits in four places:
AI assistants: ChatGPT, Claude, Gemini or Copilot seats used for research, emails and proposals.
AI note-takers and call analysis: tools that record, transcribe and summarise sales calls.
AI built into the CRM: lead scoring, email drafting, deal insights and forecasting features, often switched on by default.
AI inside automations and agents: LLM steps in Zapier, Make or n8n workflows, chatbots and AI agents that qualify leads or book meetings.
Record what each use case receives, produces and is allowed to change. A tool that drafts a follow-up email has very different requirements from an agent that can send it or update a deal.
Assess each use case
Test representative examples, including difficult ones, and assess:
Call summaries: correct matching to the contact and deal, omissions, and invented commitments or prices.
Email and follow-up drafting: accuracy, tone and whether it reflects what was actually agreed.
Lead scoring and prioritisation: whether high scores really convert, and how ambiguous cases are handled.
Prospect research and enrichment: data accuracy, sources and compliance with UK GDPR.
Proposal drafting: correct scope and pricing from approved templates, never invented terms.
Pipeline and forecast insights: whether the evidence supports the conclusions.
Measure correction time as well as initial speed, because an AI step that needs heavy editing can add work.
Distinguish suggestions from actions. AI might suggest a qualification outcome or next task, while a salesperson confirms the decision. Define which records each assistant may read or update and which actions need approval, and remember that a polished summary is not evidence that its details are correct.
Set governance and review rules
Keep a person in the loop for anything sent to a prospect or committed on the company's behalf, especially prices, discounts and contract terms. Our guide to human-in-the-loop AI systems explains how to design this review. Confirm that call participants know when AI recording or transcription is in use, check which customer data is entered into AI tools, and agree a short acceptable-use policy. For AI connected to emails or documents, check that instructions hidden in those materials cannot override the workflow's permissions.
Get more from AI
The strongest sales uses combine AI with your CRM data and a review step: summarising every call and logging next steps automatically; drafting follow-ups and proposals from deal notes; prioritising the day's calls; and turning unstructured emails into structured deal updates by combining natural language with structured data. AI agents connected through Make, Zapier, n8n and the Model Context Protocol (MCP) will take on more of this work, relying on the same foundations as this audit. Our AI consultants help sales teams build them.
Check: list every AI tool and who reviews its output. Expected outcome: one approved set of AI tools, a written usage policy, and one or two measured AI use cases ready to pilot.
Step 11: Consolidate your sales stack
With every layer mapped, overlap becomes visible. Sales stacks often have strong consolidation potential, because modern CRMs now include capabilities that were once bought separately. Functional overlap does not automatically mean redundancy, though: a specialist tool may meet a critical requirement that the CRM handles poorly.
Common sales overlap hotspots
Check each of these against what your CRM already includes:
Scheduling: Calendly or Acuity alongside CRM booking pages. Do routing and round-robin needs justify both?
Sequences: a separate outreach tool alongside CRM sequences. Are the extra features actually used?
Proposals and quotes: PandaDoc or Proposify alongside CRM quotes. Do pricing rules, approvals and design needs justify both?
E-signature: DocuSign alongside CRM document signing.
Calling: a separate dialler alongside CRM calling.
Enrichment: several data providers bought by different people.
AI note-takers: several recording tools alongside the CRM's own call AI.
Pipeline spreadsheets: deals or forecasts tracked outside the CRM.
Decide what to keep, migrate or retire
Assess each overlapping tool on how far its features overlap with your CRM and how critical it is to winning business:
Low criticality, high overlap: retire now. Switch to the native feature and cancel at renewal.
High criticality, high overlap: plan a phased migration. Rebuild templates, sequences and automations in the CRM, then run both briefly in parallel.
High criticality, low overlap: keep and integrate. A specialist tool, such as a CPQ engine with complex pricing, that the CRM cannot match.
Low criticality, low overlap: review usage. Retire it if adoption stays Basic by the next renewal.
Check migration dependencies first
Before removing any tool, check active sequences, open proposals, templates, signed contract archives, call recordings and connected workflows. Verify feature depth, plan requirements and current commercial terms, and confirm you can export your history. A promised feature or future bundle is not a tested replacement. Our platform consulting services help with both consolidation and CRM selection.
Check: list every capability provided by more than one tool. Expected outcome: a keep, migrate or retire decision for each, timed to renewal dates.
Step 12: Cost, payback and a prioritised roadmap
The audit is only valuable if it produces decisions. Translate every finding into an action with an owner, dependencies and acceptance criteria, then prioritise the problems with the greatest business effect. Leads waiting days for a response matter more than simplifying a low-cost subscription.
Compare the full cost
Include CRM seats, plan upgrades, sales tools, AI subscriptions, Zapier tasks or Make operations, and the hours your team spends on data entry, proposals and reporting. For any change, add migration, rebuilding templates and sequences, training, maintenance and any period of running two tools in parallel. Align cancellations with renewal dates to avoid paying twice.
As a hypothetical example: retiring a separate scheduling tool and e-signature subscription saves £180 per month, while a CRM plan upgrade adds £60 per month, giving a net saving of £120 per month. With £1,200 of migration work, the simple payback is 10 months. Report time savings separately, and say how released capacity will be used, such as more sales calls, faster proposals or quicker onboarding.
Sequence the roadmap
Work through improvements in this order, adjusting for your own priorities:
Immediate: fix lead response and assignment. Make sure every new lead is assigned and contacted within the agreed time. Evidence: test enquiries are assigned, acknowledged and alerted correctly.
Immediate: clean the pipeline. Give every open deal an owner, value, close date and next action, and close out dead deals. Evidence: the data quality check from Step 5 passes.
Near term: agree stages, handovers and access. Evidence: written stage definitions, assignment rules and handover requirements, with current users only.
Near term: automate follow-up and handovers. Evidence: no open deal without a next action, and won deals reach delivery and finance without re-keying.
Planned: fix reporting and forecasting. Evidence: core reports produced from the CRM and a forecast that reconciles with actual results.
Pilot: introduce a governed AI use case, such as call summaries. Evidence: representative tests show acceptable accuracy and review effort.
Planned: consolidate overlapping tools. Evidence: required features, records and integrations pass migration checks before anything is cancelled.
Give each change a rollback plan, and check dependencies before retiring any tool, because it may power a booking page, template or report that is easy to overlook.
What your completed sales audit should contain
A useful audit leaves the team with documentation that supports day-to-day decisions and implementation:
Sales application register: every tool with its owner, usage, cost, renewal date and dependencies.
Capability matrix: features used, available but unused, upgrade required or unsuitable, with overlaps highlighted.
Data ownership rules: the source of truth for contacts, deals, prices, quotes, contracts and invoices, plus record-matching rules.
Process maps: the sales journey from lead to handover, plus pipeline management and renewals.
Pipeline and data quality review: stage definitions, required fields and the findings of the data check.
Integration register: every workflow with its triggers, checks, actions, test status and recovery.
Reporting and forecast review: core reports, revenue definitions and forecast accuracy.
People and access review: owners, assignment and handover rules, CRM adoption and skills gaps.
Automation and AI registers: every workflow and AI use, with health, review responsibilities and next opportunities.
Consolidation plan and roadmap: keep, migrate or retire decisions, with cost, payback, owners and acceptance criteria.
Each recommendation should explain the issue, its business effect, the proposed change and how success will be checked.
Keep your sales stack under review
Update the register whenever you introduce a tool, change a pipeline stage, launch a significant workflow or a salesperson joins or leaves. Reassess the wider stack before major renewals, and check every new tool or AI subscription against existing capabilities before it is bought. Use the original baseline to confirm that improvements have sped up response, strengthened follow-up and made the forecast more dependable.
Book a sales tech stack audit with OpScaling
OpScaling runs Automation and AI Audits across every business function. For sales, we map your systems, processes, apps, integrations, people, automation and AI, then design and build the improvements in HubSpot and HighLevel, Make, Zapier and n8n.
Explore our sales automation services and customer success automation, see the results in our case studies, or book your free consultation to discuss your sales systems.
Continue the series with the marketing tech stack audit, or return to the complete tech stack audit guide for the framework across every department.
