Skip to Content

AI Readiness

AI for Sales and CRM in Odoo

Sales is where AI has the biggest upside and the loosest arithmetic. Time saved is certain and countable. Deals won because you quoted first are real, but they are an argument, and we do not build a business case on an argument.

Three workflows carry most of the value: turning enquiries into quotes, working out which leads deserve today, and making sure nothing goes cold because somebody forgot. The first is the flagship, and it is the one most teams feel immediately.

Book a call What AI readiness means

The three workflows we build

Scoped, quantified and approved before anything is built, so you know what each one is worth to you before you commit.

1. Quotes drafted from enquiry emails

An enquiry arrives. Someone reads it, works out what the customer actually means, finds the products, checks the price list and builds a quotation by hand. Then does it again a hundred times a month. This is the flagship, because it is both the most repetitive and the one where speed changes the outcome.

  • Verbs: Extract, then Draft
  • Where a person confirms: The quote is drafted; your salesperson checks pricing and terms and sends it. Never auto-sent, at any point.
  • What it needs from your data: A clean product catalogue and current price lists. This is the one workflow where dirty product data genuinely blocks you, and we check it first.

2. Lead enrichment and scoring

Manually researching inbound leads and guessing which ones deserve today. Most teams do this on instinct, which works right up until the volume gets away from them.

  • Verbs: Extract, Classify, then Suggest
  • Where a person confirms: The model enriches and ranks; your salesperson decides who to chase. The score has to be explainable — if it cannot say why, it does not ship.
  • What it needs from your data: Enough historical win and loss data to have something to learn from, and lead records that capture where they came from.

3. Quote follow-up and pipeline nudges

Remembering to chase an unviewed quote, updating the CRM after a call, noticing a deal has gone quiet. The admin nobody enjoys and everybody skips when they are busy, which is exactly when it matters.

  • Verbs: Summarise, then Suggest
  • Where a person confirms: Follow-ups are drafted and timed; the salesperson approves before anything goes out.
  • What it needs from your data: Quotations and activities actually recorded in Odoo. If the pipeline lives in somebody’s head or a spreadsheet, fix that first.

Worth knowing: Odoo already tracks whether a quote has been viewed and can notify on it. We check what the platform does natively before building anything, because native is cheaper, safer and survives upgrades. Sometimes half of the third workflow is already there, switched off.

The honest bit about sales value

Sales is the area where AI vendors quote the biggest numbers, and it is the area where those numbers deserve the most scepticism. We will show you both halves and we will tell you which one to trust.

What is countable

The time. Quotes built per month, times minutes per quote, times your loaded hourly cost. Lead triage time. CRM admin per rep per week. These are hours your team is spending now, and you can verify every input from your own records.

What is arguable

The extra deals. Quoting first genuinely wins business — the research is not controversial and every salesperson has felt it. But when we tell you that responding in an hour instead of two days wins three more deals a month, a reasonable finance director will say "we would have won some of those anyway", and they are right to. We cannot separate the two cleanly, so we do not price on it.

So the number we ask you to make a decision on is the time saved. The extra deals are shown, argued for, and then deliberately left out of the payback calculation. If the program only works once you count them, it is not ready.

The same discipline runs through the accounting program, where almost all of the value happens to be the countable kind. That is why we usually suggest starting there — not because sales matters less, but because the first build should be the one that is easiest to prove.

How we test it before you trust it

A quote drafter that is right most of the time is not good enough, because the failures are the expensive part. So we test the edges deliberately, not just the happy path:

  • A standard enquiry. Right products, right prices, sensible layout — and it stops for review rather than sending.
  • An enquiry mentioning something you do not sell. It must flag the unknown line for a person, not quietly guess at the nearest match.
  • An ambiguous, multi-product enquiry. Best attempt, plus a clear flag that it was ambiguous.
  • A lead from an unfamiliar industry. The score has to be explainable and the enrichment accurate, not invented.
  • An enquiry for the wrong territory. Routed correctly or flagged. A silent misroute is worse than no routing at all.

Your sales manager signs off each scenario. Done is not that it runs — it is that the team has stopped double-checking every output, which usually takes a few weeks after go-live and is the thing we watch for.

Common questions

Will it email customers without us seeing?

No. Quotes are drafted, follow-ups are drafted, and a person sends. Anything customer-facing keeps a human on the send button, and that is not a setting we quietly relax later.

Our product data is a mess. Are we out?

For quote drafting specifically, product and price list data is the constraint, and we will tell you honestly if it is not there. But that is one workflow of three — lead scoring and pipeline nudges do not depend on it, so there is usually somewhere sensible to start while the catalogue gets sorted.

How is a lead score different from what we already do?

Mostly in consistency, not brilliance. A good salesperson prioritises well on a good day. The score does it the same way every day, at 8am, across every lead, and it shows its reasoning so you can disagree with it.

Will our salespeople actually use it?

They use it when it removes the work they hate and leaves the part they are good at. Drafting the quote is the boring half; the conversation is the half they want. Design it the other way round and it gets quietly ignored — we have watched that happen.

Can we start with just the quoting?

Yes, and most teams should. One workflow, measured before and after. The second is a much easier decision once the first has a number attached to it.

Start with the quotes

It is the most repetitive job in most sales teams and the one where speed changes the outcome. A short call is enough to work out what it is costing you now.

Book a call How AI readiness works