Internal · Do not share

How to sell the co-design

The conversation guide that pairs with the product brochure and the design partner offer: the pitch, the demo order, the objections with counters that have worked in real calls, the ROI math for defending €50K, the design-partner qualification funnel, and the deal mechanics. Source: the Prudential architect call (June 2026) and the founding team's deal notes.

#01 The pitch in 30 seconds

Position: we are the intelligence layer that tells an enterprise where AI can actually be applied. Not an RPA vendor, not a consulting study, not a mining tool that needs a year of connectors.

Language discipline

• Say process discovery or process intelligence. Avoid "process mining" as a self-description — mining means reading system logs (Celonis territory), and it drags the conversation into a comparison we don't want. Use "mining" only to describe what they already have.
• Agents are the execution layer, mentioned after the intelligence story, never before it.

#02 The demo playbook

Demo link: process-mining-demo.pages.dev. Four screens, in this order. Show, pause, let them react — the Prudential call worked because we "did not frame much, just opened the prototype and noticed what was most exciting to him."

1 · ONTOLOGY

"Here is what we understand about your organisation"

The full map: processes, systems, KPIs, teams, all connected. Point at the plain-English query bar. This is the screen a technical executive has never seen before — give it air time. If they lean in here, the deal is alive.

2 · PROCESS CANVAS

"And here is one process, made visible"

Steps, decision points, the human review gate — and the evidence-backed confidence on every step. The message: we always know what we understand and what we don't. "Here's what we understand" was the exact moment the Prudential architect called it powerful.

3 · RUNBOOK

"This is what your engineers get"

The operating knowledge in plain language plus training and test cases. One line: "not a report — a package an AI agent learns from and is graded against." Keep this screen short unless the audience is technical.

4 · COCKPIT

"And this is how leadership decides what's next"

Value vs. effort, projected savings, payback, owners. End here — it's the screen that turns a technical demo into a business case, and it was rated the strongest part of the Prudential call.

#03 Objections and counters

The first two happened verbatim in the Prudential call — and the counters below are what actually turned them.

#04 Battlecards

The three names you will actually hear in deals. Never open by attacking them — acknowledge what they are good at, then place us on a different layer. Wedges are v1: validated at positioning level, sharpen them with what you hear live.

Log-based process mining

The process mining market leader. Reads event logs from SAP/ERP/CRM systems to reconstruct and analyse process flows. Many large enterprises already own it.

Where they're strongBrand and analyst standing, deep SAP integration, solid when the process lives inside well-instrumented systems that log every step.

Where we winThey only see what systems log — decisions, workarounds and cross-system work are invisible. Integration takes months before the first insight. Output is analysis; automation is still someone else's project. Not built to feed agents.

A line that works"Celonis tells you a claim sat open for 50 days. We tell you why — and hand your engineers the blueprint to fix it. If you already run Celonis, that data can feed our cockpit. Nothing is wasted."

Plant this question"How do they capture the process knowledge that never hits an event log? And how long did integration take before you saw the first insight?"

If the prospect is happy with Celonis, don't fight it — sell the discovery + agent layer on top. The complement story closes; the replacement story stalls.

Klarity Closest direct competitor

US platform, "Transforming Transformation." Captures processes from screen recordings, videos and AI interviews, builds a process map ("context graph"), and recommends improvements with ROI estimates. Marquee logos: DoorDash, Stripe, ServiceNow.

Where they're strongNearly identical capture motion (AI interviews + screen recording), fast SOP generation at scale, strong US enterprise logos, deep finance & accounting heritage.

Where we winTheir engagement ends in a map and recommendations; ours ends in a running agent. Runbooks carry training and test data into Beam's own agent platform — discovery and execution are one motion, no handoff. Plus European enterprise fit: private deployment, granular data controls, works-council-aware delivery.

A line that works"Mapping is where they stop. It's where we start — the first process is running on AI before the engagement ends, on the same platform that mapped it."

Plant this question"After their analysis, who builds and operates the automation — and on whose platform does it run?"

Their site also claims agent-building. Expect that counter; hold the line on native execution — our agent platform is the product, not a service bolted on. Validate this wedge in the first live deal against them.

Consultancies Accenture, Big 4, boutiques

People-powered process documentation and AI transformation programmes. Often already inside the account with board-level trust.

Where they're strongExecutive relationships, delivery capacity, industry benchmarks, and the safety of a name nobody gets fired for hiring.

Where we winSpeed and cost: weeks and €50K versus months and seven figures. And permanence — consultant knowledge leaves with the consultants; our map lives in the client's hands and keeps updating. A deck describes the work; our twin runs it.

A line that works"A consultancy rents you analysts. We leave you a living map your own team operates — with the first automation already running when we're done."

Plant this question"When the engagement ends, where does the process knowledge live? And what does it cost to keep the deck current next year?"

Consultancies can also be channel, not competition — if the prospect's consultancy is in the room, position us as the tool their programme runs on.

#05 ROI math for defending €50K

Anchor on their numbers, not ours. The pattern from the Prudential call: pick one process they know is slow, quantify the delay, price the delay.

The architect's phrasing, reusable: "If we can reduce it by ten or even twenty days, that has real cash value to the business." The €10M figure is illustrative — always swap in the prospect's own volume. Cash-freed estimate = receivables × (days saved ÷ 365).

Three anchors to price against

#06 Who to sell to, and who blocks

Known blockers

• IT security / data protection — engage early with the security answers from #03; in Germany, ask whether a works council needs to be informed about employee interviews and observation. Raise it before they do — it builds trust.
• Middle managers guarding team time — the executive sponsor must mandate the ~4 hours weekly; if the sponsor won't, that's a qualification red flag, not a negotiation point.

#07 Qualification: the design-partner funnel

Every prospect runs through the same funnel, and "not a design partner" is a routing decision, not a rejection. Source: Strategic Development Partner Evaluation Framework (Chris Fennell, 3 July 2026), adapted for sales use.

The four gates (pass or stop)

Yes/no. One "no" stops the design-partner conversation — no score rescues a failed gate.

Discovery-call questions, mapped to gates

Scope guidance for the first build

Candidate: ______________________

GATES     G1 exec mandate ...... PASS / FAIL
          G2 stability ......... PASS / FAIL
          G3 mutual commitment . PASS / FAIL
          G4 output & reason ... PASS / FAIL   → any FAIL: stop, route to sell-to

STRATEGIC (50%)                    score 1-5 × weight
  S1 Access / network .......... [ ] ×0.12   S4 Brand (risk-adj) .. [ ] ×0.08
  S2 Industry-coverage fit ..... [ ] ×0.12   S5 Ambition .......... [ ] ×0.06
  S3 Partner upside ............ [ ] ×0.12

TECHNICAL (50%)
  T1 Engineering bench ......... [ ] ×0.14   T4 Open / API-first .. [ ] ×0.10
  T2 Process docs / SOPs ....... [ ] ×0.10   T5 Agentic maturity .. [ ] ×0.06
  T3 Data legibility ........... [ ] ×0.10

RISK      R1 size [ ]  R2 compliance [ ]  R3 IP/vendor [ ]  R4 bias [ ]

VERDICT:  ⬤ Build partner   ◐ Maybe / conditional   ○ Sell-to only

#08 Deal mechanics

What we need from the client to start

• An executive sponsor with the mandate to allocate the time
• Three named process owners (one per deep-mapped process)
• Security/procurement contact and their checklist
• Agreement on the ten candidate processes (we co-select in scoping)

Negotiation guardrails

• €50K is fixed. Do not discount. If a prospect pushes, the flexible levers are payment terms and which processes make the ten — never the price. Anything beyond that, escalate to Aqib.
• The 8-week schedule in the offer doc is the default. Stretching to fit their calendar is fine; compressing below 8 weeks is not — the interviews need the weeks.
• Never quote rollout pricing. The line that works: "the engagement is €50K fixed; the rollout is something we design together in the week-8 readout."
• Everything they keep is already in the offer doc (maps, runbooks, ROI overview) — don't add commitments beyond it, especially not fee credits toward a rollout.

#09 Proof points

From a June 2026 call with the Chief Software Architect of a Fortune 500 insurer. Pending clearance — confirm with Aqib before quoting externally; safe to paraphrase in conversation.

Additional signal usable in conversation (not as a quote): a major North American insurer validated the core problem to us directly — they do not understand their own processes well enough to deploy AI confidently.

#10 Open gaps in this playbook