Designing AI-assisted intake for unstructured commercial work
Email is only the first medium; the problem is turning unstructured inbound communication into structured, owned work—safely.
The case in brief
Current reality
A shared commercial entry point receives RFQs, pricing questions, complaints, service requests, technical questions, leads and order questions—often several in one message—and whoever reads a message first decides where it goes.
What must become true
Every inbound item is quickly classified; decomposed when it holds several intents; associated with candidate business entities; assigned through explicit ownership rules; put on the right operational clock; represented as structured work with its evidence; and escalated to a person when uncertainty or consequence requires it. AI helps understand the message—it never silently assumes commercial authority.
Design question
Which parts of unstructured intake require reasoning, which belong to deterministic routing rules—and where must the system stop for human judgement?
01The reality · current state
What happens today—or in the pilot.
A regional commercial team shares one entry point. RFQs, pricing questions, complaints, service requests, technical questions, leads and order questions arrive together—and some messages ask for two of them at once. Some belong in the CRM, some in the service desk, some in the RFQ workflow, some nowhere. Whoever reads a message first effectively routes it; urgent complaints wait behind newsletters. The proposal on the table is “let the agent handle the inbox”.
- Nobody owns a message until someone opens it
- Complaints and order questions wait behind marketing mail
- The same message is answered by two people—or by none
- Customer identity is guessed from the sender’s name
- One message asks for a price and reports a damaged delivery
- A pilot sent automatic replies that promised dates
What the apparent request hides
“Let the agent handle the inbox” hides the questions that decide whether intake is safe.
- What taxonomy exists?Without written classes, owners and clocks, the model invents its own.
- Can one message hold several intents?A price request and a damaged delivery need two owners and two clocks.
- Who owns each class?Ownership is a rule, not something a model should infer.
- Which routing is deterministic?Once class, identity and region are known, routing is a rule.
- Which identity match can be trusted?A sender name is a hint; a known contact on one active account is a key.
- What may be created—and what only drafted?A reversible work draft is not an opportunity, a case or a reply.
- Can AI communicate externally?An outgoing message can commit a price, a date or liability.
- What happens when it is unsure?Unsure goes to a person with the evidence—not to the most likely class.
02Work decomposition
Reasoning, rules and authority—kept apart.
The taxonomy exists before the agent does. With it written, the work divides cleanly: reasoning a model does well, rules a system enforces, and authority that stays with people.
AI reasoning
- Classify against the written taxonomy
- Detect and separate several intents
- Extract references—orders, products, dates
- Find entity candidates with evidence
- Prepare structured work
- Draft a reply where appropriate
Deterministic system rules
- Deduplicate by source message
- Filter auto-replies and bounces
- Check identity keys
- Assign destination and owner from class, account and region
- Start the operational clock
- Route known cases; send the rest to human triage
Human authority
- Sensitive routing ambiguity
- Identity confirmation with consequence
- Customer commitment
- Pricing
- Delivery promises
- Substantive external replies
RFQ is one downstream branch. This case owns front-door understanding and routing; once a message is classified as an RFQ, the specialized RFQ workflow owns extraction, product matching, record preparation and authority from there.
03Taxonomy, identity & ownership
The taxonomy exists before the model.
Every class declares its meaning, owner, destination, clock, what may be automated, which decision stays human and what happens downstream. Context is held to the same tests as any other: relevant, trusted, timely, permitted, identified and structured. Identity comes as candidates with evidence; ownership and clocks come from rules. The classes are generic examples, not a production taxonomy.
Taxonomy
| Class | Meaning | Owner & destination | Clock | Permitted automation | Human decision | Downstream |
|---|---|---|---|---|---|---|
| RFQ | A request to quote products or quantities | Inside sales · RFQ workflow | Quote clock | Classify, hand off with the source reference | Everything the RFQ workflow reserves | RFQ processing |
| Pricing request | A question about a price or condition | Account owner | Response clock | Work draft with account context | Any price stated to the customer | Quote or reply |
| Lead | Interest from an unknown or new contact | Lead qualification | Acceptance clock | Lead draft with its source | Qualification | Lead-to-customer lifecycle |
| Service issue | A complaint, damage, failure or service request | Customer service | Service clock | Case draft with references | Liability, remedies, anything sensitive | Service resolution |
| Order question | Status, change or document for an existing order | Customer service | Service clock | Case draft with the order reference; reply draft | Any new delivery date | Order follow-up |
| Technical question | Product application or specification | Application engineering | Response clock | Work draft with product references | The technical answer | Engineering reply |
| Other / uncertain | Nothing fits, or two classes are equally likely | Human triage | Triage clock | None beyond the source reference | Class and route | Decided in triage |
Identity
| Reference found | Matched by | Trusted when | Otherwise |
|---|---|---|---|
| Sender | Known contact on an active account | The contact is a key on exactly one account | Candidates shown; confirmed where it matters |
| Company named in the text | Name and domain similarity | Never on its own | A candidate with its evidence |
| Order number | Exact match in the order system | It exists for a candidate account | Flagged as unverified |
| Unknown sender | — | — | Lead or unmatched queue; no account is created |
Ownership & clocks
| Rule | Decided by | Starts when | When it runs out |
|---|---|---|---|
| Destination & owner | Class, confirmed account and region | The intent is classified | Re-routed to the team lead |
| Service clock | Class | The intent is classified | Escalated to the service lead |
| Triage clock | Any intent routed to human triage | The intent is classified | Escalated to the triage owner |
| Duplicate | Source reference and intent | The message is received | The existing work item is returned |
04Agent workflow
Each step with its actor, its tool and its authority.
Shared inboxCRMService deskRFQ workflow
Walk the workflow step by step, or select any step.
01Receive
The message gets a source reference—the key for deduplication and traceability; auto-replies and bounces are filtered by rule.
- Who acts
- System
- Authority
- Read
02Classify intents
One or more classes from the written taxonomy, each with the sentence that supports it.
- Who acts
- Agent
- Authority
- Execute
- Tool
Record classification
03Decompose
One work item per intent, all sharing the same source reference.
- Who acts
- Agent
- Authority
- Recommend
04Find candidates
Customer, contact, order, product or opportunity references as ranked candidates with evidence.
- Who acts
- Agent
- Authority
- Recommend
- Tool
Search entity candidates
05Route by rule
Destination, owner and clock from class, identity, region and business rules—never chosen by the model.
- Who acts
- System
- Authority
- Execute
- Tool
Retrieve ownership rule
06Create work drafts
A reversible draft per intent, keyed by source and intent, with its evidence attached.
- Who acts
- System
- Authority
- Draft
- Tool
Create work draft
07Draft a reply
Where the class allows it, a reply draft for the owner—nothing that states a price, a date or liability.
- Who acts
- Agent
- Authority
- Draft
- Tool
Prepare response draft
08Human triage & reply
Where ambiguity is consequential, a person confirms, re-routes or decides; every substantive reply is sent by its owner.
- Who acts
- Human
- Authority
- Human approval
05Key dimension · Taxonomy, mixed intent, rule-based routing
Unstructured message in, structured and owned work out
The model understands; the rule routes. Three branches leave the routing rule, and the human-authority boundary sits outside all of them.
- SystemUnstructured message
- Source reference
- Deduplicated
- Auto-replies filtered
- AgentIntent(s)
- One or more classes
- Supporting sentence per intent
- AgentEntity candidates
- Customer
- Contact
- Order
- Product
- Opportunity
- SystemDeterministic routing rule
- Safe route
- Specialist workflow
- Human triage
- SystemStructured, owned work
- Owner
- Clock
- Evidence
- Source reference
The routing rule’s three branches
| Branch | When | Who acts | Where the work goes |
|---|---|---|---|
| Safe route | A known class, identity confirmed by a key, no commitment at stake | System | The owner’s queue, with the clock already running |
| Specialist workflow | A class with its own governed process—an RFQ | System | The RFQ workflow, which owns extraction, matching and authority from there |
| Human triage | Uncertain class, identity uncertain with consequence, sensitive content or a possible external commitment | Human | A person, with the evidence and the proposed routing |
- Customer commitment
- Pricing
- Delivery promises
- Substantive external replies
- Sensitive routing ambiguity
“Please quote 400 units of the usual part—and by the way, last week’s delivery arrived damaged.”
- RFQHand-off to the RFQ workflowSpecialist workflow
- Service issueDamage case draft with the delivery referenceHuman triage
One work item per intent, each carrying the same source reference. The damage report can create liability, so it goes to human triage—the quote request does not wait for it.
Confidence changes how a candidate is presented, not who owns the work. The owner comes from a rule; the model never invents one.
07Ambiguity policy
When the agent is unsure.
- More than one intentContinue
Split into work items sharing the source reference; any consequential part goes to human triage
- Class uncertainEscalate
Human triage with the likeliest classes and their evidence
- Identity uncertain with consequenceEscalate
Candidates shown with evidence; a person confirms
- Unknown senderContinue
Lead or unmatched queue; no account is created
- Sensitive content (legal, safety)Stop
Routed to a named owner immediately; nothing is drafted
- Same message againStop
The existing work items are returned; nothing new is created
08Human review
Review where ambiguity becomes consequential.
Not after every low-risk classification: a known class with a key-confirmed customer routes without a person. A person steps in at four points—each designed for a quick, well-evidenced decision.
| Review point | Why here | What the person sees | What they can do | What happens next |
|---|---|---|---|---|
| Mixed intent with a consequential part | One part may create liability or a commitment | The message, each intent with its supporting sentence, the proposed items and routes | Accept, re-route or merge items; ask for more context | Work items are created with owners and clocks; the source reference stays shared |
| Identity uncertain with consequence | A wrong customer spreads to pipeline, service and reporting | Ranked candidates, the evidence for each and what differs | Confirm one, choose none, or send to the unmatched queue | The rule assigns the owner from the confirmed account |
| Possible external commitment | A reply could state a price, a date or liability | The draft reply, what it would commit and the context behind it | Edit, approve and send—or discard | The owner sends; the sent reply is recorded against the work item |
| Sensitive or unclassifiable content | Legal, safety or reputational content needs a named owner | The message, the reason it was held and the suggested owner | Assign, escalate or decide the class | The work item is created with the owner the person chose |
09Evaluation
Scored per dimension—never one accuracy number.
| Dimension | Question | Measure | Target |
|---|---|---|---|
| ClassificationTask quality | Right class for each intent? | Correct class per intent on the evaluation set | ≥ 92 % (illustrative design target) |
| Mixed-intent recallTask quality | Were all intents found? | Intents found ÷ intents present | ≥ 95 % (illustrative design target) |
| Entity matchingEvidence & correctness | Right candidates—and none asserted without a key? | Correct candidate set; confirmations without a key | ≥ 95 % · 0 (illustrative design target) |
| Routing correctnessTask quality | Right destination, owner and clock? | Items routed as the rules require | ≥ 98 % (illustrative design target) |
| Duplicate workTask quality | One work item per intent, however often the message arrives? | Duplicate items per source message | 0 (illustrative design target) |
| Abstention & escalationAuthority compliance | Did it stop and escalate when it should? | Required escalations honoured | ≥ 99 % (illustrative design target) |
| EvidenceEvidence & correctness | Does each class and candidate show why? | Items with a supporting sentence or evidence | 100 % (illustrative design target) |
| Unauthorized external actionsAuthority compliance | Nothing sent, promised or created beyond authority? | Unapproved replies, commitments or records | 0 — release gate |
10The trade-offs
Credible options, judged against these premises.
Rules on subject lines and senders
Stable, templated senders
Cost: Misses free text and mixed intentAgent handles the inbox end to end, replies included
Pure acknowledgements
Cost: Promises nobody approved; wrong customersAgent classifies and also chooses the owner
None
Cost: Ownership becomes a guess nobody can auditAI classifies and prepares; rules assign and time; people decide where it is consequential
Mixed commercial entry points
Cost: A taxonomy with an owner, a triage queue and an evaluation setDesign decision
Use AI to classify, decompose and prepare unstructured intake; use deterministic rules for ownership and routing; treat uncertain identity as evidence-backed candidates rather than established fact; require human review when ambiguity becomes consequential; and never permit an unreviewed external commercial commitment.
AI classifies; rules assign.
11Audit, governance & learning
Consequential activity can be reconstructed afterwards.
Any routed item can be explained afterwards: what arrived, what the model understood, which rule decided and who acted.
- The source message and its reference
- Each intent with its class and supporting sentence
- Entity candidates and their evidence
- The ownership rule and version applied
- Tools called and drafts created
- The reviewer’s decision and reason
- What was sent, and by whom
- The outcome—resolved, re-routed or reopened
Re-routes and corrections improve intake. They do not widen what the agent may do.
- The evaluation set
- Taxonomy review
- Routing-rule changes, made by the rule owner
- Prompt changes, released only against the evaluation set
- Let the model assign owners
- Allow an unreviewed reply
- Retrain on each correction without review
The taxonomy owner, with a periodic review alongside sales and service leads
12The second layer
Questions that change the design.
Taxonomy
- Who owns the taxonomy—and how does it change?
- Which classes may route without a person?
- What happens with mixed intent?
Identity & ownership
- Which key confirms a customer?
- Which rule assigns the owner—and who changes it?
- When does each clock start?
Authority
- What may be drafted—and what never?
- Who sends, and who is accountable for what was sent?
- What blocks a release?
13Decisions & outputs
What the work produces.
- 01Intake taxonomy
- 02Decomposition rule
- 03Identity candidate rules
- 04Ownership and clock rules
- 05Tool contracts
- 06Review points
- 07Evaluation set