This part­ner insight was authored by Parid­hi Gup­ta, Chief Mar­ket­ing Offi­cer at SEW​.AI. 

Afford­abil­i­ty was once viewed pri­mar­i­ly through the lens of cus­tomer assis­tance and reg­u­la­to­ry pol­i­cy. It now shapes board­room dis­cus­sions about rev­enue resilience, cus­tomer trust, oper­a­tional per­for­mance, and long-term grid invest­ment. Ener­gy demand is grow­ing, elec­tri­fi­ca­tion is accel­er­at­ing, and house­hold bud­gets face increas­ing pres­sure, mak­ing afford­abil­i­ty a strate­gic pri­or­i­ty that touch­es every part of the business.

Afford­abil­i­ty strate­gies have his­tor­i­cal­ly focused on cus­tomers whose hard­ship has already sur­faced, so the response comes after a pay­ment is missed or an account enters col­lec­tions, by which point ordi­nary finan­cial strain has often hard­ened into cri­sis. Lead­ing util­i­ties are now rec­og­niz­ing that the great­est oppor­tu­ni­ty lies much earlier.

Very lit­tle of this shows up in a sin­gle bill. It builds from a series of events that are unre­mark­able in iso­la­tion. Con­sump­tion begins to climb and pay­ment behav­ior changes, sea­son­al demand dri­ves the bill high­er, and even­tu­al­ly the cus­tomer calls with a billing ques­tion or becomes less engaged across dig­i­tal chan­nels. Each is a rou­tine oper­a­tional event on its own, and only togeth­er do they reveal emerg­ing afford­abil­i­ty risk, which is what makes it a blind spot.

The chal­lenge is inter­pre­ta­tion, and ear­li­er inter­ven­tion depends on turn­ing scat­tered indi­ca­tors into a pic­ture front­line teams can act on.

Fig­ure 1. Source: EIA RECS / NCLC analysis.

The Busi­ness and Soci­etal Cost of Affordability

For house­holds, afford­abil­i­ty now func­tions as a mea­sure of finan­cial resilience. Ris­ing ener­gy costs are forc­ing fam­i­lies into dif­fi­cult trade-offs, with near­ly one in four house­holds report­ing that they have reduced spend­ing on essen­tials such as food or med­i­cine to pay ener­gy bills. Low-income house­holds car­ry an ener­gy bur­den near­ly three times high­er than oth­ers, and that bur­den affects health, safe­ty, and qual­i­ty of life beyond just bill payment.

Fig­ure 2. Source: NEA­DA Win­ter Out­look 2025 – 26

Util­i­ties feel that pres­sure from the oth­er side. Ris­ing arrears land on col­lec­tions and cus­tomer ser­vice oper­a­tions while strain­ing assis­tance pro­gram capac­i­ty and com­pli­cat­ing reg­u­la­to­ry com­mit­ments. Recov­er­ing unpaid bal­ances is only part of the chal­lenge, which extends to the cost of learn­ing about a customer’s sit­u­a­tion late, and that can lead to high­er oper­a­tional costs, increased ser­vice com­plex­i­ty, and reduced cus­tomer trust.

Fig­ure 3. Source: U.S. EIA Res­i­den­tial Util­i­ty Dis­con­nec­tions Report, April 2026

The Intel­li­gence Gap Behind the Afford­abil­i­ty Gap

Util­i­ties have invest­ed heav­i­ly in cus­tomer pro­grams, pay­ment assis­tance, dig­i­tal engage­ment, and self-ser­vice capa­bil­i­ties, and those invest­ments remain essen­tial. Afford­abil­i­ty is now expos­ing a gap in intel­li­gence along­side them.

Most util­i­ties already hold the infor­ma­tion need­ed to rec­og­nize emerg­ing pres­sure. Pay­ment his­to­ry sits in the billing sys­tem and detailed con­sump­tion insight in advanced meter­ing infra­struc­ture, while cus­tomer ser­vice records car­ry billing inquiries and stat­ed pay­ment con­cerns. Dig­i­tal chan­nels reflect chang­ing engage­ment pat­terns, and the assis­tance pro­grams hold the cus­tomers who have already sought support.

Each sys­tem records the part of the cus­tomer it was built to record, and each is accu­rate with­in its own scope. None shows a house­hold approach­ing hard­ship. Con­nect­ing the sys­tems alone does not close that dis­tance, because read­ing one function’s data along­side another’s is inter­pre­tive work, and con­nect­ed data is not the same thing as con­nect­ed intelligence.

Not Every Cus­tomer Needs the Same Help

A per­sis­tent mis­con­cep­tion in afford­abil­i­ty is that an unpaid bill means the same thing wher­ev­er it appears. Some cus­tomers are finan­cial­ly stretched and still com­mit­ted to pay­ing. Oth­ers qual­i­fy for assis­tance and nev­er enroll, either unaware the pro­gram exists or assum­ing they are inel­i­gi­ble. A third group is absorb­ing tem­po­rary bill shock dri­ven by sea­son­al weath­er or chang­ing ener­gy con­sump­tion, and a fourth has the capac­i­ty to pay and sim­ply puts the util­i­ty bill last. Iden­ti­cal treat­ment makes oper­a­tions sim­pler at the cost of results, and lead­ing util­i­ties are replac­ing one-size-fits-all strate­gies with approach­es that read the con­text behind the account.

• The finan­cial­ly stretched cus­tomer is best served by out­reach before arrears build, usu­al­ly mean­ing proac­tive enroll­ment in assis­tance and a flex­i­ble pay­ment arrangement.
• The sta­ble cus­tomer needs less. Dig­i­tal self-ser­vice, clear billing, and time­ly noti­fi­ca­tions are enough to sus­tain the pay­ment behav­ior already in place.
• The under­served cus­tomer may be reach­able only through com­mu­ni­ty agency coor­di­na­tion, mul­ti­lin­gual engage­ment, or a sim­pli­fied enroll­ment pathway.
• The high­er-risk cus­tomer requires tar­get­ed col­lec­tions and clear pay­ment expec­ta­tions, applied in a way that pro­tects rev­enue with­out harm­ing the relationship.

His­tor­i­cal report­ing and sta­t­ic busi­ness rules will not iden­ti­fy which sit­u­a­tion an account is in. It takes intel­li­gence that reads sig­nals togeth­er con­tin­u­ous­ly and rec­om­mends an action while there is still time for it to matter.

Turn­ing Afford­abil­i­ty into an Oper­at­ing Model

Util­i­ties make thou­sands of deci­sions every day that shape afford­abil­i­ty. Offer­ing a pay­ment arrange­ment is one. So is mov­ing an account onto a rate plan that suits it bet­ter, notic­ing that a con­sump­tion spike traces to a new­ly installed EV instead of a house­hold in trou­ble, or flag­ging a cus­tomer who will prob­a­bly need help before the next bill. These get made in the ordi­nary course of work and togeth­er form a utility’s afford­abil­i­ty strat­e­gy, whether or not any­one has described it that way.

Those deci­sions are dis­trib­uted across func­tions that each hold a dif­fer­ent piece of the account. The incon­sis­ten­cy that fol­lows is what the cus­tomer expe­ri­ences. From where the cus­tomer sits, there is only one utility.

The util­i­ties mak­ing the most progress have start­ed treat­ing afford­abil­i­ty as an enter­prise capa­bil­i­ty, mea­sur­ing suc­cess by how reli­ably the right cus­tomer is iden­ti­fied, engaged, and supported.

Peo­ple + Ver­ti­cal AI sup­ports that mod­el. Ver­ti­cal AI does not replace the judg­ment of a cus­tomer ser­vice rep­re­sen­ta­tive, a col­lec­tions spe­cial­ist, or a com­mu­ni­ty part­ner; it gives that judg­ment more to work with, con­nect­ing cus­tomer behav­ior, pay­ment his­to­ry, ener­gy con­sump­tion, pro­gram eli­gi­bil­i­ty, tar­iff struc­tures, and reg­u­la­to­ry con­text so that every cus­tomer-fac­ing team is decid­ing from the same picture.

In prac­tice, that looks like a lev­elized pay­ment plan rec­om­mend­ed before bill shock turns into delin­quen­cy or an assis­tance pro­gram sur­faced for a cus­tomer who nev­er knew it exist­ed. Afford­abil­i­ty shifts from reac­tive to proactive.

Proof in Prac­tice: DTE Energy

Low-income assis­tance is struc­tured much the same way across the indus­try. A com­mu­ni­ty action agency, rather than the util­i­ty, holds the cus­tomer rela­tion­ship and man­ages the paper­work for state pro­grams like Michigan’s ener­gy assis­tance fund. The utility’s own col­lec­tions process runs in par­al­lel, on a sep­a­rate sys­tem. When the two do not share data in real time, a pledge to hold a dis­con­nec­tion while an appli­ca­tion is processed can sit in the agency’s records with­out yet appear­ing in the utility’s time­line. The gap is struc­tur­al. It has shaped how assis­tance gets deliv­ered across much of the sec­tor, and it is not a fail­ure spe­cif­ic to any one utility.

The SAP and SEW​.AI part­ner­ship address­es that gap direct­ly. SAP sup­plies the enter­prise foun­da­tion, con­nect­ing cus­tomer, finan­cial, and oper­a­tional process­es so that a sin­gle view of the account car­ries across billing, col­lec­tions, ser­vice, pay­ments, and assis­tance. On top of that foun­da­tion, SEW​.AI applies Peo­ple + Ver­ti­cal AI, which turns the enter­prise data into real-time deci­sion intel­li­gence. Cus­tomer ser­vice teams, col­lec­tions spe­cial­ists, field oper­a­tions, and com­mu­ni­ty part­ners all draw on it, using it to iden­ti­fy need ear­li­er, per­son­al­ize engage­ment, and coor­di­nate the next best action across the cus­tomer lifecycle.

DTE Ener­gy closed it by chang­ing where the two sys­tems meet. Agen­cies that deter­mine eli­gi­bil­i­ty now work from a shared, live view of pledge sta­tus, tied direct­ly into the utility’s dis­con­nec­tion and col­lec­tions sys­tems. An approved pledge stops a shut­off the moment it is entered, with­out wait­ing for some­one to relay it. Ver­i­fi­ca­tion and approval hap­pen inside the same plat­form the agency already uses. Over the past three years, that change in sequenc­ing has helped direct more than $160 mil­lion in finan­cial assis­tance and pre­vent­ed shut­offs for more than 397,000 house­holds. DTE Energy’s broad­er ener­gy effi­cien­cy pro­grams, aimed at low­er­ing usage rather than sub­si­diz­ing bills direct­ly, have reached 1.8 mil­lion res­i­den­tial cus­tomers and 500,000 businesses.

DTE Ener­gy offers the indus­try a replic­a­ble mod­el of util­i­ty-led afford­abil­i­ty lead­er­ship, one built on coor­di­na­tion and con­nect­ed intel­li­gence between the sys­tems and agen­cies that deter­mine whether eli­gi­ble house­holds are reached before the shut­off notice arrives.

The Next Afford­abil­i­ty Chal­lenge Is Already Tak­ing Shape

Afford­abil­i­ty will get hard­er over the next decade. Elec­tric vehi­cles, heat pumps, dis­trib­uted ener­gy resources, indus­tri­al elec­tri­fi­ca­tion, and AI-pow­ered data cen­ters are all chang­ing the shape of elec­tric­i­ty demand, and meet­ing it will require sus­tained grid invest­ment bal­anced against afford­abil­i­ty, reli­a­bil­i­ty, and long-term sustainability.

Util­i­ties are ask­ing how to rec­og­nize chang­ing cir­cum­stances in time to keep afford­abil­i­ty from becom­ing a cri­sis at all, which takes intel­li­gence capa­ble of inter­pret­ing cus­tomer, oper­a­tional, and rev­enue sig­nals in con­text and on an ongo­ing basis.

Some of this is already under­way. Util­i­ties can now iden­ti­fy eli­gi­bil­i­ty proac­tive­ly, where it used to be some­thing cus­tomers had to dis­cov­er for them­selves, and reg­u­la­tors are plac­ing greater empha­sis on whether eli­gi­ble house­holds actu­al­ly par­tic­i­pate, along­side the fund­ing of the pro­grams them­selves. Def­i­n­i­tions of equi­ty have expand­ed past income to include lan­guage, acces­si­bil­i­ty, dig­i­tal inclu­sion, and the abil­i­ty to engage every cus­tomer in the way they are most like­ly to respond. Stan­dard out­reach is giv­ing way to personalization.

Peo­ple make the deci­sions that build trust; the technology’s con­tri­bu­tion is get­ting those deci­sions to hap­pen ear­li­er and with more con­text behind them. Util­i­ties that con­sis­tent­ly decide ear­li­er strength­en cus­tomer afford­abil­i­ty, and they strength­en cus­tomer trust, oper­a­tional resilience, and long-term busi­ness per­for­mance along with it.

Parid­hi Gup­ta is the Chief Mar­ket­ing Offi­cer at SEW​.AI.