Why reps can't find the case study they need mid-deal

The naming convention was never the fix

Someone has already tried to solve this at your company. Probably marketing. There's a shared drive with folders by sector, a spreadsheet index of every case study with columns for industry, deal size and outcome, and a rule that every won proposal gets filed as Client-Sector-Year. For about six weeks it works beautifully.

Then a rep saves a deck to their desktop because they're on a train. A proposal gets written in Google Docs and never leaves the deal folder. The index stops being updated when the person who owned it moves to a different team. Two years later you have a sales knowledge base that is, technically, complete — and functionally unusable, because finding anything in it requires already knowing it exists.

You've got good reps. You've got genuinely good past work, written up properly, with real numbers in it. The failure isn't discipline. It's that filing and finding are two different problems, and everyone keeps trying to solve the second one by doing more of the first.

Your rep isn't searching for a word. They're searching for a situation.

Here's what actually happens on the call. A prospect — a mid-sized logistics firm, three depots, running a fleet management system nobody likes — asks the question every buyer asks: "Have you done this for anyone like us?"

Your rep knows the answer is yes. There was a project, eighteen months ago, a distribution business, similar mess, and it went well. They just can't remember the client's name.

So they search. "Logistics". Nothing useful, because the case study was written up as "third-party distribution". "Fleet". Two hits, both irrelevant. "Depot". Nothing. Keyword search matches the words in the document, not the situation the rep is describing, and the person who wrote that case study eighteen months ago was not writing it to be found by this prospect's vocabulary. They were writing it to sound good.

What the rep needs is to describe the prospect in their own words and get the closest past work back. Meaning matched against meaning, with the keyword matching still doing its job underneath for names, product codes and numbers (in technical speak, hybrid search). That's a different mechanism from a folder tree, and no amount of tidying gets you there.

It's a findability problem, and it has a category name now

The thing that fixes this is an AI knowledge base for sales: your case studies, won proposals, statements of work and project write-ups indexed together, with a chatbot over the top that your reps type into like a colleague. "Prospect is a 200-person logistics firm, three depots, worried about implementation downtime. Have we done anything close?"

The answer comes back in seconds, with citations — the actual document, the actual page. That last part matters more than the speed. An answer a rep can't verify is an answer they won't put in front of a buyer, and rightly so.

Two years ago this was a data science project. Someone had to build it, host it, and keep it alive. That's no longer true, which is the part most sales leaders haven't caught up on. You point it at Google Drive, Dropbox, Notion, or upload the PDFs directly, and it's answering questions the same afternoon.

The buyer is going to check your rep's answer anyway

There's a reasonable objection here: if buyers are all using AI to research vendors themselves, why does it matter what your rep can retrieve?

Gartner surveyed 645 B2B buyers and found 69% prefer to validate AI-generated insights with a sales rep [1]. Buyers are running the AI research and then coming to your rep to check whether it's true. The moment your rep can't produce the specific, relevant proof, they've failed the exact test the buyer set up for them.

The same research found reps outperforming generative AI by 39 percentage points on understanding buyer needs and 32 percentage points on buyer confidence in the decision [2]. The rep is the differentiator. They just need the filing cabinet to answer them.

“Sixty-nine percent of B2B buyers prefer to validate AI-generated insights with sales reps. Sales reps outperformed GenAI by 39 percentage points on understanding buyer needs, 32 percentage points on buyer confidence in decisions, 28 percentage points on advancing purchase process steps, and 21 percentage points on quantifying benefits.”

Gartner

Survey of 645 B2B buyers, press releases 20 May 2026

Where we'd push back on our own argument

Sometimes the honest answer is that you haven't done it before. A good assistant will tell you that — nothing relevant came back — and a rep who trusts the tool will believe it and stop looking. That's a genuine risk if half your best work was never written up, lives in a client's SharePoint, or is under an NDA that means you can't name them anyway. Garbage in, confident-sounding nothing out.

So before you index anything, spend an hour listing the ten deals you're proudest of and checking whether each one exists as a document you're allowed to show a prospect. If four of them don't, that's a documentation problem, and no tool fixes it. Write those four up first. You'd want them written up regardless.

The other honest caveat: time saved is not automatically time earned. Gartner's survey of 210 sales leaders found AI tools save sellers an average of 4.8 hours a week, and 72% of sales organisations report low reinvestment of that time into high-value selling [3]. Handing a rep back twenty minutes an afternoon doesn't do anything on its own.

“AI tools save sellers an average of 4.8 hours per week, yet 72% of sales organizations report low reinvestment of those time savings into high-value sales activities. Sales organizations that achieve moderate to large AI time savings and reinvest that time into high-impact sales activities are 2.2x more likely to exceed customer growth goals and 3.1x more likely to exceed lead-to-opportunity conversion goals.”

Gartner

Survey of 210 CSOs and senior sales leaders, press release 19 May 2026

What a single source of truth actually moved

Allianz Trade is far larger than you are, so take the mechanism rather than the numbers. Their stated problem was that they lacked "a single source of truth" for sales content, which caused version control issues and inconsistent messaging. After consolidating, they reported a 20% increase in quota attainment, 15 hours saved per week for reps, a 10% improvement in win rate and 98% adoption within the first few months [4].

The adoption figure is the interesting one. Content libraries fail at adoption because using them costs the rep time they don't have mid-deal. Something that answers in seconds, in the words the rep typed, costs them nothing.

LESSON ONE: Check whether your best past work exists as a shareable document before you buy anything to search it.

LESSON TWO: Measure retrieval by whether a rep uses it during a live call, not by how many documents you've uploaded.

LESSON THREE: Decide in advance what the reclaimed hours are for, or they'll quietly go into admin.

Start with the deals you're already losing on proof

Most AI adoption so far has been thin. The ONS found UK businesses with 10 or more employees now use an average of 1.6 AI technologies, up from around 1.4 since late 2023, with large language models the most common at 18% [5]. Plenty of your competitors have ChatGPT open in a tab and nothing indexed against their own past work.

Pick the next five deals where a prospect asks whether you've done it before. Time how long it takes your rep to produce a relevant, named example they're allowed to send. If the honest answer is "they send the generic deck", that's your case for a sales knowledge base, and it's an afternoon's work to test rather than a quarter's project.

The proof already exists. It's sitting in Drive with a filename nobody remembers.

Sources

[1] Gartner, 'Gartner Survey Finds 69% of B2B Buyers Turn to Sales Reps to Validate AI-Generated Insights', press release, 20 May 2026 — https://www.gartner.com/en/newsroom/press-releases/2026-05-20-gartner-survey-finds-sixty-nine-percent-of-b-two-b-buyers-turn-to-sales-reps-to-validate-ai-generated-insights

[2] Gartner, 'Gartner Survey Finds Sales Organizations That Provide AI-Enabled Next Best Actions Are 2.6x More Likely to Achieve Commercial Growth', press release, 20 May 2026 — https://www.gartner.com/en/newsroom/press-releases/2026-05-20-gartner-survey-finds-sales-organizations-that-provide-ai-enabled-next-best-actions-are-two-point-six-times-more-likely-to-achieve-commercial-growth

[3] Gartner, 'Gartner Survey Finds AI Saves Sellers Nearly 5 Hours Per Week, Yet 72% of Sales Organizations Fail to Reinvest Time in High-Value Activities', press release, 19 May 2026 — https://www.gartner.com/en/newsroom/press-releases/2026-05-19-gartner-survey-finds-ai-saves-sellers-nearly-five-hours-per-week-yet-seventy-two-percent-of-sales-organizations-fail-to-reinvest-time-in-high-value-activities

[4] Highspot, 'Allianz Trade Case Study: Increase Quota Attainment' (vendor-published customer story) — https://www.highspot.com/success-stories/allianz-trade/

[5] Office for National Statistics, 'Artificial intelligence in UK businesses: 2023 to 2026', 20 July 2026 — https://www.ons.gov.uk/businessindustryandtrade/business/businessservices/articles/artificialintelligenceinukbusinesses/2023to2026